aei housing market indicators (hmi) › rs › 475-pbq-971 › images ›...
TRANSCRIPT
AEI Housing Market Indicators (HMI)
Edward Pinto ([email protected]) and
Tobias Peter([email protected])
AEI Housing Center
AEI.org/housing
April 29, 2019
1
We grant permission to reuse this presentation, as long as you cite as the source:
AEI Housing Center, www.aei.org/housing.
AEI Housing Market Indicators: An Introduction• Provide accurate and timely metrics for the housing market. These include:
• Mortgage Risk/Leverage– Particular focus is paid to agency first-time buyer volume and risk
• House prices and appreciation trends
• Housing sales – New and existing sales whether institutionally financed, cash, or other-financed
• Months inventory
• The housing market is influenced by many different levers. All need to be
connected and considered to better understand market trends.– AEI HMI adds geography and price points to the broad set of metrics:
• Geography: national, state, and selected metros – House prices down to the census tract
• Price points: low, low-medium, medium-high, and high price tiers– Price tiers are defined based on the availability of leverage for borrowers at certain geographies.
• Expanded Housing Market Indicators use and connect many different datasets:• HMDA• Public Records Data• National Mortgage Risk Index (agency MBS data)• CoreLogic’s LLMA and Black Knight’s McDash (servicer data)• Fannie Mae’s Loan Performance data and Freddie Mac’s Loan-Level Data (acquisition data)• FHA Snapshot data (endorsement data)• Data from Zillow on existing home sales and unique listings
• Advantages of the AEI Housing Market Indicators:– Most in-depth resource for key housing data and trends (select data available online for download)
– Accurate, timely, and in-depth coverage of purchase trends
– Connects the dots for many housing indicators, yielding the most comprehensive analysis of the
housing market and boom/bust cycles
• Detailed Methodologies are available after “Reminaing Briefing Dates” slide. 2
HMI Key Takeaways: Tracking the Home Price Boom
• Only AEI’s Housing Market Indicators offer consistent and market-based measures.
– The media’s presentation of housing data has been very confusing over the last 6 months.
– There has been a media whiplash from seller’s to buyer’s back to seller’s market.
– The media has also presented conflicting messages from existing and new home sales.
• Given the interest rate drop since November, the rate of house price appreciation (HPA) is again rising.
– At our January 7, 2019 briefing we predicted: “Given the rate drop since November, we would expect a modest
pickup in the HPA rate.”
– Preliminary numbers for March 2019 indicate national HPA of 4.4% (yoy), which is up from 4.1% in January 2019.
– HPA remained strongly bifurcated. House prices in the low price tier appreciated at 6.8% (yoy), while prices in the
high price tier fell 1.4% (yoy).
– We expect HPA to remain strongly bifurcated, with HPA in the low tier continuing at a rate of 2 times wage growth.
• Mortgage risk continued to increase, but there may be a silver lining:
– The composite Purchase National Mortgage Risk Index (NMRI) set a series’ high for the month of January.
– The index was up 0.4 percentage point from January 2018.
– It’s too early to tell, but changes to FHA’s Total Scorecard may soon be reigning in some of FHA’s worst practices.
• GSE programs unrelated to buying a primary residence have seen large declines over the last year :
– Cash out refi volume is down 38% from January 2018.
– Investor and Second Home volume is down 36% and 19% respectively over the same time period.
– These trends should make it easier to wind these programs down.
• Housing Market Indicators for the Nation and 60 largest Metros are now available on the web:
– Visit http://www.aei.org/multimedia/national-and-metro-housing-market-indicators/ to explore.
– All data are available for download in an Excel file and will be updated quarterly.
• AEI’s new construction sales numbers will be covered in depth in a briefing on May 9th at 11am EST.
– Initial results indicate that they are generally somewhat higher than the Census Bureau’s new home sale counts.
– Metros such as Austin and Raleigh, with a high level of employment growth combined with a high level of new
construction sales (particularly at the entry-level) have been better at keeping house price appreciation in check .3
Whiplash: From a Buyer’s Market to a Sellers’ Market
• After finally admitting in late 2018 that there is a house price boom, the
media quickly changed its tune in early 2019 to: It’s a buyers’ market and
this will help entry-level buyers– “Real estate agents say housing market is favoring buyers” Housing Wire, 3.25.19
– “Out of the Seller’s Market, Into the Buyer’s Market” DSNews, 3.21.19
– “Homebuyers gain edge in this year's housing market” USA Today, 3.7.19
– “A Buyer’s Market? Hopes Rise With Falling Rates, More Homes for Sale” WSJ,
3.31.19
• But less than a month later, the tune has changed once again:– “2019's Housing Market Is Likely to Be Stronger Than We Thought—Here's Why”,
Realtor.com, 4.23.19
• Avoid this media whiplash by following AEI’s Housing Market Indicators:– On January 7, 2019 we advised:
• “Rumors of the end of Housing Boom 2.0 are greatly exaggerated. Unsustainable rate of house
price appreciation continues, particularly for the low price tier. Inventories remain very tight at
the lower end, which implies that house prices will continue to increase, thereby worsening
affordability.”
• “Loosening mortgage underwriting standards continue with FHA and First-Time Buyers again
setting the pace, leading to greater market bifurcation.”
• The data we present today confirms our informed, market-based
predictions4
Update: NMRI for Agency Home Purchase Loans
*Change from January 2013 to January 2019.
Source: AEI Housing Center, www.AEI.org/housing. RHS is Rural Housing Service.
Composite index has consistently been trending up since mid-2013, with FHA leading the way. While index growth has paused, it is too early to confirm a definite trend.
FHA’s March 2019 Total Scorecard changes are expected to result in a decline in the FHA and composite indices. We will report the actual outcome once April and May
data become available.
5
4%
8%
12%
16%
20%
24%
28%
32%
4%
8%
12%
16%
20%
24%
28%
32%
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17 Sep-18
Stressed default rate
FHA share of purchase loans: Nov-18 22.2%; Dec-18 22.9%; Jan-19 23.9%
FHA: +7.5 ppts, from 21.1 to 28.6%*
Composite: +1.8 ppts, from 11.4% to 13.2%*
VA: +1.7 ppts, from 10.9% to 12.6%*
Fannie: +3.0 ppts, from 5.3% to 8.3%*
Freddie: +1.8 ppts, from 4.8% to 6.6%*
FTB Purchase Loan NMRI: Credit Easing Continues
Note: Includes all types of NMRI purchase loans (primary owner-occupied, second home, and investor loans).Source: AEI Housing Center, www.AEI.org/housing.
The First-time Buyer MRI continued to increase. FHA’s First-time Buyer MRI stood at 29.0% in January, up 1.6 ppts from a year earlier. While the rate of increases for FHA and Fannie has slowed over recent months, these increases are coming off of higher
levels. We remain hopeful that FHA’s Total Scorecard changes will reign in some of the worst practices.
6
-1.0%
-0.5%
0.0%
0.5%
1.0%
1.5%
2.0%
2.5%
-1.0%
-0.5%
0.0%
0.5%
1.0%
1.5%
2.0%
2.5%
Feb-14 Feb-15 Feb-16 Feb-17 Feb-18
Change from 12 months earlier, in percentage points
FHA FTBs
Fannie FTBs
Easing
TighteningFreddie FTBs(blue)
All Agency FTBs
Fannie’s Unhealthy Competition with FHA Is Moving
Both Out the Risk Curve
Note: Includes all types of NMRI purchase loans (primary owner-occupied, second home, and investor loans).
Source: AEI Housing Center, www.AEI.org/housing.
A closer look at the risk distribution of Fannie and Freddie reveals that while Freddie has loosened underwriting moderately, Fannie has been more aggressive. The share pickup for
Fannie, Freddie, and FHA shows that Fannie is increasingly competing with FHA for loans with a risk score between 8-24. This segment currently accounts for 35% of both Fannie and FHA’s business. FHA is replacing this lost business with much higher risk loans, those with an MRI
of greater than 32%.
7
0%
10%
20%
30%
40%
50%
0-4 4-8 8-12 12-16 16-20 20-24 24-28 28-32 >32
Mortgage Risk Index
January 2019 Distribution Freddie Fannie FHA
Percent of loans
-10%
-5%
0%
5%
10%
15%
0-4 4-8 8-12 12-16 16-20 20-24 24-28 28-32 >32
Mortgage Risk Index
Change in Distribution, January 2017 to January 2019Freddie Fannie FHA
Percentage points
Origination Shares by Credit Score Bin,
Purchase Loans
Source: AEI Housing Center, www.AEI.org/housing.
The story in the media has been of too tight credit holding back buyers. The reality is the long-term trend has been towards looser credit and higher levels of volume until very recently (especially for FTB). Especially noteworthy is the large
influx of borrowers with subprime credit scores below 660.
8
10%
15%
20%
25%
30%
10%
15%
20%
25%
30%
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17 Sep-18
660-699
700-739
<660
740-779
≥780
Percent of loans
FHA Total Scorecard Change
Note: Includes all types of NMRI purchase loans (primary owner-occupied, second home, and investor loans).Source: MBA Weekly Application Survey, and AEI Housing Center, www.AEI.org/housing.
The latest MBA Weekly Application Survey continues to indicate an application decline of about 10% since FHA made Total Scorecard changes in March 2019. FHA and VA
application shares usually track fairly closely, especially over short time periods. Since the changes took effect, FHA application volume, relative to VA volume, is down about
13%. This decline may indicate that FHA is getting the desired result from the Total Scorecard changes.
9
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
3.08.19 3.15.19 3.22.19 3.29.19 4.5.19 4.12.19 4.19.19
Ratio of FHA and VA Apps (VA Indexed to 1.00)
For the week ending 4.19.19, FHA application volume was down by 11% relative to VA volume, compared to a decline of 17% for the prior week. FHA volume is benchmarked to VA volume for the weeks ending 3.8.19 and 3.15.19.
3.15.19 was last application date prior to the 3.18.19 effective date of the FHA Total Scorecard change.
No-Cash-Out Refi and Cash-Out Refi NMRIs
Despite a Refi NMRI near its series’ high, expected refi stressed defaults have declined slightly in recent months. While individual loans are becoming riskier (esp.
cash outs), rising rates in 2017 and 2018 have driven down refi volume. Cash-Out NMRI is largely driven by growth in volume and risk on FHA and VA guaranteed loans.
10Source: AEI Housing Center, www.AEI.org/housing.
6%
8%
10%
12%
14%
16%
6%
8%
10%
12%
14%
16%
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17 Sep-18
Refi NMRI
Cash-out NMRI
No-cash-out NMRI
Stressed default rate
Red markers show January stressed default rate in each year.
Expected Refi Stressed Defaults by Origination Date
Jan 2013 – 50,728Jan 2014 -- 16,981Jan 2015 – 20,830Jan 2016 – 20,822Jan 2017 – 23,266Jan 2018 – 20,822Jan 2019 – 13,722
Greater House Price Volatility at the Lower End
11
In the past, increasing leverage has fueled unsustainable house price trends. Since the advent of expanded “affordable housing” efforts, these trends have become more pronounced at the lower end of the market with higher peaks and lower troughs.
Currently the low price tier is increasing at 2 times wage growth.
Dovish Fed = Monetary Punchbowl Getting Spiked Again
12
As predicted in our last briefings, the year-over-year rate of house price appreciation (HPA) has picked up again. The national rate of HPA for March was 4.4%. This is down from its
recent peak of 5.7% in March 2018, but up from the 4.1% in January 2019. This coincides with recent movements in mortgage rates. Rates, after having increased by 116 basis points from
September 2017 to early November, have since declined by 74 basis points.
Note: Data are for the entire country. Data for 2019:Q1 are preliminary.
Source: AEI Housing Center, www.AEI.org/housing.Source: Freddie Mac
5.7%
4.4%
0%
1%
2%
3%
4%
5%
6%
7%
Jan
-13
Jul-
13
Jan
-14
Jul-
14
Jan
-15
Jul-
15
Jan
-16
Jul-
16
Jan
-17
Jul-
17
Jan
-18
Jul-
18
Jan
-19
Year-over-Year Rate of House Price Appreciation
4.94
4.20
3.00
3.50
4.00
4.50
5.00
5.50
10
/3/2
01
3
4/3
/20
14
10
/3/2
01
4
4/3
/20
15
10
/3/2
01
5
4/3
/20
16
10
/3/2
01
6
4/3
/20
17
10
/3/2
01
7
4/3
/20
18
10
/3/2
01
8
4/3
/20
19
30-year Fixed Rate Mortgage
Red markers show late April rate in each year.
National House Price Appreciation (HPA) by Price Tier
13
In March, the low price tier not only continued, but reaccelerated its unsustainable trend (left
panel). In March 2019, house prices in the low price tier appreciated at 6.8% year-over-year
(yoy) - the strongest rate of growth since January 2016 (right panel). In the low-medium and
medium-high tiers, they increased at 4.2% and 4.1%, respectively. House prices in the high tier
(about 8% of the market) continued to decline at a yoy rate of 1.4%.
Note: Data for 2019:Q1 are preliminary. Price tiers are set at the metro level and are defined as follows: Low: all sales at or below the 40th percentile of FHA sales
prices; Low-Medium: all sales at or below the 80th percentile of FHA sales prices; Medium-High: all sales at or below the 125% of the GSE loan limit; and High: Rest.
HPAs are smoothed around the times of FHFA loan limit changes.
Source: AEI Housing Center, www.AEI.org/housing.
90
100
110
120
130
140
150
160
Home Price Appreciation by Tier
Overall
Low
Low-Med
Med-High
High
Index: Jan-2012 = 100
-4%
-2%
0%
2%
4%
6%
8%
10%
Year-over-Year HPA - by Tier
Overall
Low
Low-Med
Med-High
High
Wage Growth relative to House Price Growth
Source: Current Population Survey, Bureau of Labor Statistics, and Federal Reserve Bank of Atlanta Calculations. AEI Housing Center, www.AEI.org/housing.
Affordability has worsened as gains in house prices have far outpaced gains in wages. This wedge between prices and wages is most pronounced for the low price tier. With
house price appreciation picking up steam again, this wedge will only further increase. This trend has been worsened through the availability of leverage, which has enabled
less credit-worthy buyers to stay in the market and drive up prices.
14
90
100
110
120
130
140
150
160
Jan
-12
Jul-
12
Jan
-13
Jul-
13
Jan
-14
Jul-
14
Jan
-15
Jul-
15
Jan
-16
Jul-
16
Jan
-17
Jul-
17
Jan
-18
Jul-
18
Jan
-19
Low Tier House Price Index
Wage Index
Index: Jan-2012 = 100
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
Jan
-13
Jul-
13
Jan
-14
Jul-
14
Jan
-15
Jul-
15
Jan
-16
Jul-
16
Jan
-17
Jul-
17
Jan
-18
Jul-
18
Jan
-19
Low Tier House Price Appreciation
Wages
Cumulative Growth Annual Growth Rate
Shrinking the GSE Footprint: Administrative Reform Options
Source: AEI Housing Center, www.AEI.org/housing.
Administrative steps to shrink the GSE footprint should include phasing out programs unrelated to financing a primary residence. Interestingly, these programs have seen
large declines in volume since last January 2018 making them easier to wind down now. Early 2018 was about the time policy proposals gained currency which suggested
administrative action as a way to reduce the GSEs’ bloated footprints. Cash out refi volume is down 38% from January 2018. Investor and Second Home volume is down
36% and 19% respectively over the same time period.
15
0
10,000
20,000
30,000
40,000
50,000
60,000
70,000
Cash Out Investor Second
Jan-18
Jan-19
Decline Jan-2018to Jan-2019:
Cash Out: -38%Investor: -36%Second Homes: -19%
# of GSE loans
16
Try Out the National and Metro Housing Market Indicators
Interactive ToolVisit http://www.aei.org/multimedia/national-and-metro-housing-market-indicators/ to explore
Housing Market Indicator data for the nation and the 60 largest US metros using our new interactive tool. All data are available for download in an Excel file and will be updated
quarterly.
17
Examples from the National and Metro Housing Market
Indicators Interactive Tool
Visit http://www.aei.org/multimedia/national-and-metro-housing-market-indicators/ to explore Housing Market Indicator data for the nation and the 60 largest US metros using our new interactive tool. All data are available for download in an Excel file and will be updated
quarterly. Here are two examples for the Washington, DC metro:
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
20
13
:Q1
20
13
:Q3
20
14
:Q1
20
14
:Q3
20
15
:Q1
20
15
:Q3
20
16
:Q1
20
16
:Q3
20
17
:Q1
20
17
:Q3
20
18
:Q1
20
18
:Q3
Overall
Entry Level
Move-Up
Months’ Supply: Washington, DC Metro
Source: AEI Housing Center, www.AEI.org/housing.
0%
5%
10%
15%
20%
25%
30%
20
12
:Q1
20
12
:Q3
20
13
:Q1
20
13
:Q3
20
14
:Q1
20
14
:Q3
20
15
:Q1
20
15
:Q3
20
16
:Q1
20
16
:Q3
20
17
:Q1
20
17
:Q3
20
18
:Q1
20
18
:Q3
Overall
Entry level
Move up
New Construction Share of Sales:Washington, DC Metro
18
AEI New Construction Sales vs
Census Bureau New Residential Sales
Comparing AEI’s new construction sales to the census bureaus’ shows that AEI sales generally exceed Census totals (adding ~280,000 units or about 8% to Census totals over
the period). However, the difference between both series has been narrowing. AEI’s series tracks home sales through the public records, while the Census Bureau tracks
them through builder surveys, which can be incomplete or suffer from survey bias.
Source: AEI Housing Center, www.AEI.org/housing, and U.S. Census Bureau.
0
100,000
200,000
300,000
400,000
500,000
600,000
700,000
2012 2013 2014 2015 2016 2017 2018
Census new home sales
AEI New construction sales
# of sales
House Price Appreciation (HPA), New Construction, &
Employment Growth: 100 Largest MetrosMetros such as Austin and Raleigh, with a high level of employment growth combined with a high level of new construction sales (particularly at the entry-level) have been
better at keeping house price appreciation in check. The same applies to metros with medium employment growth. This relationship is missing in metros without much
employment growth.
19Source: AEI Housing Center, www.AEI.org/housing.
20
Try Out the NMRI Interactive Tool
Visit https://www.aei.org/housing/mortgage-risk-index/ to explore National Mortgage Risk Index data using our new interactive tool.
Scroll over “Indexes & Indicators” and then click on “Mortgage Risk Index” on the drop down menu
Latest data will go live at 10 am on the day of the briefing call.
Remaining Briefing Dates for 2019
• Announcing: new briefing on new construction and housing supply metrics to take place on Thursday, May 9.
• Next briefing on Housing Market Indicators on Tuesday, May 28.
• The remaining briefings for 2019 are listed below:
21
• All briefings take place at 11 AM ET.
Thursday May 9
Tuesday May 28
Monday July 1
Monday July 29
August – no briefing
Monday September 30
Monday October 28
Monday November 25
Monday January 6, 2020
Please note that this month’s appendix slides have not been updated.
Price Tier Methodology
• Goal: create leverage-based price tiers.
• Rational: segmenting the market by price tier is important because housing policies, new
construction activity, and access to leverage vary by these price tier. Thus, these factors can
create very different home price appreciation trends depending on the price tier.
• 4 Price Tiers:– Low: all sales below the 40th percentile of FHA sales prices
– Low-medium: all sales at or below the 80th percentile of FHA sales prices
– Medium-high: all sales at or below 125% of the GSE loan limit
– High: all other sales
• Data Inputs:– Public Records (near-real time with latency and coverage problems).
– FHA Snapshot (monthly dataset of all FHA endorsements; released around mid-month with a one month
lag).
– FHFA loan limits at the county level.
• Assumptions and Construction:– On average, the difference between loan origination and endorsement is one month. ( We have confirmed
this on aggregate by comparing monthly FHA Snapshot to NMRI counts.)
– Price Tiers are set quarterly at the metro level. When there are fewer than 50 FHA loans in a quarter, we
pool all FHA loans at the non-metro state level.
– For the demarcation between medium-high and high tier, we multiply a perspective's county loan limit by
1.25 to account for an 80% LTV, which is the median LTV of loans taken out at the loan limit.
• Result:
22
2018Price Tier
OverallLow Low-Med Med-High High
Mortgage Risk Index 16.0% 14.6% 8.8% 3.2% 11.2%
Market Share* 26% 28% 38% 7% 100%
*For institutionally financed sales. May not round to 100% due to rounding.
National Mortgage Risk Index (NMRI): A Quick Primer
• Overall goal: – Monitor market stability through accurate, real-time tracking of leverage that, if left unchecked,
would result in destructive housing booms/busts.
• Principles behind the NMRI – NMRI is a stress test, similar to a car crash safety rating or hurricane rating for buildings.
– The NMRI’s stress event is the financial crisis from 2007.
• Basics of index construction– The NMRI is a standardized quantitative index for mortgage risk (leverage)
– Places loans in risk buckets and assesses default risk based on the performance of the 2007
vintage loans with similar characteristics
• Advantages of the NMRI– Near-complete census of gov’t-guaranteed loans,
– Accurate, timely, and in-depth coverage of purchase mortgage trends
– NMRI provides significant signals of market trends without the noise of other indices
• What does an increasing or decreasing NMRI mean?– Increasing NMRI = increasing leverage = looser lending– Decreasing NMRI = decreasing leverage = tighter lending
23
Stressed Default Rates, Home Purchase Loans
• Takeaway: Huge spread of default rates across risk buckets
• All 320 risk buckets for home purchase loans are shown at Periodic Table – Purchase
• Analogous tables for cash-out and no-cash-out refi loans are at Periodic Tables –Refinance
• Additional loan risk factors are applied to VA loans and to ARMs, investor loans, second homes, 15 year terms, and 20 year terms
24
Risk Bucket Credit Score CLTV Total DTI Default Rate
Very Low ≥ 770 61-70% ≤ 33% 0.8%
Low 720-769 76-80% 34-38% 4.2%
Medium 690-719 81-85% 39-43% 9.3%
High 660-689 91-95% 44-50% 22.7%
Very High 620-639 > 95% > 50% 45.8%
Note: Default rates represent cumulative defaults through year-end 2012 for Freddie Mac’s
2007 vintage of acquired loans. The loans included in the calculation are all primary owner-
occupied, 30-year fixed-rate, fully amortizing, fully documented, home purchase loans.
Home Sales Methodology• Data Inputs
– Public Records (near-real time with latency and coverage problems).
– HMDA (annual dataset of institutionally financed sales (IFS); covers around 99% of loans; released with
lag).
– FHA Snapshot (monthly dataset of all FHA endorsements; released around mid-month with a one month
lag).
– National Mortgage Risk Index (NMRI) (covers 99% of Agency loans; two months lag).
• Assumptions– Recorder offices process transactions in random order; latency in reporting applies equally across all sales
types.
– FHA loans are properly recorded (stamp on mortgage document).
– On average, the difference between loan origination and endorsement is one month. ( We have confirmed
this on aggregate by comparing monthly FHA Snapshot to NMRI counts.)
– Conventional loans have same seasonal pattern as GSE loans.
• Construction– Aggregation from the county level up.
– Use FHA Snapshot for all FHA sales.
– When HMDA is available: Use HMDA for remaining IFS when available: • Impute cash and other financed sales as a percent of IFS (assume state average for counties with latency
problems);
• Impute seasonal pattern from either public records or NMRI.
– When HMDA is not yet available: Use Public records with adjustments:• Limited to ~ 700 counties that account for ~80% of sales (remove counties with insufficient FHA counts or
breaks in series);
• Gross up all sales in that county by the ratio of FHA Public Records loans to FHA Snapshot loans;
• Assume same rate of change for ~2400 counties with ~20% of sales -> still working on improving this
assumption.
• As a robustness check of this, we compare state VA and RHS totals to the NMRI and adjust totals.
25
House Price Appreciation (HPA) Index: A Quick Primer
• Overall goal: – Monitor market stability through accurate, real-time tracking of house prices.
• Basics of index construction– Most widely known HPA Indices are repeat sales (i.e. Case Shiller or FHFA) or hedonic (Zillow)
indices.– AEI’s HPA is a “quasi” repeat sales index with a hedonic element.– Index measures HPA by constructing an artificial sales pair consisting of one actual sale and one
“artificial” sale as measured by the property’s AVM.– The AVM (Automated Valuation Model) approximates a property’s sale price at a given point in
time. The AVM used for AEI’s HPA Index is unbiased and accurate.
• Advantages of AEI’s HPA Index:– Combines the best of repeat and hedonic models.
– Unlike a true repeat sales index, which is limited to repeat sales and may therefore be biased,
AEI’s index includes the entire universe of sales.
– Unlike a true hedonic index, which incorporates every property (even unsold ones), it reduces the
amount of errors since at least one sale of the transaction pair actually occurred.
– Allows for an index construction by price tier and fine geographic levels (down to census tract).
• Data for the HPA index– National Public Records data and AVM for Dec-2017 come from First American via DataTree.com.– Uses virtually all institutionally financed sales back to January 2012.– Data are weighted at the county level to make them representative.– HPAs for the medium-high and high price tiers are spliced around the time of loan limit changes.
26
New Construction Methodology
• Data Inputs– Public Records (Deed & Assessor files)
– Zillow API and/or Listings data
• Identification of New Constructions– Year Built in Assessor data
• We also just received Effective Year Built so going to analyze in future
– If Year Built is missing:
• Seller name (we have assembled a list of over 400 builders with their subsidiaries and key words to identify smaller builders.) If a
seller is a builder and the Year Built is missing, then it is most likely a new construction that has not yet been assessed.
• Ping Zillow API for Year Built and Use Code. Data not perfect, but even some information helps us determine status.
• Sellers with multiple sales that are not individuals/gov’t/lender/other corporation are most likely builders. (Relatively small number.)
– Count only first sale of home as a new construction.
– Still working on identifying owner-built homes without a long lag.
• Verification– Random sampling and checking of new constructions and existing homes using Zillow data, Google street
view/satellite images.
– Find very few false positives and false negatives. (Random sample found <4% were false positive, <3% of remaining
88% found to be new construction.)
• Final dataset allows us to:– Monitor new constructions at the property level with minimal lags,
– Accurately estimate new home sales at very fine geographic levels when combined with Home Sales numbers,
– Estimate additions to the existing housing stock when combined with Assessor data,
– Estimate sales by builder and track builder,
– Combine new construction numbers with Months’ Supply and house price appreciation,
– Much more.
27
List of Abbreviations (cont’d)
Term DescriptionGSE A Government-Sponsored Enterprise (GSE) is an entity created by Congress that operates under a
government-defined mission and charter. There are two housing-related GSEs: Freddie Mac and Fannie
Mae. They purchase mortgages on the secondary market and subsequently pool them into mortgage-
backed securities (MBS), which are purchased by government and private investors.
Fannie Mae The Federal National Mortgage Association (FNMA), known as Fannie Mae, was founded in 1938 as
part of the New Deal legislation.
Freddie Mac The Federal Home Loan Mortgage Corporation (FHLMC), known as Freddie Mac, was created in
1970 to complement Fannie Mae.
Ginnie Mae The Government National Mortgage Association (Ginnie Mae) is a federal government corporation
that aims to promote homeownership for low- and moderate-income families. It ensures the timely
payment of principal and interest on mortgage-backed securities formed from mortgages that are
guaranteed or insured by FHA, VA, RHS, or smaller programs for Native Americans. Ginnie Mae was
created in 1968. Prior to 1968 its role was performed by Fannie Mae.
FHA The Federal Housing Administration (FHA), founded in 1934, is a federal agency that today provides
mortgage insurance for residential loans made to high-risk borrowers. The borrower pays an upfront
mortgage insurance premium as well as monthly insurance premiums for the service. In return, FHA
covers 100% of the lender’s loss in case of the borrower’s default.
RHS The Rural Housing Service (RHS) is a program within the U.S. Department of Agriculture that
guarantees mortgages in rural areas. The borrower pays an upfront annual fee for the service. In return,
RHS covers 100% of lender’s loss in case of the borrower’s default.
VA The Department of Veterans Affairs (VA) guarantees mortgages to eligible veterans and generally
pays 25% of lender’s loss in case of the borrower’s default. The borrower pays an upfront annual fee for
the service.
HUD FHA has been overseen by the Department of Housing and Urban Development (HUD) since its
creation in 1965.
28
List of Abbreviations
Term DescriptionMRI The Mortgage Risk Index (MRI) measures how the loans originated in a given month would perform if
subjected to the same stress as loans originated in 2007, which experienced the highest default rates as
a result of the Great Recession.
NMRI The National Mortgage Risk Index (NMRI) currently covers home purchase and refinance loans
(except for VA refinances) that have been (1) acquired and securitized by Fannie Mae or Freddie Mac or
(2) insured or guaranteed by the Federal Housing Administration (FHA), the Department of Veterans
Affairs (VA), or the Rural Housing Service (RHS).
SMRI The State-level Mortgage Risk Index (SMRI) measures mortgage risk on a state level. It employs
exactly the same stress-test methodology as the national index.
FBMSI The First-time Buyer Mortgage Share Index (FBMSI) equals the number of loans made to first-time
buyers divided by the number of all home purchase loans excluding those made to investors and second
home buyers for any given month (see first-time buyer (FTB) definition below). The agency FBMSI
covers government-guaranteed loans, while the combined FBMSI covers both government-guaranteed
and private-sector loans. The agency loans are from the same database used for the NMRI, while the
private-sector component of the combined FBMSI come from AEI’s National Housing Market Index
(NHMI) and assumptions believed to be reasonable.
FBMRI The First-time Buyer Mortgage Risk Index (FBMRI) is calculated using the same methodology as for
the NMRI. The only difference is that the set of included loans is restricted to first-time buyers.
FTB AEI uses the federal government’s definition of a first-time homebuyer (FTB). A FTB is an individual
borrower who (1) is purchasing the mortgaged property, (2) will reside in the mortgaged property as a
primary residence, and (3) had no ownership interest (sole or joint) in a residential property during the
three-year period preceding the date of the purchase of the mortgaged property. Investment properties,
second homes, and refinance transactions are not eligible to be considered first-time homebuyer
transactions. Other organizations such as the National Association of Realtors (NAR) use a different
definition of FTB based on self-identification.
RB Repeat Buyers (RB) are all home buyers that are not first-time buyers.
29
List of Abbreviations (cont’d)
Term DescriptionFICO® The FICO Credit Score is a statistical credit evaluation score developed by Fair, Isaac and Co. The
FICO score attempts to measure a borrower’s risk of default through his or her personal financial history.
FICO scores range from a high default-risk score of 300 to a low default-risk score of 850. The term
“credit score” is used to connote a generic score.
LTV / CLTV The Loan-to-Value Ratio (LTV) is the ratio of the 1st lien loan amount to the property’s value. Since the
down payment on a purchase transaction is the property’s value minus the loan amount, the LTV is
inversely related to the down payment. The Combined Loan-to-Value (CLTV) is the ratio of all loan
amounts at 1st lien origination to the property’s value. Both ratios are a measure of a borrower’s skin in
the game.
DTI The total Debt-to-Income Ratio (DTI) gauges the ability of a borrower to repay a mortgage by
measuring the amount of income consumed for repayment of all outstanding debts of the borrower.
ARM An Adjustable-Rate Mortgage (ARM) is a mortgage whose interest rate varies over the lifetime of the
loan based on market conditions. ARMs have on average a higher default risk than FRMs.
FRM A Fixed Rate Mortgage (FRM) maintains the interest rate at origination throughout the lifetime of the
loan.
MSA A Metropolitan Statistical Area (MSA) is a geographical region with a population of at least 50,000
inhabitants at its core and close economic ties throughout the region.
PCE price index The Personal Consumption Expenditure (PCE) price index measures the prices of goods and
services purchased by consumers in the U.S. economy. It is published monthly by the Bureau of
Economic Analysis in the Department of Commerce. The PCE price index is the measure of inflation
targeted by the Federal Reserve.
SLOOS The Senior Loan Officer Opinion Survey on Bank Lending Practices (SLOOS) is a survey of lending
conditions conducted quarterly by the Federal Reserve among roughly eighty large domestic banks and
twenty-five U.S. branches and agencies of foreign banks.
30
List of Abbreviations (cont’d)
Term DescriptionQM/QRM The Qualified Mortgage (QM) and the Qualified Residential Mortgage (QRM) are mortgage terms
created under the Dodd-Frank Act. A mortgage that meets the QM requirements provides legal
protection for lenders against a claim that the loan was made without due consideration of the
borrower’s ability to repay. The QRM designation relates to the securitization of mortgages. If the loans
in a mortgage-backed security are QRMs, the securitizing agent is not required to retain any risk position
in the security. Although the initial proposed QRM definition was relatively strict, the final definition was
watered down to be equivalent to the looser QM definition. The five guarantee agencies (Fannie Mae,
Freddie Mac, FHA, VA, and RHS are exempt from substantial portions of the QM rules and entirely from
the QRM rules. (For Fannie and Freddie, this exemption applies only while they are in conservatorship).
MIP The Mortgage Insurance Premium (MIP) is a payment to compensate for the risk of default on the
mortgage. As noted above, FHA mortgages carry both upfront and monthly MIP payments. Fannie Mae
and Freddie Mac generally require mortgage insurance for loans they guarantee with LTVs above 80%;
borrowers with these GSE-guaranteed loans may make monthly MIP payments depending on the
premium plan.
TRID The TILA-RESPA Integrated Disclosure (TRID) rule – commonly also known as Know Before You
Owe – requires lenders to summarize and more prominently display the loan terms on the mortgage
form. It also institutes a three-day waiting period before closing to allow borrowers time to review the
contract. The form change is currently suppressing sales volume as it is delaying loan closings by
creating additional burdens on lenders. TRID was mandated by the Consumer Financial Protection
Bureau (CFPB) and applies to mortgage applications filed on or after October 3, 2015.
31
AppendixAdditional slides and those not included every month:
– National Mortgage Risk Index: Loan Totals
– Background: Financial Crisis and AEI’s Response
– Principles of Housing Finance over 125 years (1850-1975)
– Pinto’s Principles of Housing Finance
– Definition of Low-Risk Loans
– Current State of the Housing Market
– Recent Steps by the GSEs, the FHA, and Regulators Add Fresh Fuel to the Long Running and Accelerating House Price Boom
– Agency Origination Shares, Purchase Loans
– Changes in the >95% CLTV Purchase Loan Market
– Agency Origination Shares, Cash-Out Loans
– Agency Origination Shares, No Cash-Out Loans
– Agency Total, Refi, and Purchase Loan Counts
– DTI Distributions, Agency Primary Purchase Loans
– Prime, Subprime, and Nearprime, Purchase Loan Counts
– GSEs: Large Lender Market Share and Relative Risk Share, Purchase Loans
– FHA: Large Issuer Lender Type Market Share and Relative Risk Share, Purchase Loans
– Combined, Purchase, and Refi NMRIs
– No-Cash-Out Refi and Cash-Out Refi NMRIs
– FHA’s Pro-Cyclical Policies Continue to Fuel the Boom
– Which Risk Factors Have Driven Up the Purchase NMRI?
– FHA’s NMRI for Home Purchase and Refinance Loans
– What explains FHA’s riskiness?
– How wide is the FHA credit box?
– Risk Overlap between FHA and the GSEs
– FHA NMRI by Risk Decile, Home Purchase Loans
– FHA Median Downpayment and Sales Price
– FHA’s Should Begin Reining in Its Pro-Cyclical Policies 32
Appendix (cont’d)
33
– BCFP’s Pro-Cyclical QM Patch Continues to Fuel the Boom
– Pro-Cyclical Parallels to the Last Boom
– Purchase Loans with Total DTI Greater than 43%
– It Is Time for the BCFP to Announce that the Temporary GSE QM Patch Will Be Allowed to Sunset in January 2021
– The Role of Leverage
– Ratio of Sales Price for First-time to Repeat Buyers
– Median Sale Price by Market Segment, FTB Purchase Loans
– Measure Market Behavior in Four Leverage Based Price Tiers
– House Price Trends Impacted by Leverage
– Constant-quality Prices by Guarantor Type: Low Price Tier
– High risk home purchase lending is fueling home price appreciation
– Scatterplots: Introduction
– Strong Positive Correlation Between Mortgage Risk & Home Price Appreciation
– Strong Positive Correlation Between Mortgage Risk & Tract Income
– Measured Steps Now Would Moderate Unsustainable Home Price Increases, Not Lead to Home Price Declines
– DTI Distributions, GSE & FHA Purchase Loans
– Share of Fannie Purchase Loans by DTI Bucket
– RHS Reduced Borrower DTIs from 2013 to 2018, while the FHA Kept Increasing DTIs
– The FHA’s and the GSEs’ Rising DTIs Have Been Pro-Cyclically Fueling the House Price Boom
– Origination Shares Issuer Lender Type, FHA and RHS Purchase Loans
– MRIs by Issuer Lender Type, FHA and RHS Purchase Loans
– Origination Shares and MRIs by Seller Lender Type, GSE Refinance Loans
– Origination Shares and MRIs by Issuer Lender Type, FHA Refinance Loans
– State NMRI and FHA Share, Purchase Loans
– State NMRI Change, Purchase Loans
Appendix (cont’d)
34
– Pricing Changes, Home Purchase Loans
– Stressed Default Rates by Loan Type
– Median Downpayments
– Volume Growth in Counts and Dollars, Purchase Loans
– A closer look at RHS’ October 2016 MIP cut
– No-Cash Out Refi Demand and 30-yr Mortgage Rate
– Low-Risk Origination Shares, Purchase Loans
– Calibrating Mortgage Safety
– Credit Conditions: 1990 to 2013-14
– Role of Income Leverage During Housing Boom
– Fed Tightening and Efforts to Maintain Buying Power
– Appraisals Should Be the Guard Rail Against Speculative Booms
– Cross-subsidies Return to the GSEs
– Change in Agency Purchase Loan Volume
– GSEs: Ratio of NMRI for Loans with Total CLTV > 95% to Loans with Lower Total CLTVs
– FICO® Score Distribution
– Median Credit Score on Primary Purchase Loans
– Aggregate Default Risk Surge for Home Purchase Loans Is Over Five Years Old
– The Effect of April 2016 PMI Price Change
– Credit Score Distribution & MRIs, Purchase Loans
– Purchase Loans with Down Payment of 5% or Less
– Composite Origination Shares and MRIs by Channel, Purchase Loans
– Large Bank Origination Shares and MRIs by Channel, Purchase Loans
– Fed’s Senior Loan Officer Survey is Badly Flawed
– Urban Myth: Tight Credit Keeping“Creditworthy” Borrowers Out of Market
– FHA Perpetuates This Myth
Appendix (cont’d)
35
– FHA Is All about Moral Hazard
– While FHA’s Capital Reached Required 2% Statutory Level for 1st Time since 2008, It Is Insufficient
– Share of States with Increase in SMRI for Purchase Loans from Year-Earlier Period
– CBSA NHMI: Investor Type for Home Purchase Loans
– Riverside/San Bernardino: A Case in Point
– House Price Volatility, 51 Largest Metro Areas
– Median Values of Risk Factors by Loan Type
– Risk Shares for Home Purchase Loans
– DTI Distributions and MRIs, Primary Purchase Loans
– Cash-Out Share and Home Equity
– Agency Cash-Out Share and Defaults
– Nonbank Origination Shares and MRIs by Channel, Purchase Loans
– Greater House Price Volatility at the Lower End
– Tale of two markets: Low end and entry level and high end repeat buyers (high price tier)
– Tale of two markets: Low end and entry level and high end repeat buyers (low price tier)
– Comparison: Home Sale Transactions (New and Existing)
– Home Sales, by Metro Market Size
– Mortgage Risk Indices by Lender Type, Purchase Loans
– Jumbo portfolio-GSE spreads (in bps)
– Homeowners Can’t Count on House Price Gains to Build Wealth
– Evaluating the GSEs 2017 Business
– Evaluating the GSEs 2017 Business (cont.)
– Evaluating the GSEs 2017 Business (cont.)
– Update: John Burns Intrinsic Home Values
– GSEs: Large Lender Market Share and Relative Risk Share, Refinance Loans
– FHA: Large Issuer Lender Type Market Share and Relative Risk Share, Refinance Loans
– Leverage Fueled Housing Demand Continues to Climb
– Agency Origination Shares by Risk, Purchase Loans
Appendix (cont’d)
36
– What Does this Mean for the Broader Market?
– Borrowing at the Conforming Loan Limit, GSE Purchase Loans
– The CFPB’s Qualified Mortgage Policy and GSE QM Patch Allowed for Credit Easing While Supply Is Constrained, a Direct and Continuing Cause of the Current House Price Boom
– Number of Investors Flipping Houses Creeping Up
– Average House Price Change by Zip (%, annual avg.)
– Raising the Conventional Loan Limit – A Prediction
– Raising the Conventional Loan Limit – A Good Idea?
– Cash-Outs and the Economy
– Cash-Out Share and Expected Defaults
– FTB slides begin
– Punchbowl 1: Mortgage Rate Changes Applicable to FTBs and RBs
– Punchbowl 2: Eased Underwriting StandardsOnly Available to Agency First-time Buyers
– Agency Purchase Loan Demand Remains Strong
– Market Segmentation: Median Sales Price for First-time and Repeat Buyers
– Constant Quality Prices Outlook for the Bifurcated Market:Slowing Price Appreciation for at the Higher End, Continued Robust Appreciation for at the Lower End
– Outlook for Bifurcation of Market – Quality Changes
– Outlook for a Bifurcated Market –Transaction Prices by State & Changes in Transaction Volume
– Outlook for a Bifurcated Market – Transaction Prices by State & Changes in Median Transaction Prices
– While FHA’s Forward Program Capital Is at 3.9%, in Excess of Statutory Minimum of 2%, It Should be 7%
– FHA Cash Out Count and MRI
– Average Credit Score and DTI: FHA Purchase Loans
– FTB Purchase Loan NMRI: Credit Easing Continues
Appendix (cont’d)
37
– Which Risk Factors Have Driven Up the FTB NMRI?
– Share of GSE FTB Purchase Loans w. DTIs of 46-50%
– FTB Purchase Loans, by Level of Downpayment
– Agency First-time Buyer Purchase Loan Share
– Government Housing Policy Creates an Economics Free Zone
– Definition of Low-Risk / Prime Loans
– Agency First-Time Buyer Loan Count
– Agency Origination Shares, FTB Purchase Loans
– Origination Shares by Credit Score Bin, First-time Buyer Purchase Loans
– Agency Origination Shares, FTB Purchase Loans by Market Segment
– Originations by Market Segment, FTB Purchase Loans
– Combined First-Time Buyer Mortgage Share Index
– The NAR’s first time buyer series is fatally flawed. After removing seasonality, most of what remains is noise
– Changes in the >95% CLTV Purchase Loan Market
– Share of States with Rise in First-time Buyer Loan Volume and Share from Year-Earlier Period
– Profiles of GSE and FHA First-time Buyers with >95% CLTV
– Characteristics of Mortgages Taken Out by First-Time and Repeat Homebuyers
– Rising Prices Have Disparate Effects on Buyers
– Agency-Specific First-Time Buyer MortgageShare Indices
– DTI Distributions, Agency FTB Purchase Loans
– The Effect of FHA Mortgage Insurance Premium Cut
– Income and Debt Growth by Income Group: FHA Purchase Loans
– House Price Appreciation (HPA) by Price Tier
– House Price Appreciation (HPA) by Price Tier: 73 Metros
– Supply-Demand Imbalance in the Market Is Driving Prices Up
– Affordability Worsens in a Seller’s Market
Appendix (cont’d)
38
– Fannie vs. Freddie Risk Index, GSE Purchase Loans
– History Repeats Itself: the “Quiet” Battle for Subprime (High Risk >95 CLTV Purchase Loans) among Fannie, Freddie & FHA
– Leverage Fueled Housing Demand Pauses Due to Higher Rates
– Agency First-time and Repeat Buyer Mortgage Risk Indices
– Origination Shares and MRIs by Seller Lender Type, GSE Purchase Loans
– Origination Shares and MRIs by Issuer Lender Type, FHA Purchase Loans
– Mortgage Risk Indices by Lender Type, Refi Loans
– FHA’s March 2018 Action to Address Excessive Loan Risk Is a Positive Step, but Too Early to Tell How Significant
– Credit Easing = Punchbowl Spiking Continues, Led by FHA
– Supply-Demand Imbalance in the Market Is Driving Prices Up
– Cash Out Refi Volume by Agency
– Compositional Change of Cash-Out Refis
– Unforgiving Home Price Cycles: Booms Fueled by Increasing Leverage in a Seller’s Market, Followed by Mean Reversion
– Housing Volatility Index
– Supply-Demand Imbalance Is Greatest in the Low Price Tier
– Comparing the Supply-Demand Imbalance: 100 Largest Metros
– Months’ Supply: A Tale of Two Markets
– Annualized Home Sales (New vs Existing)
– Quarterly Home Sales (New and Existing): by Type
– Origination Shares Based on Purchase Loan Counts
– Agency Refi and Purchase Loan Counts
– Agency First-Time Buyer Mortgage Share
– Ratio of Sales Price for First-time to Repeat Buyers
– Median Sale Price by Risk Segment*, FTB Purchase Loans
National Mortgage Risk Index:
Loan Totals
• The November 2018 NMRI covers over 35.9 million agency loans dating back to Sept. 2012. These data are used to construct the NMRI, First-Time Homebuyer Indices, and the National Housing Market Indexes (NHMI).
• This total consists of nearly 18.2 million agency purchase loans and over 17.7 million agency refinance loans
• NMRI and other risk indices published for:
– Purchase loans, with separate indices for first-time and repeat buyers
– Refinance loans, with separate indices for no-cash-out and cash-out refis
– Composite of purchase and refinance loans
– Purchase loan NMRI is the primary measure for monitoring mortgage risk and the impact of housing policy, particularly with respect to first-time buyers
– Refinance loan NMRI contributes to overall assessment of changes in leverage
39
Background: Financial Crisis and AEI’s Response
• Financial crisis largely stemmed from a failure to understand buildup of housing risk:
– Mortgage risk
– House-price (collateral) risk
– Capital adequacy
• AEI’s Housing Center (AEI.org/housing) addresses this problem by undertaking evidence-based research that expands the body of knowledge concerning housing markets and finance:
– Provides objective and transparent mortgage risk measures• Risk indices published monthly
– Provides objective and transparent housing market indicators• Market indicators published quarterly
– Provides objective and transparent house price appreciation measures
40
Principles of Housing Finance over 125 years (1850-1975)•At all times, but especially in the last few years, people have dreamt of universalizing wealth by universalizing credit.… Now, in no country is it possible to transfer from one hand to another more products than there are(1850)1
•Since value depends on location, & location on convenience, & convenience on nearness, the intermediate steps may be eliminated & say that value depends on nearness. (1903)2
•If a new utility does not arise, [sales] prices may advance & recede, while intrinsic values do not change. If a new utility arises, both [sales] prices & intrinsic values will alter their levels. (1903)2
•[s]peculative elements cannot be considered as enhancing the security of residential loans [rather they] enhance the risk of loss to mortgagees [if] permit[ed] to creep into valuations….(1938)3
•Because situations of scarcity [seller's market] or over-supply [buyer's market] do not last indefinitely they cannot be considered as phenomena the affect valuations for long-term use…. & not truly indicative of value for mortgage insurance purposes. (1947)4
•The sequence of [market cycle] events is fairly predictable, though the period of the phases of the cycle & the amplitude of the variations are not subject to dependable forecasting. (1949)5
•[I]nflationary construction costs, home purchase prices, & land prices not only loan disproportionate financial burdens upon the owners at time of acquisition but also form the bloated base upon which the major costs of occupancy [including property taxes] are determined for the entire term of ownership. (1949)5
•The essential nature of housing demand is changeability; the nature of housing supply is rigidity. (1949)4
•[I]n a seller's market, when choice is restricted & the seller virtually dictates sales terms, more liberal credit is likely to be [capitalized] in price with probably a reduction in housing standards. (1951)6
•[Transitioning] from buyer's to seller's market, maximum terms become so commonly used they tend to be considered the minimum. (1951)6
•The parallel between the increases in the “costs” of new housing units & increases in the amount & percentage of needed funds that could be obtained by lengthening their terms & [reducing] downpayments raises the radical question of whether the disbursements made to assist purchasers & (renters) have not benefited others more than those whom they were intended to relieve. The largest groups to whom it is sometimes suggested some of the benefits may have flowed are the builders, building labor, the suppliers of building materials, & real estate brokers & speculators. (1975)71 Frederic Bastiat, What Is Seen and What Is Not Seen, 1850, 2 Hurd, The Principles of City Land Values, 1903, 3 FHA 1938 Underwriting Manual (main authors: Ernest Fisher and Frederick Babcock)4 FHA 1947 Underwriting Manual5 Ratcliffe, Urban Land Economics, 19496 Fisher, Financing Home Ownership, NBER, 19517 Fisher, Housing Markets and Congressional Goals, 1975
41
Pinto’s Principles of Housing Finance
1. Corollary to Fisher’s capitalization rule: capitalization is added to land price
2. Uncertainty Principle: Can’t simultaneously set an asset’s credit risk & risk weight
– A low risk designation and corresponding low capital weight (greater leverage) unleashes demand pressures causing it to no longer be low risk (think GSEs, private MBS, Greek sovereign debt)
3. Dual Underestimation Principle: Never underestimate the government’s willingness & ability to (i) add leverage to stimulate the market & (ii) ignore its impact on raising home prices and default risk under stress
– Housing debt & default risk have increased with over 60 years of housing policies focused on increasing leverage
4. Law of the Marginal Buyer: Home prices will keep rising so long as the marginal buyer, who sets the price for all, has access to higher leverage (see #3)
5. Corollary: Historically the government has endeavored to add leverage in both buyer’s & seller’s markets; but the latter has potential for dangerous buildup of risk (see #1)
– Result is an economics free zone promoting demand, while supply is restricted by regulation
• FHA neither prices nor underwrites for risk
• Government policies increase leverage regardless of rates going up or down
• Low capital entities (FHA and GSEs) compete with each other over loosening credit
• Affordable housing goals and duty to serve policies promote risky lending 42
Definition of Low-Risk Loans
• We define low-risk loans as those with a stressed default rate of less than 6%. Why?
• Low-risk definition calibrated from two sources
– Original QRM proposal to implement Dodd-Frank
– FHA underwriting standards over 1935-55
– Both yield an average stressed default rate of ≈ 3%
• This is consistent with a maximum stressed default rate of ≈ 6% on individual loans, assuming a uniform distribution starting near 0%
• Hence the use of 6% as the highest stressed default rate for a low-risk loan
43
• The current house price boom is about 6 years old and rate of house priceincreases is accelerating
• “Home Values Climbing at Fastest Rate in 12 Years….The median U.S. home value rose 8.7 percent to $215,600 in April, the fastest year-over-year climb since June 2006” Zillow, 5.24.17
• “Start of year sees strongest home price growth since 2005. … About 60% of all U.S. metros saw an acceleration in the rate of price increases through February this year.” (Housing Wire, 5.7, 2018)
• “Housing confidence hits record high as home prices skyrocket. Consumer confidence in housing jumped to its highest level on record in April, according to Fannie Mae. Those who think home prices will move even higher rose the most, and those who think now is a good time to sell came in second.” (CNBC, 5.7.18)
• “Mortgage lenders are making it easier for you to buy a house. But are they repeating last decade's mistakes? Dana Wade, the acting Federal Housing Administration commissioner, minced few words in testimony last month before a U.S. House of Representatives committee. The FHA, the federal housing agency that insures mortgages made to first-time and lower-income buyers, has seen “certain trends and indicators of potential defaults.” Philadelphia Inquirer 5.4.18
• A house price boom is when prices rise faster than fundamentals
• The boom is driven by too much money chasing too few properties– When the market is supply constricted (a seller's market), credit easing is likely to be capitalized in price.
– FHA, Fannie, Freddie, and the VA are all pro-cyclically fueling the boom
• The length and acceleration of the boom adds urgency to shrink the GSEs and FHAby administrative action
44
Current State of the Housing Market
Recent Steps by the GSEs, the FHA, and Regulators Add Fresh Fuel to the Long Running and Accelerating House Price Boom
• “Freddie Mac takes aim at FHA with widespread expansion of 3% down mortgages…. But now, Freddie Mac is about to supercharge its 3% down program and launch a widespread expansion of the offering.” (Housing Wire, 4.26.18)
• “Credit scores may jump starting this month…. Because of improved standards [from regulators] for utilizing new and existing public records, the three major credit reporting companies are now excluding all tax liens from credit reports. That means some scores will head higher, for some by as much as 30 points.” (CNBC, 4.12.18)
• “Manufactured housing giant endorses HUD's call for regulatory relief…. But the FHA has suffered major losses from insuring manufactured loans in the past and is unlikely to increase its role in this sector.” (National Mortgage News, 4.3.18)
• “Will The Gig Economy Change Mortgage Lending?...[r]ather than two years of iron-clad documentation, [GSEs] now say as little as 12 months of self-employment are enough, as long as the applicant’s previous employment is in the same field and his or her income remains steady.” (Mortgage Orb, 7.26.17)
• “If the lender obtains documentation to evidence the actual monthly payment is $0, the lender may qualify the borrower with the $0 payment as long as the $0 payment is associated with an income-driven repayment plan.” (Fannie Mae Selling Guide, 7.25.17)
• “Fannie Mae will ease financial standards for mortgage applicants next month… Fannie will be raising its DTI ceiling from the current 45 percent to 50 percent as of July 29.” (Washington Post, 6.6.17)
45
Agency Origination Shares, Purchase Loans
Source: AEI Housing Center, www.AEI.org/housing.
Major market shifts are often related to pricing changes. The largest effect was from FHA’s mortgage insurance premium (MIP) cut in January 2015, which boosted FHA’s market share from 23% to 30%. Recently, FHA’s share has declined, returning FHA back to its pre-MIP cut level. As Freddie’s MBS execution price has improved, it has
recently picked up share.
46
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
Sep-12 Feb-13 Jul-13 Dec-13 May-14 Oct-14 Mar-15 Aug-15 Jan-16 Jun-16 Nov-16 Apr-17 Sep-17 Feb-18 Jul-18
Freddie
VA
RHS
Fannie
FHA
Changes in the >95% CLTV Purchase Loan Market
47
While FHA continues to dominate this market segment with a 73% share, this is down from a 92% share in Oct. 2015. Fannie continues its dominance over Freddie, coming in at a 20% share, up from 7% in Oct. 2015, compared to Freddie’s 7%, up from 1% in Oct.
2015. This is contributing to Fannie’s Risk Index sprinting ahead of Freddie’s.
Note: Excludes loans made by VA and RHS.
Source: AEI Housing Center, www.AEI.org/housing.
0
25,000
50,000
75,000
100,000
0
25,000
50,000
75,000
100,000
Sep-12 Apr-13 Nov-13 Jun-14 Jan-15 Aug-15 Mar-16 Oct-16 May-17 Dec-17 Jul-18
Number of loans with CLTV > 95%
FHA
Freddie
Fannie
Total (FHA, Fannie, Freddie)
Text not updated
Agency Origination Shares, Cash-Out Loans
Source: AEI Housing Center, www.AEI.org/housing.
Market share for cash-out refis has shifted from the GSEs to the FHA and VA. FHA and VA accounted for less than 10% of market share in 2012. In October 2018, they
accounted for 33%, with FHA’s share surging recently. This increase has powered the increase in the riskiness in the cash-out index.
48
0%
10%
20%
30%
40%
50%
60%
70%
0%
10%
20%
30%
40%
50%
60%
70%
Sep-12 Mar-13 Sep-13 Mar-14 Sep-14 Mar-15 Sep-15 Mar-16 Sep-16 Mar-17 Sep-17 Mar-18 Sep-18
Fannie
Freddie
FHA
VA
RHS
Nov 2018 Cash-Out NMRIComposite 15.8%Fannie 11.4%Freddie 10.3%FHA 27.1%VA 23.4%
Agency Origination Shares, No Cash-Out Loans
Source: AEI Housing Center, www.AEI.org/housing.
VA and FHA were both losing market share as early as 2018. However, the trend between the two agencies diverges around May 2018, right around the time the VA
was subjected to a new statute designed to reign in predatory no cash-out refi lending. The VA share is now near its series’ low dating back to October 2013.
49
0%
10%
20%
30%
40%
50%
60%
70%
0%
10%
20%
30%
40%
50%
60%
70%
Sep-12 Mar-13 Sep-13 Mar-14 Sep-14 Mar-15 Sep-15 Mar-16 Sep-16 Mar-17 Sep-17 Mar-18 Sep-18
Fannie
Freddie
FHA
VARHS
Nov 2018 No Cash-Out NMRIComposite 12.1%Fannie 10.4%Freddie 8.0%FHA 26.8%VA 22.8%
Agency Total, Refi, and Purchase Loan Counts
All agency volume was down 29% from a year ago. Refi volume has slowed since the end of 2016. As interest rates have moved sharply higher since November 2016, refi volume,
and especially no-cash out refi volume, has contracted.
50Source: AEI Housing Center, www.AEI.org/housing.
0
100,000
200,000
300,000
400,000
500,000
600,000
700,000
800,000
0
100,000
200,000
300,000
400,000
500,000
600,000
700,000
800,000
Sep-12 May-13 Jan-14 Sep-14 May-15 Jan-16 Sep-16 May-17 Jan-18 Sep-18
Refis
Purchase loans
No-Cash-Out Refis
Cash-Out Refis
TotalCash-out share of refis
Nov-12: 19%Nov-13: 30% Nov-14: 33%Nov-15: 40%Nov-16: 37%Nov-17: 55%Nov-18: 71%
Red markers show November count in each year.
DTI Distributions, Agency Primary Purchase Loans*
*Data pertain to all agency purchase loans for primary owner-occupied properties. Source: AEI Housing Center, www.AEI.org/housing.
DTIs have been shifting higher as the rise in house prices has been outpacing income gains. The share of DTIs below 34% has declined sharply, offset by a much
greater share of DTIs above 40%. While bullish for home prices in the near term, this presents long-term sustainability problems for both homeowners and the FHA.
51
California shows how the shift could intensify as affordability worsens.
0%
1%
2%
3%
4%
5%
6%
<20 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65
DTI Distribution for Entire U.S.
Feb. 2013
GSE November 2018
Ginnie November 2018
DTI Feb. 2013 Nov. 2018Difference
<34% 39% 25% -13 ppts
34-40% 26% 22% -3 ppts
>40% 36% 53% 17 ppts
Shift in DTI Distribution
0%
2%
4%
6%
8%
<20 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65
DTI Distribution, November 2018
U.S. excluding CA
Ginnie CA
GSE California
DTI CA U.S. ex. CA Diff (CA - U.S. ex. CA)
<34% 14% 27% -13 ppts
34-40% 19% 23% -4 ppts
>40% 66% 49% 16 ppts
0
20,000
40,000
60,000
80,000
100,000
120,000
140,000
160,000
0
20,000
40,000
60,000
80,000
100,000
120,000
140,000
160,000
Sep-12 May-13 Jan-14 Sep-14 May-15 Jan-16 Sep-16 May-17 Jan-18 Sep-18
Near-prime
Prime
Prime, Subprime, and Nearprime, Purchase Loan Counts
Source: AEI Housing Center, www.AEI.org/housing.Note: Prime loans are defined as having a stressed default rate less than 6%; near-prime loans are between 6 to 12%; subprime loans are greater than 12%.
While growth in purchase lending volume has paused, it has not paused equally across the risk spectrum. Historically, most of the growth in volume has come from the near-
prime and especially the subprime segment.
52
Subprime
Large banks Large nonbanks
53
NMRI
Higher GSE risk share(relative to market share)
Lower GSE risk share(relative to market share)
Larger circle represents larger market share. Lenders shown represent the 8 largest banks and 15 largest nonbanks by origination share in 2017.
30% 20% 10% 5% 1%
Wells Fargo
Citizens Bank
USAA
JP Morgan Chase
Bank of America
BB&T
Flagstar Bank
U.S. Bank
Pennymac
Caliber Home
Nationstar
Amerihome
New Penn Financial
Lakeview
Franklin American
Guild
Quicken Loans
LoanDepot
United Shore
Fairway Independent
Money Source
Ditech
GSEs: Large Lender Market Share and Relative Risk Share, Purchase Loans
25+% -25+%15 to 25% -15 to -25%-5 to -15%5 to 15% 5 to -5%
2013 2014 2015 2016 20172018/Q1-Q3
2018/Oct.
5.4% 6.0% 6.2% 6.5% 6.8% 7.3% 7.4%
2013 2014 2015 2016 20172018/Q1-Q3
2018/Oct.
5.4% 6.0% 6.2% 6.5% 6.8% 7.3% 7.4%
Freedom Mortgage
Not updated
Large banks Large nonbanks
54
NMRI
Higher FHA risk share(relative to market share)
Lower FHA risk share(relative to market share)
FHA: Large Issuer Lender Type Market Share and Relative Risk Share, Purchase Loans
Larger circle represents larger market share. Lenders shown represent the 8 largest banks and 15 largest nonbanks by origination share in 2017.
30% 20% 10% 5% 1%
US Bank
JPMorgan Chase
Wells Fargo
Citizens
Flagstar
BB&T
Bank of America
Nationstar
Amerihome
Freedom Mortgage
Ditech
Money Source
United Shore
Quicken Loans
PennyMac
Guild
LoanDepot
Caliber
New Penn
Fairway Independent
Lakeview
25+% -25+%15 to 25% -15 to -25%-5 to -15%5 to 15% 5 to -5%
2013 2014 2015 2016 20172018/Q1-Q3
2018/Oct.
22.0% 24.1% 23.8% 24.6% 26.0% 27.8% 28.2%
2013 2014 2015 2016 20172018/Q1-Q3
2018/Oct.
22.0% 24.1% 23.8% 24.6% 26.0% 27.8% 28.2%
Not updated
Combined, Purchase, and Refi NMRIs
The Combined Purchase and Refi NMRI set a series’ high in November. There has been a sharp trend reversal on refis, which tend to follow feast-and-famine cycles depending
on the mortgage rate. The Refi series is pulling away steeply from the Purchase one after having converged at elevated levels.
55Source: AEI Housing Center, www.AEI.org/housing.
8%
9%
10%
11%
12%
13%
14%
15%
8%
9%
10%
11%
12%
13%
14%
15%
Sep-12 Mar-13 Sep-13 Mar-14 Sep-14 Mar-15 Sep-15 Mar-16 Sep-16 Mar-17 Sep-17 Mar-18 Sep-18
Purchase NMRI
Refi NMRI
Stressed default rate
Purchase and Refi NMRI
Red markers show November stressed default rate in each year.
FHA’s Pro-Cyclical Policies Continue to Fuel the Boom
• Pro-cyclical policies support the housing market when the market is going
up, and withdraw support when the market is going down. Therefore,
such policies push the market further away from its long-term mean,
which ends up prolonging booms and worsening busts.
• FHA’s mortgage risk indices jumped in Nov. setting new series’ highs
– Purchase index at 28.5%
– Refi index at a high for the month of November, with the Cash-Out Refi NMRI at a series’
high.
• Higher NMRI indicates FHA continues to increase leverage to maintain
levels of mortgage activity and to further their “affordable housing”
mission.
– FHA’s credit box is wide, therefore credit for entry-level buyers is not tight.
– FHA continues to loosen at a breath-taking pace.
– FHA is adding mostly high risk borrowers, whose risk index keeps climbing through the
effects of risk layering.
56
Which Risk Factors Have Driven Up the Purchase NMRI?
• Since 2013, all the key risk factors have contributed, which has magnified the effect on the NMRI through risk layering. An increasing share of loans have:
– Subprime credit scores
– High DTIs
– High CLTVs
– 30-year terms
• Major upward moves by each are in red font.
• Over the past two years, DTIs have moved higher, promoting risk layering. 57
Primary home purchase loansFreezing FHA at its
Oct-12 shares
Risk factor Oct-12 Oct-14 Oct-16 Oct-18 Oct-18
Credit score < 660 13% 16% 16% 18% 13%
DTI > 43% 21% 24% 27% 38% 33%
CLTV ≥ 95% 57% 56% 58% 58% 58%
30-year term 94% 94% 95% 96% 95%
Risk layering 20% 26% 28% 34% 30%*Risk layering is defined as having at least 3 of the 4 features presented in the table above present in a loan.
Note: Calculated for primary home purchase loans with a government guarantee and reported risk factor. Data for the last column
hold FHA’s shares for each risk factor constant at their Oct-2012 level, thereby assuming no credit easing for FHA.
Source: AEI Housing Center, www.AEI.org/housing.
Not updated
FHA’s NMRI for Home Purchase and Refinance Loans
Source: AEI Housing Center, www.AEI.org/housing.
All of FHA’s indices have consistently been trending up since early-2013 (earliest data available). For comparison purposes, Rural Housing Services’ Purchase
MRI has been flat. Unless FHA makes policy changes, its current credit box will continue to lean into the current housing boom, thereby leading the way in the
promotion of unsustainable home price increases.
58
16%
18%
20%
22%
24%
26%
28%
30%
16%
18%
20%
22%
24%
26%
28%
30%
Feb-13 Aug-13 Feb-14 Aug-14 Feb-15 Aug-15 Feb-16 Aug-16 Feb-17 Aug-17 Feb-18 Aug-18
FTB
Refi CO
Repeat Buyer
Refi NCO
RHS FTB
Stressed default rate
What explains FHA’s riskiness?
• Across all risk factors FHA is more risky than the rest of the Agency Market.
• Over the past 2 years an increasing share of FHA loans has had higher DTIs and lower credit scores.
• With term and CLTV basically maxed out, further FHA loosening will have to come from subprime credit score borrowers (<660) or higher DTIs.
59
Primary Home Purchase Loans
FHA Rest of Agency
Risk factor Nov-16 Nov-18 Nov-16 Nov-18
Credit score < 660 36% 46% 8% 9%
DTI > 43% 48% 60% 20% 32%
DTI > 50% 19% 30% 3% 4%
CLTV ≥ 95% 91% 90% 45% 47%
30-year term 99% 100% 93% 94%
% high risk loans 88% 93% 24% 29%
Note: Calculated for primary home purchase loans with a government guarantee and reported risk factor.
Source: AEI Housing Center, www.aei.org/housing/.
How wide is the FHA credit box?
60
*2018 data for downpayment assistance are from September. Source: AEI Housing Center, www.aei.org/housing, and FHA Snapshot data.
FHA borrowers are the marginal borrowers. FHA’s credit box is wide and its riskiest portions are being used more and more. It spans as low as a 580 credit score, as high
as a 57 DTI, and generally a 98.2 CLTV. In addition, about 1/3 of FHA borrowers make no downpayment. Therefore, the credit box for the marginal buyer is not tight, it is loose.
FHA Primary Home Purchase Loans
PercentileCredit Score DTI (in %) CLTV
Nov-12 Nov-18 Nov-12 Nov-18 Nov-12 Nov-18
5 631 600 54 57 99 103
10 642 614 52 56 99 99
25 659 636 47 52 99 99
50 / median 688 664 41 46 99 99
75 731 699 34 39 97 99
90 770 737 27 32 95 95
average 697 670 40 45 97 98
Share that received
downpayment assistance25 32
Risk Overlap between FHA and the GSEs
61Source: AEI Housing Center, www.aei.org/housing.
While there is a clear separation between FHA and the GSEs at the high and low ends
of the risk spectrum, there is substantial competition for borrowers with MRIs between
8-20%. Over the past year, the GSEs have moved out the risk curve and therefore
gained market share from FHA. On the other hand, FHA has even further moved out its
risk curve and has therefore been picking up borrowers with MRIs > 32%.
0
10
20
30
40
50
0-4 4-8 8-12 12-16 16-20 20-24 24-28 28-32 >32
November 2018 Distribution
GSEs FHA
Percent of loans
-8
-4
0
4
8
12
16
20
0-4 4-8 8-12 12-16 16-20 20-24 24-28 28-32 >32
Change in Distribution, November 2017 to November 2018
GSEs FHA
Percentage points
FHA NMRI by Risk Decile, Home Purchase Loans
All FHA loans are increasing in risk, but alarmingly, it is the riskiest FHA loans that
are getting even riskier. As house prices and leverage continue to rise, it will be
largely borrowers at the lower end of the market that will continue to add on risk
and drive up house prices for everyone.
62Source: AEI Housing Center, www.AEI.org/housing.
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
1 2 3 4 5 6 7 8 9 10
Risk Decile (1 = lowest, 10 = highest)
Stressed default rate
Nov-2018
Nov-2012
FHA Median Downpayment and Sales Price
63
With both equity and income leverage increasing, the long running boom in home
prices is not only shows no signs of abating, but is rather accelerating. One sign of
growing equity leverage is the fact that for home buyers guaranteed by FHA, the dollars
of initial equity has stayed roughly the same since September 2012, while home prices
over the same period have increased by 27%.
Note: In April 2017, Ginnie Mae started including the FHA upfront mortgage insurance premium in the LTV. Due to this switch and lagged reporting of loans for March 2017, this month’s median downpayment is imputed by averaging the median downpayments for February and April 2017, which are largely unaffected by Ginnie Mae’s reporting change.Source: AEI Housing Center, www.aei.org/housing.
$0
$2,000
$4,000
$6,000
$8,000
$10,000
$100,000
$120,000
$140,000
$160,000
$180,000
$200,000
$220,000
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17 Sep-18
Median Sales Price (left axis)
Median Downpayment(right axis)
% of FHA borrowers receiving downpayment assistance:
Oct-12: 25%Oct-13: 30%Oct-14: 33%Oct-15: 35%Oct-16: 33%Oct-17: 32%Sep-18: 31%
FHA’s Should Begin Reining in Its Pro-Cyclical Policies
• Start by taking immediate steps to reduce the risk posed to it and its
borrowers by an excessively risky credit box:
– Eliminate DTIs above 50% on 30-year term loans with credit scores below 660
– Reduce seller concessions to 3% on 30-year term loans
– Eliminate cash out refinances
– Crowd in 20-year term loans by lowering mortgage insurance premium and allowing
somewhat expanded DTIs and seller concessions
64
BCFP’s Pro-Cyclical QM Patch Continues to Fuel the Boom • In 1.13, “Ability-to-Repay and Qualified Mortgage Standards” rule was issued, effective
1.10.14, pursuant to the Dodd-Frank Act’s calling for minimum mortgage standards
• The Bureau noted it will “protect consumers from irresponsible mortgage lending.” • The rule effectively set a maximum debt-to-income (DTI) limit of 43% for the private sector.
• Temporary GSE QM Patch (the QM Patch exempted the GSEs and their automated underwriting systems from this provision for seven years.
• Similarly, FHA, the VA and the Department of Agriculture’s Rural Housing Services (RHS) , were exempted for up to seven years or until these agencies issued their own rules codifying their own lending practices (which all subsequently did).
• It was to make sure “prime” loans will be made responsibly, yet it sets no minimum down payment, no minimum standard for credit worthiness, and no maximum debt-to-income ratio (for government agencies)
• Under this definition of “prime”, a borrower can have no down payment, a credit score of 580, and a debt-to-income ratio over 50% as long as they are approved by a government-sanctioned underwriting system.
• It was foreseeable that this rule would promote an unsustainable home price boom:• In 2013: “Booms are fueled by excessive leverage” and “this rule does little to limit borrower leverage and lays
the foundation for the next bust.”* **
• In 1951: “[In transitioning] from a buyer's to a seller's market, maximum terms become so commonly used they tend to be considered the minimum.”***
• Flaw 1: The QM Patch does not operate counter-cyclically so as to “take the punch bowl away” during a leverage-fueled price boom.
• Flaw 2: The QM Patch has crowded out the private market, leaving it more risky scraps. *Pinto, “CFPB’s new ‘qualified mortgage’ rule: The devil is in the details”, http://www.aei.org/publication/cfpbs-new-qualified-mortgage-rule-the-devil-is-in-the-details/
**Wallison and Pinto, “New Qualified Mortgage rule setting us up for another meltdown” https://www.washingtontimes.com/news/2013/mar/3/wallison-and-pinto-new-qualified-mortgage-rule-set/
***Fisher, Financing Home Ownership, NBER, 1951
**** WSJ, No Pay Stub? No Problem. Unconventional Mortgages Make a Comeback, 1.23.19 65
66
Pro-Cyclical Parallels to the Last Boom
In the last 20 years we have experienced two gigantic house price booms.* It is no coincidence that rising debt-to-income ratios have provided the fuel for both the last
and current house price boom. Going back to at least the 1930s, there is no other time when DTIs were so high (or interest rates so low).
* Shiller, The Housing Boom Is Already Gigantic. How Long Can It Last?, New York Times, 12.7.18
Note: Rate for 1988-1991 is conservatively estimated at 5 percent, and is likely well below that rate.
Source: AEI Housing Center, www.AEI.org/housing, Fannie Mae, BCFP, and FHFA.
60
70
80
90
100
110
120
130
140
150
160
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
GSE Loan Share with DTI ≥ 42%(blue, left axis)
Real house prices indexed to 100 in 2000(red, right axis)
GSE Loan Share with DTI ≥ 42% and Real House Prices
67
Purchase Loans with Total DTI Greater than 43%As we have been predicting, the share of loans with DTI > 43% is now growing rapidly to
compensate for faster home price increases compared to incomes, a trend most pronounced for Fannie (+3.9 ppts over past 12 months) and FHA (+6.7ppts). Despite
Fannie’s announcement in March to update its Desktop Underwriting, after it had first raised the DTI limit to 50 in August 2017, there is little evidence that it has actually
reigned in this segment. The only exceptions to the trend are RHS and Portfolio lenders.
Note: Data pertain to purchase loans for primary owner-occupied properties. Data for the portfolio line come from LLMA and McDash after removing duplicative loans. The data are weighted
by loan amount buckets and origination year using HMDA weights (lag due to time needed to allow for sales to GSEs). Weights for 2018 are assumed to be identical to 2017.
* A seller’s market, defined by the National Association of Realtors (NAR) as a home inventory supply of 6 months or less, has been present since Sept. 2012.
Source: AEI Housing Center, www.AEI.org/housing, CoreLogic, and Black Knight.
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
55%
60%
65%
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
55%
60%
65%
Feb-13 Jul-13 Dec-13 May-14 Oct-14 Mar-15 Aug-15 Jan-16 Jun-16 Nov-16 Apr-17 Sep-17 Feb-18 Jul-18
FHA
Fannie
VA
Agency composite
Freddie
RHS
Portfolio
It Is Time for the BCFP to Announce that the Temporary GSE QM Patch Will Be Allowed to Sunset in January 2021
• As is noted on Slide 19, the percentage of agency loans with DTIs greater than 43% has exploded since the QM rule was announced in 2013, a period that coincides with the current home price boom
• The BCFP, in its January 2019 report, found*:– The continued prominence of Temporary GSE QM originations is contrary to the Bureau’s
expectations at the time of the rulemaking, and certain goals of the Rule have therefore not been met.
– In accounting for the continued prominence of Temporary GSE QM originations, two factors can be distinguished. First, the scope of GSE-eligible loans is broad, and it grew even broader for a period of time after the Rule became effective as the GSEs loosened their credit eligibility in various respects.
– In contrast, the underwriting guidelines and DTI limits for General QM loans have remained static since they were issued.
• BCFP should immediately take steps to:– Announce that the GSE patch will not be renewed.
– Provide guidance to GSEs that they should immediately begin reducing industry’s reliance on patch in a measured manner, thereby reducing any market impacts between now and the 2021 expiration of the patch.
– Coordinate with HUD/FHA on reductions to its DTI policies as part of a broader effort to counter-cyclically slow down the home price boom.
– Indicate it will be looking at changes to the QM rule so that, in the future, it has a counter-cyclical component.
* BCFP, Ability-to-Repay and Qualified Mortgage Rule Assessment Report, January 10, 2019
68
The Role of Leverage
Despite worsening affordability, leverage is allowing lower price tier borrowers to forego a quality adjustment. The same concept applies when mortgage rates rise.
69
Cumulative Constant-quality and Market Expenditure House Price Appreciation Indices (Oct-2012 = 0%)
Note: HPIs are smoothed around the times of FHFA loan limit changes.Source: AEI Housing Center, www.AEI.org/housing.
Quality offset
Quality offset
0%
10%
20%
30%
40%
50%
60%
Ho
use
Pri
ce A
pp
reci
atio
n
Low Price Tier
Constant-quality
Market Expenditure
0%
10%
20%
30%
40%
50%
60%
Ho
use
Pri
ce A
pp
reci
atio
n
High Price Tier
Constant-quality
Market Expenditure
Ratio of Sales Price for First-time to Repeat Buyers
The trend upward is towards higher first-time buyer (FTB) prices relative to repeat buyers (RBs). FTBs have access to the leverage punchbowl, thereby greatly reducing
the tendency to make downward quality adjustments to offset rapid home price appreciation. RBs without access to this punchbowl, tend to make downward quality
adjustments to offset home price appreciation. This adds to demand at lower price tiers.
70
Source: AEI Housing Center, www.AEI.org/housing.
70%
71%
72%
73%
74%
75%
70%
71%
72%
73%
74%
75%
Feb-13 Feb-14 Feb-15 Feb-16 Feb-17 Feb-18
Red markers show ration in November in each year.
Ratio of FTB to RB sale price
Median Sale Price by Market Segment*, FTB Purchase Loans
* We define prime loans as low-risk (with a stressed default rate of less than 6%), and subprime as high risk (with a stressed default rate of 12% or greater).Source: AEI Housing Center, www.AEI.org/housing.
Higher risk borrowers are being provided additional leverage which is fueling rapidly increasing home prices. Market prices for subprime borrowers have increased 25 percent since Feb-2013, while market prices for prime borrowers have only increased 11 percent.
71
95
100
105
110
115
120
125
130
95
100
105
110
115
120
125
130
Feb-13 Aug-13 Feb-14 Aug-14 Feb-15 Aug-15 Feb-16 Aug-16 Feb-17 Aug-17 Feb-18 Aug-18
Subprime (4.0% avg. annual growth)
Nov-18: $218,000
Prime(1.9% avg. annual growth)
Nov-18: $304,000
Prime 270,000$
Subprime 170,000$
Mean price Feb-13Prime Subprime
Feb-13 3.0% 20.4%
Nov-18 3.2% 22.7%
NMRI
Feb-2013 = 100
Measure Market Behavior in Four Leverage Based Price Tiers
Note: Dare are for largest 73 CBSAs and consist of 8.5 million sale transaction study covering 5-years of home price appreciation (HPA) for 41,000 census tracts. Weighting
based on HMDA. Shares based on count. Low & med-low price tiers defined respectively as <=40th & >40th to <=80th percentile of FHA sales prices & med-high & high price
tiers defined respectively as >80th percentile of FHA sales prices & <= 125% of GSE limit & > 125% of GSE limit, all at county-level. Mortgage Risk (Leverage) Loan Grades:
High risk = >12%, Medium risk = >6%-12%, Low risk = <=6%
Source: AEI Housing Center, www.AEI.org/housing.
One of AEI’s innovations to track home price appreciation is to use four price bins, because the
market behaves differently in each price bin.
• “Low” bin has all sales priced less than the bottom 40% of sales prices for FHA insured homes.
• The “Low-Medium” bin has all sales priced in the next 40% of sales prices for FHA insured
homes.
Most first time buyers (FTB) in the bottom two bins, and their mortgage loans are much riskier.
By contrast, the top two bins have relatively fewer FTBs, and buyers have much less risky loans.
72
73
House Price Trends Impacted by Leverage
On a constant-quality basis and market price basis, prices of low and low-medium priced homes have increased much faster than medium-high and high priced homes. With easy access to government-supplied leverage, buyers in low and low-medium
tiers have had to make little compromise on quality.
Source: Dare are for largest 73 CBSAs and consist of 8.5 million sale transaction study covering 5-years of home price appreciation (HPA) for 41,000 census tracts. Weighting based on
HMDA. Shares based on count. Low & med-low price tiers defined respectively as <=40th & >40th to <=80th percentile of FHA sales prices & med-high & high price tiers defined
respectively as >80th percentile of FHA sales prices & <= 125% of GSE limit & > 125% of GSE limit, all at county-level. HPIs are smoothed around the times of FHFA loan limit changes.
Source: AEI Housing Center, www.AEI.org/housing.
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
2012:Q4 2013:Q4 2014:Q4 2015:Q4 2016:Q4 2017:Q4
Cumulative Constant-Quality HPI, by Price Tier (2012:Q4 = 0%)
Low
Low-Med
Med-High
High
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
2012:Q4 2013:Q4 2014:Q4 2015:Q4 2016:Q4 2017:Q4
Cumulative Market Expenditure HPI, by Price Tier (2012:Q4 = 0%)
Low
Low-Med
Med-High
High
Not updated
74
Constant-quality Prices by Guarantor Type: Low Price Tier
FHA, GSE, & Private HPI for the low priced tier all went up about the same amount over 5 years—
45%. Buyers with high mortgage risk set the price in this and low-medium market segment. VA &
RHS had lower price gains, likely due to differing appraisal practices & DTI limitations.
Source: Dare are for largest 73 CBSAs and consist of 8.5 million sale transaction study covering 5-years of home price appreciation (HPA) for 41,000 census tracts. Weighting based on
HMDA. Shares based on count. Low & med-low price tiers defined respectively as <=40th & >40th to <=80th percentile of FHA sales prices & med-high & high price tiers defined
respectively as >80th percentile of FHA sales prices & <= 125% of GSE limit & > 125% of GSE limit, all at county-level. HPIs are smoothed around the times of FHFA loan limit changes.
Data for RHS are not available in years for which HMDA data has not yet been published.
Source: AEI Housing Center, www.AEI.org/housing.
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
2012:Q4 2013:Q2 2013:Q4 2014:Q2 2014:Q4 2015:Q2 2015:Q4 2016:Q2 2016:Q4 2017:Q2 2017:Q4 2018:Q2
Cumulative Constant-quality House Price Index, by Guarantor Type: Low Price Tier (2012:Q4 = 0%)
FHA
GSE
Portfolio
VA
RHS
Not updated
75
High risk home purchase lending is fueling home price appreciation
In the largest 73 metros, currently 41% of agency purchase lending is high risk. FHA accounts for 57% of this high risk lending, which is down from 74% in 2012. Significantly, the GSEs account for nearly all of this high risk share shift. Their high risk share has increased from 10% in 2012 to 30% in 2018.
Source: Dare are for largest 73 CBSAs and consist of 8.5 million sale transaction study covering 5-years of home price appreciation (HPA) for 41,000 census tracts. Weighting based on
HMDA. Shares based on count. Low & med-low price tiers defined respectively as <=40th & >40th to <=80th percentile of FHA sales prices & med-high & high price tiers defined
respectively as >80th percentile of FHA sales prices & <= 125% of GSE limit & > 125% of GSE limit, all at county-level. HPIs are smoothed around the times of FHFA loan limit changes.
Data for RHS are not available in years for which HMDA data has not yet been published.
Source: AEI Housing Center, www.AEI.org/housing.
76
Scatterplots: Introduction
The scatter charts on the two slides that follow show correlations at the census tract level relating to mortgage risk which measures expected default rates under stress (x-axis) and: • The ratio of tract home price appreciation (HPA) to county HPA, • Income as a percent of metro area income.
The scatterplots are binned to better show the trend. Instead of a standard scatterplot, which plots all the data points, the binned scatterplot only plots the binned data points.
The scatter dots for each chart are color coded based on the percentage of high risk purchase loans as a share of all purchase loans in the tract. • Those from the green color palette have a high risk share of less than 30%.• Those from the blue color palette have a high risk share of greater than or equal to 30%
There is a strong positive correlation between higher mortgage risk (higher expected default rates under stress) and higher home price appreciation, lower home prices, and
lower income.
Source: Dare are for largest 73 CBSAs and consist of 8.5 million sale transaction study covering 5-years of home price appreciation (HPA) for 41,000 census tracts.
Source: AEI Housing Center, www.AEI.org/housing.
77
Strong Positive Correlation Between Mortgage Risk &
Home Price Appreciation
Note: Instead of a standard scatterplot, which plots all the data points, the binned scatterplot only plots the binned data points. It first groups the x-axis variable into equal-sized bins and
then computes the mean of the x and y-axis variables within each bin thereby simplifying the plot while keeping the relationship between x and y variable intact. High risk loans are
defined as loans with a Mortgage Risk Index ≥12%.
Source: AEI Housing Center, www.AEI.org/housing.
House price appreciation increases with a census tract’s mortgage risk index:
• For the dark green dots (MRI < 15%), the median ratio of tract to county house
price appreciation is 0.86
• For the dark purple dots (MRI ≥ 60%), ratio is 1.19—a 38% higher level of price
appreciation
Together the blue color palette tracts (MRI ≥ 30%) represented about 50% of all sale
transactions.
78
Strong Positive Correlation Between Mortgage Risk & Tract Income
Note: Instead of a standard scatterplot, which plots all the data points, the binned scatterplot only plots the binned data points. It first groups the x-axis variable into equal-sized bins and
then computes the mean of the x and y-axis variables within each bin thereby simplifying the plot while keeping the relationship between x and y variable intact. High risk loans are
defined as loans with a Mortgage Risk Index ≥12%.
Source: AEI Housing Center, www.AEI.org/housing.
• Dark green dots on the right, with <15% high risk loans, had low average tract MRIs (about 3-6%) and dark purple dots on the right, with >=60% high risk loans, had high tract MRIs (about 17-23%)
• For the dark green dots, the median tract income was 158% of metro area income, while for the dark purple dots, the median tract income was 89% of metro area income
• 75% of the census tracks with median income below 120% of metro area income had average tract MRIs of 9% or greater
Measured Steps Now Would Moderate Unsustainable Home Price Increases, Not Lead to Home Price Declines
Unlike FHA, rural housing services (RHS) has not moved out risk curve during boom
2.0, keeping housing more affordable for RHS buyers. RHS’ stressed default rate is
unchanged over the last 5+ years, while FHA’s First-Time Buyer (FTB) risk index has
increased from 21.5% to 28.9%. (The same increases apply to other FHA risk indices.)
79Source: AEI Housing Center, www.AEI.org/housing.
18%
20%
22%
24%
26%
28%
30%
18%
20%
22%
24%
26%
28%
30%
Feb-13 Aug-13 Feb-14 Aug-14 Feb-15 Aug-15 Feb-16 Aug-16 Feb-17 Aug-17 Feb-18 Aug-18
FHA FTB
RHS FTB
Median saleprice
November 2013 November 2018 Change from November 2013 to November 2018
RHS -$2,200 -$800 6%
FHA $3,700 $3,900 29%
Median downpayment
DTI Distributions, GSE & FHA Purchase Loans
Source: AEI Housing Center, www.AEI.org/housing.
DTIs have been shifting higher as the rise in house prices has been outpacing income gains. The credit easing race between the GSEs and FHA continues. After
Fannie (and Freddie) eliminated compensating factors in July 2017, virtually all GSE borrowers, not just those around the previous DTI limit of 45 percent, have shifted to
higher DTIs. We expect FHA volume to continue to shift to higher DTIs.
80
0%
1%
2%
3%
4%
5%
6%
1 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50
DTI
GSE
July 2017
Nov 2018
0%
1%
2%
3%
4%
5%
6%
1 20 24 28 32 36 40 44 48 52 56
DTI
FHA
July 2017
Nov 2018
Share of Fannie Purchase Loans by DTI Bucket
Source: AEI Housing Center, www.AEI.org/housing/.
Despite Fannie’s announcement in March to update its Desktop Underwriting after it had first raised the DTI limit to 50 in August 2017, there is little evidence that it has actually
reigned in this segment. Compared to Feb-2018, the pullback was minor and the share of loans with a DTI in excess of 44 is still much greater than just a year ago.
81
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
<20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
DTI
Fannie Purchase Loans by DTI Bucket Jul-17
Fannie Purchase Loans by DTI Bucket Feb-18
Fannie Purchase Loans by DTI Bucket Nov-18
Share of Fannie Purchase Loans
RHS Reduced Borrower DTIs from 2013 to 2018, while the FHA Kept Increasing DTIs
DTIs limits act as counter-cyclical friction to slow the increase of house prices when
supply is tight. Remove the friction and house prices increase, fueling a boom.
82
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
<20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58
Purchase Loans by DTI Bin: February 2013
FHA
RHS
In 2013 RHS appears to have had a semi-hard stop at 43%.
RHS also allowed DTIs up to 48%, based on compensating factors. DTIs above 48% were rare.
In 2013 FHA appears have had a semi-hard stop at 50% DTI.
FHA also allowed higher DTIs, generally up to 57%, with compensating factors. DTIs above 57% were rare.
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
< 20 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57
Purchase Loans by DTI Bin: February 2018
RHS
FHA
By 2018, RHS was acting counter-cyclically against the house price boom by lowering its semi-hard DTI limit from 43% to 41%.
More importantly, RHS's hard stop was reduced to 46% from 48% As a result, since 2013 the percentage of loans with DTIs greater then 43% DECLINED from about 20% to about 10%.
However FHA was pro-cyclically fueling the house price boom. Loans with DTIs greater than 43% increased from 37% in 2013 to 55% in 2018. Those above 50% to 57% more than doubled to 27%. Additionally there is no evidence of the use of compensating factors.
Source: AEI Housing Center, www.AEI.org/housing.
Not updated
The FHA’s and the GSEs’ Rising DTIs Have Been Pro-Cyclically Fueling the House Price Boom
Under QM, their credit boxes allow for DTIs well above 43%. As a result, DTIs have
increased dramatically. It is the use of compensating factors that reduces risk
layering, which is an important policy during a boom. However, the use of
compensating factors has been reduced markedly.
83
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
<20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56
Purchase Loans by DTI Bin: February 2013
FHA GSE
In 2013 FHA appears have had a semi-hard stop at 50% DTI.
In 2013 the GSEs had a semi-hard stop at 45% DTI.
In 2013 the GSEs also allowed relatively few DTIs up to a hard stop of 50% with compensating factors.
FHA also allowed many DTIs up to 57% with limited use of compensating factors.
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
<20
21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57
Purchase Loans by DTI Bin: February 2018
FHA GSE
FHA was also acting pro-cyclically. DTIs >43% increased from 37% of loans in 2013 to 55% in 2018. Those >50% up to 57% more than doubled to 27%. There is no strong evidence indicating the use of compensating factors.
Over 2013-2018 the GSEs were pro-cyclically fueling the boom. DTIs >43% increased from 13% in 2013 to 27% in 2018.
The GSEs' requirement for compensating factors was removed in 2017. As a result, DTIs >45% up to 50% increased from 3.5% of loans to 19%.
Source: AEI Housing Center, www.AEI.org/housing.
<20 GSE bin has a value of 12%
Not updated
Origination Shares Issuer Lender Type,
FHA and RHS Purchase Loans
84
Similar dramatic market shifts occurred from large banks to nonbanks for both FHA and RHS loans. Today nonbanks account of 80% of FHA and RHS originations.
*Origination shares do not show shares for State Housing Finance Agencies and Credit Unions which account for about 4% of the FHA Purchase market and 1% of the RHS Purchase market.Source: AEI Housing Center, www.AEI.org/housing.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
Sep-12 May-13 Jan-14 Sep-14 May-15 Jan-16 Sep-16 May-17 Jan-18
FHA Purchase Origination Shares*
Large banks
Nonbanks
Other banks
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
Sep-12 May-13 Jan-14 Sep-14 May-15 Jan-16 Sep-16 May-17 Jan-18
RHS Purchase Origination Shares*
Nonbanks
Large banks
Other banks
Not updated
MRIs by Issuer Lender Type,
FHA and RHS Purchase Loans
85
In the case of the FHA, migration to nonbanks has boosted overall risk level, as its wide-open credit box encourages higher risk lending by nonbanks who originate what they can sell and sell what they originate. This has not happened with RHS, apparently
due to a risk management approach that monitors risk so as not to lean into the current house price boom. Counter-cyclical policies are key to not promoting a boom.
Source: AEI Housing Center, www.AEI.org/housing.
19%
21%
23%
25%
27%
29%
Sep-12 May-13 Jan-14 Sep-14 May-15 Jan-16 Sep-16 May-17 Jan-18
FHA Purchase Mortgage Risk Indexes
Composite
Large banks
Nonbanks
Other banks
Nonbank MRI at 28%, up from 23% 5 years prior
14%
15%
16%
17%
18%
19%
20%
21%
22%
Sep-12 May-13 Jan-14 Sep-14 May-15 Jan-16 Sep-16 May-17 Jan-18
RHS Purchase Mortgage Risk Indexes
Nonbanks
Large banks
Other banks
Composite shown by blue line Nonbank MRI at 19%, down from 20% 5 years prior
Not updated
Origination Shares and MRIs by Seller Lender Type,
GSE Refinance Loans
Note: Data for most recent months may understate large-bank share by perhaps 2 percentage points, as large banks are slower to move recent originationsto the guarantee agencies for securitization and our market shares are based on securitized loans. MRI for state housing agencies not shown because loanvolume is nil.*Origination shares do not show shares for State Housing Finance Agencies and Credit Unions which account for about 3% of the GSE Refi market.Source: AEI Housing Center, www.AEI.org/housing.
86
Shift away from large banks in GSE refi market has mirrored that in GSE purchase market. Banks (both large and other) have lower risk profile than nonbanks.
0%
10%
20%
30%
40%
50%
60%
70%
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17 Sep-18
Refi Origination Shares*
Other banks
Large banks
Nonbanks
5%
7%
9%
11%
13%
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17 Sep-18
Refi Mortgage Risk Indexes
Other banks
Nonbanks
Composite shown by blue lineLarge banks shown by black line
Origination Shares and MRIs by Issuer Lender Type,
FHA Refinance Loans
Note: Data for most recent months may understate large-bank share by perhaps 2 percentage points, as large banks are slower to move recent originationsto the guarantee agencies for securitization and our market shares are based on securitized loans. MRI for state housing agencies and credit unions not shown because loan volume is nil.*Origination shares do not show shares for State Housing Finance Agencies and Credit Unions which account for about 1% of the FHA Refi market.Source: AEI Housing Center, www.AEI.org/housing.
87
Massive shift from large banks to nonbanks in FHA refi market. Nonbanks now have 94% of the market, along with a higher risk profile than large banks.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Sep-12 Jun-13 Mar-14 Dec-14 Sep-15 Jun-16 Mar-17 Dec-17 Sep-18
Refi Origination Shares*
Other banks
Large banks
Nonbanks
15%
17%
19%
21%
23%
25%
27%
29%
Sep-12 Jun-13 Mar-14
Dec-14Sep-15 Jun-16 Mar-17
Dec-17Sep-18
Refi Mortgage Risk Indexes
Other banks
Large banks
Composite
Nonbanks
88Source: AEI Housing Center, www.AEI.org/housing.
State NMRI and FHA Share, Purchase Loans
The share of FHA purchase loans in a state is heavily correlated with overall lending risk. FHA, as the riskiest lender by far, is accounting for a significant portion of risk, but is
also moving the risk curve out for other agencies.
AK
ALAR
AZ CA
CO
CT
DC
DE
FL
GA
HI
IAID
IL
IN
KS
KY
LA
MA
MD
ME
MI
MN
MO
MT
NC
ND
NE
NHNJ
NM
NV
NY
OH
OK
OR
PA
RI
SC
SD
TN
TX
UT
VA
VT
WAWI
WV
WY
R² = 0.679
6%
7%
8%
9%
10%
11%
12%
13%
14%
15%
16%
5% 10% 15% 20% 25% 30% 35% 40%
Stat
e N
MR
I
FHA Share of Agency Purchase Loans
State NMRI and State FHA Share of Agency Loans, August-October 2018
State NMRI Change, Purchase Loans
89
The states with the largest FHA and greatest risk levels have experienced faster growth in risk. All but three states have seen their risk levels increase over the past 5 years.
Source: AEI Housing Center, www.AEI.org/housing.
Change in State NMRI:
Jan. 2013 to Jan. 2018 (in ppts)
Not updated
Pricing Changes, Home Purchase Loans
• The GSEs find themselves in a multi-faceted competitive situation• At one end is the FHA which neither prices nor underwrites for risk• At the other end, the GSEs have risk-based loan level fee adjustments and private
mortgage insurers are required to hold capital in a manner that more accurately reflects risk
– The recently implemented PMI premium changes lowered cost for borrowers with higher credit scores (>720) and increased cost for borrowers with lower credit scores (<700)
• To meet affordable housing goals in this difficult competitive environment, the GSEs are resorting to heavy subsidies
– However, stiff competition from FHA and due to the new PMI premium structure, the GSEs have been forced to fill affordable housing quotas with higher credit score loans (median of 737)
90
Agency Date Effect
RHS Oct. 2014 RHS raised monthly fee from 40 bps to 50 bps
FHA Jan. 2015 FHA lowered annual MIP from 135 bps to 85 bps
RHS Oct. 2015 RHS raised upfront fee from 200 bps to 275 bps
GSEs Apr. 2016
As a result of GSE-imposed private mortgage insurer (PMI) capital
requirements, industry revised premium structure to focus more on
borrower’s credit score
RHS Oct. 2016RHS lowered upfront MIP from 275 bps to 100 bps and lowered
monthly fee from 50 bps to 35 bps
Stressed Default Rates by Loan Type
Compared to an identical purchase loan, refis have higher stressed default rates across all CLTV buckets. Cash-out refis are even riskier than no-cash-out refis.
At its current level, the average CO is as risky as a >90% purchase loan and the average NCO is as risky as a mix of 81-90% and >90% purchase loans.
Reasons: weakness of appraisal process and borrower self-selection.
91
Note: All stress default rates computed for credit score of 720-769 and DTI of 39-43%.
Source: AEI Housing Center, www.AEI.org/housing.
Median Downpayments
• For agency market as a whole, median downpayment is small (5%, $10,700)
• Median is even smaller for first-time buyer loans, especially for Ginnie loans (1.8%, $2,800). Ginnie accounts for almost 60% of agency first-time buyer volume
• Traditional 20% downpayment is the norm only for Fannie/Freddie repeat buyers. Ginnie repeat buyers typically put down barely more than first-time buyers
• Hence, in today’s market, little saving or accumulated equity is needed to buy a home, particularly a first home
92
Median downpayment on primary home purchase loans,
April 2018
Guarantee agencyAll
buyersFirst-time buyers Repeat buyers
Composite 5.0% / $12,100 3.0% / $5,800 10.2% / $33,300
Fannie, Freddie 13.0% / $33,700 7.0% / $18,400 20.0% / $50,500
Ginnie (FHA, VA, RHS) 1.8% / $2,900 1.8% / $2,900 1.8% / $3,200
Note: Calculated for primary home purchase loans with a government guarantee.
Source: AEI Housing Center, www.aei.org/housing.
Not updated
Volume Growth in Counts and Dollars, Purchase Loans
As prices have been rising, dollar volume has been outgrowing count volume.
Credit easing, particularly by the FHA, is fueling this trend. This creates a vicious
cycles of price appreciation and credit easing.
Solution: dial back flow of money into housing system.
93
Source: AEI Housing Center, www.AEI.org/housing.
-10%
-5%
0%
5%
10%
15%
20%
25%
30%
35%
40%
-10%
-5%
0%
5%
10%
15%
20%
25%
30%
35%
40%
Sep-13 Jan-14 May-14 Sep-14 Jan-15 May-15 Sep-15 Jan-16 May-16 Sep-16 Jan-17
YoY Change Growth Rates
Dollar Volume
Count
Difference
Not updated
A closer look at RHS’ October 2016 MIP cut
As expected, RHS’ purchase volume jumped immediately after its MIP cut in October
2016. Since the cut, RHS has grown faster than FHA, its most direct competitor. In
January, its growth surpassed all other agencies.
94
Source: AEI Housing Center, www.AEI.org/housing.
-20%
-15%
-10%
-5%
0%
5%
10%
15%
20%
25%
30%
-20%
-15%
-10%
-5%
0%
5%
10%
15%
20%
25%
30%
Feb-16 Apr-16 Jun-16 Aug-16 Oct-16 Dec-16 Feb-17 Apr-17 Jun-17
RHS
FHA
RHS premium cut
Not updated
No-Cash Out Refi Demand and 30-yr Mortgage Rate
In our last NMRI briefing we wrote that in response to higher rates “refi volume could drop by 40% to 150,000 per month.” In January refi volume was down 40% from its peak in October. Refi demand, especially no-cash outs, and the mortgage rate are strongly
correlated.
95Source: AEI Housing Center, www.AEI.org/housing, and Freddie Mac.
3.0
3.2
3.4
3.6
3.8
4.0
4.2
4.4
4.6
4.8
5.00
50,000
100,000
150,000
200,000
250,000
300,000
350,000
400,000
450,000
500,000
Sep-12 May-13 Jan-14 Sep-14 May-15 Jan-16 Sep-16 May-17 Jan-18
30-yr FRM, scale inverted (right axis)
No-Cash Refinance loans (left scale)
Not updated
Low-Risk Origination Shares, Purchase Loans
Source: AEI Housing Center, www.AEI.org/housing.
Fannie’s low-risk (prime) share has recently dropped below 50% for the first time in the history of the series. The low risk percentage gap between Fannie and Freddie is
also the widest in series history. VA’s low-risk share is well below the GSEs’.
96
48%
50%
52%
54%
56%
58%
60%
62%
64%
66%
68%
70%
72%
74%
Sep-12 May-13
Jan-14 Sep-14 May-15
Jan-16 Sep-16 May-17
Jan-18 Sep-18
Freddie Mac
Fannie Mae
Combined
20%
22%
24%
26%
28%
30%
32%
34%
36%
38%
40%
Sep-12 May-13
Jan-14 Sep-14 May-15
Jan-16 Sep-16 May-17
Jan-18 Sep-18
VA
FHA/RHS low-risk share (not shown) averages about 2%
Calibrating Mortgage Safety
• NMRI captures the complex interplay of changes in three types of leverage: property (LTV and term), income (DTI, ARM vs. FRM, and term), and credit score
• Composite index substantially above 1990 level, but not approaching 2007 level when underwriting was exceptionally lax
• Fannie/Freddie index somewhat above 1990 level
• FHA index is extremely high. Sharp contrast with safe underwriting during 1935-55.
• VA index less than half the level of FHA, both recently and in 2007
97
NMRI – purchase loansLatest
date
Latest
Value
1935-1955
vintages (est.)
1990 vintage
(est.)
2007 vintage
(est.)
Composite index Jul 12.8% NA 6% 19%
Fannie and Freddie Jul 7.3% NA 4% 13%
FHA Jul 28.0% 3% 15% 33%
VA Jul 11.9% NA NA 15%
An index value of less than 6 is indicative of conditions conducive to a stable market.
Not updated
Credit Conditions: 1990 to 2013-14
• Clear buildup of risk from 1990-92 to 2005-07
• 2013-14 loan cohort less risky than 2005-07 cohort due to smaller percentage of loans with high DTIs, higher median credit score, and very few low/no doc loans
• Still, 2013-14 cohort is substantially riskier than 1990-92 cohort. Main differences are sharp increase in loans with high CLTVs and high debt ratios, notwithstanding much lower interest rates 98
1990-92 2000-03 2005-07 2013-14
% Loans with DTI ≥ 42% 5-10% 28% (2003) 43% (2007) 28%
Median borrower credit score 735 701 705 735-740
% Loans with credit scores < 640 9.5% 25% NA 5%
% Loans with CLTV > 90% 26% 35% 40% 45%
% Loans with CLTV ≥ 97% 1% 11% 40% 30%
% Loans Low/No Doc Nil (1992) 6% (2000-02) 25% Nil
30-year fixed interest rate 10% (1990) 7% (2001) 6.5% (2006) 4% (2013)
First-time buyers as % of primary
home purchase mortgages38-42% (1990) NA NA 50%
Perfect credit (no lates) as a % of
home purchase borrowers 57-60% (1990) NA NA 60%
NMRI 6% NA 19% 10%
Compiled by AEI. Sources: CoreLogic for DTI data for 2000-03 and 2005-07 and median credit score for 2005-07. Other data from AEI, Fannie
Mae, FHA, Equifax, Freddie Mac, FICO®, and miscellaneous other sources. In general, figures shown are for the entire purchase loan market
(conventional and government guaranteed). John Burns (John Burns Real Estate Consulting) collaborated on the presentation format.
Role of Income Leverage During Housing Boom
• Less attention paid to income leverage than to property and credit leverage
– Owes to data scarcity, as the FHA and GSEs published virtually no DTI data until 2013, when FHFA released DTI trends for the period 1996 onward for both the GSEs and FHA
• Over 1996-2005, higher income leverage raised overall home purchase buying power 46%, three-quarters of the 62% rise in real home prices*
– GSE median housing DTIs (purchase transactions): 23% in 1996, 27% in 2005 ― 17% boost in buying power (based on 8% interest rate for both years)
– Median loan rates fell from 8% in 1996 to 6% in 2005 ― 20% added boost in buying power
– Low doc/no doc loan share was near 0% in 1996 with minimal income overstatement. By 2005, share was 15% with 25% income overstatement (source: CoreLogic). This increased housing DTI in 2005 another percentage point to 28% (4% added boost in buyer power).
• Push/pull of increasing leverage at work today
– Home Prices Start to Heat Up: Double-digit growth arrives in more cities, but affordability worries emerge amid thin supply (WSJ, May 12, 2015)
• NMRI tracks changes in income leverage
– Since Nov. 2012, median total DTI for all agency primary purchase loans increased from 36% to 38%, leaving buyers more highly leveraged even as income volatility increases: Cash Crunch Is, for Many, a Monthly Problem (WSJ, May 20, 2015)
* Does not take into account increases in income, home size or quality. Ex. the median new home size increased 14% from 1996 to 2005.99
Fed Tightening and Efforts to Maintain Buying Power
• Historical precedent: end of the Fed’s interest rate peg in effect from World War II
– Long-term mortgage rate rose from 4.1% in 1953 to 6% in 1962
– 5 amendments to National Housing Act (1954-61) increased FHA’s LTV and loan term limits
– These changes, along with rising housing DTIs, kept buying power constant from 1953 to 1962
• Today: with the Fed now starting to tighten, long-term interest rates will rise
– All else equal, a rise in the 30-year mortgage rate from 4% to 6% would reduce buying power by same amount as a 19% jump in home prices
• Two steps would keep buying power largely constant with no change in income-to-house price ratio
– Reduce FHA’s annual premium an additional 35 basis points to 0.50% (requires action by FHA and would further stress FHA’s capital level)
– Boost median total DTI from 41% today to 45% for FHA and from 34% today for GSEs to 38%. These changes would be QM compliant due to the agency QM exemption.
• Very risky steps. Would result in median total DTI for FHA well above the peak level in 2005-06 and for GSEs equal to the 2005-06 peak level.
• Wealth Building Home Loan provides a sustainable alternative
*FHFA’s 2014 fee report indicates that the GSEs were undercharging on high risk loans and overcharging on low risk ones and thatoverall guarantee fees were lower than needed to meet capital return levels. This is equivalent to a hidden guarantee fee cut and could be repeated in the future.
100
Appraisals Should Be the Guard Rail Against Speculative Booms
• Property valuations and appraisals should review and provide:
– A robust and transparent opinion of a property’s most likely market price based on a systematic analysis of generally available information rather than 3 subjectively chosen comparison properties
• Including a range around the most likely market price at a specified confidence level
– Trends in and nearness to key elements of utility such as employment, shopping, transportation, other infrastructure and amenities, along with zoning, density restrictions, and tax burden that impact intrinsic value and market price
– Market conditions and an assessment of whether a substantial differential between a property’s intrinsic value and market price is substantiated by a change in utility:
• At least 10-year nominal and real home price trends and a determination as to current position in market cycle relative to equilibrium
• At least a 5-year history of buyer’s market (inventory > 6 mo.) and/or seller’s market ( ≤ 6 mo.)
– Impact on buying power over last 5 years due to changes in loan leverage or prevailing interest rates
– Current land value and land share, and trends in both
– Whether real price changes are due to leverage growth, improving utility or a combination
– A property’s overall condition and a recommendation as to any readily observable repairs necessary to make it meet generally accepted minimum property requirements
101
An appraisal should provide an opinion as to the relationship between
market selling price and intrinsic or fundamental value
Cross-subsidies Return to the GSEs
• FHFA’s Report on Single-Family Guarantee Fees in 2014 disclosed numerous instances of mispricing and cross-subsidies, a significant deviation from 2013 report1,2
– High risk 30-year loans subsidized by low risk 15-years
– Borrowers with low credit scores scores subsidized by those with high credit scores
– High LTV loans subsidized by low LTV loans
– Overall guarantee fee levels were found insufficient to meet the estimated future cost of providing the guarantee
• Mispricing promotes adverse selection and increases overall risk of mortgage finance system
– Allows the GSEs to implement guarantee fee cuts in an opaque manner
– Ability to subsidize risky loans will cause progressive loosening of underwriting standards
– As was the case the last time around, this movement out the risk curve may take 5-10 years
• NMRI is designed to track these risks in real time
1 http://www.fhfa.gov/AboutUs/Reports/Pages/Fannie-Mae-and-Freddie-Mac-Single-Family-Guarantee-Fees-in-2014.aspx2For a detailed analysis of the role robust risk-based pricing plays in promoting a fair and efficient mortgage market, see Board of Governors of the Federal Reserve System, Report to the Congress on Credit Scoring and Its Effects on the Availability and Affordability of Credit, August 2007, www.federalreserve.gov/boarddocs/rptcongress/creditscore/creditscore.pdf
102
103Source: AEI Housing Center, www.AEI.org/housing.
Change in Agency Purchase Loan Volume
Since April 2016 the GSEs have again overtaken FHA as the fastest growing agencies indicating that FHA MIP cut effect from January 2015, which led to massive poaching and
some new homebuyers, has worn off.
-30%
-20%
-10%
0%
10%
20%
30%
40%
50%
60%
-30%
-20%
-10%
0%
10%
20%
30%
40%
50%
60%
Sep-13 Feb-14 Jul-14 Dec-14 May-15 Oct-15 Mar-16 Aug-16 Jan-17 Jun-17 Nov-17 Apr-18 Sep-18
Fannie/Freddie
Composite
Percent change from 12 months earlier
FHA
GSEs: Ratio of NMRI for Loans with Total
CLTV > 95% to Loans with Lower Total CLTVs
Source: AEI Housing Center, www.AEI.org/housing.104
FHFA Director Mel Watt stated that with use of compensating factors “loans with a 3 percent down payment backed by GSEs are no riskier than those with a down
payment of 10 percent …” (Jan. 27, 2015). Based on NMRIs, this is not true.
80%
100%
120%
140%
160%
180%
200%
220%
80%
100%
120%
140%
160%
180%
200%
220%
Sep-12 Feb-13 Jul-13 Dec-13 May-14 Oct-14 Mar-15 Aug-15 Jan-16 Jun-16 Nov-16 Apr-17 Sep-17 Feb-18 Jul-18
NMRI ratio: >95% CLTV loans to 86-90% CLTV loans
NMRI ratio: >95% CLTV loans to 91-95% CLTV loans
Full use of compensating factors would imply ratios of 100%
Fannie accounts for the vast majorityof GSE loans with CLTVs > 95%
FICO® Score Distribution
Source: AEI Housing Center, www.AEI.org/housing, from FICO 8 score distribution for October 2014. Distribution for FICO scores of 300-800 and 850 directly from FICO; distribution between 800 and 850 interpolated by Edward Pinto.
FHA’s minimum scores are near the bottom of the FICO credit score distribution. An FHA borrower with a 500 credit score has an NMRI of 50%, twice as risky as today’s
median FHA loan and eight times riskier than today’s median GSE loan.
105
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
500 520 540 560 580 600 620 640 660 680 700 720 740 760 780 800 820 840
800 ≈ 80th percentile750 ≈ 62nd percentile700 ≈ 46th percentile580 ≈ 19th percentile (min for FHA, 3.5% down)500 ≈ 5th percentile (min for FHA, 10% down)
Share of total population with scorebelow level shown on horizontal axis
FICO score
Median Credit Score on Primary Purchase Loans*
*Data pertain to purchase loans for primary owner-occupied properties. Percentiles based on population of all scorable individuals. Source: AEI Housing Center, www.AEI.org/housing 106
Median scores about unchanged from January 2017. FHA’s all-buyer median at 34th
percentile of scored distribution, with room to drop given FHA’s minimum scores. In current seller’s market, this will boost home prices faster than income.
660
680
700
720
740
760
780
660
680
700
720
740
760
780
Sep-12 Feb-13 Jul-13 Dec-13 May-14 Oct-14 Mar-15 Aug-15 Jan-16 Jun-16 Nov-16 Apr-17 Sep-17 Feb-18 Jul-18
Repeat buyers
First-time buyers
All buyers
FHA only, all buyers
Aggregate Default Risk Surge for Home Purchase Loans
Is Over Five Years OldAggregate default risk (which measures the combined effect of loan-level risk and
volume) continues to rise. FHA continues to account for more than half of the aggregate agency risk.
107
Source: AEI Housing Center, www.AEI.org/housing.
5,000
10,000
15,000
20,000
25,000
30,000
35,000
40,000
45,000
5,000
10,000
15,000
20,000
25,000
30,000
35,000
40,000
45,000
Sep-12 Feb-13 Jul-13 Dec-13 May-14 Oct-14 Mar-15 Aug-15 Jan-16 Jun-16 Nov-16 Apr-17 Sep-17 Feb-18 Jul-18
Number of expected defaults under stress
Red markers show October count in each year.Composite
FHA
Other agencies
The Effect of April 2016 PMI Price Change
108
Pricing for risk matters. GSE pricing for higher credit scores are now competitive with FHA, which is reflected in changes to market shares. It has also led to GSE
gaining a greater share of lower risk FHA borrowers.
Source: AEI Housing Center, www.AEI.org/housing, and Fannie Mae.
GSEs’ Market Share in CLTV > 95%, by Credit Score Bins New PI+PMI Monthly Payment Comparison
Average NMRI in CLTV > 95% Market
Apr-Oct 2015 Apr-Oct 2016 Change
GSEs 14.3% 13.9% -0.4 ppt
FHA 24.4% 25.3% +0.8 ppt
Credit
Score Bin
PMI Pricing
Change
PMI Pricing
Compared to FHA
Risk-based required
asset amount factors
>= 760 ($101) ($5) 4.83%
740-759 ($60) $35 7.60%
720-739 ($32) $82 9.84%
700-719 ($30) $136 11.55%
680-699 $18 $187 14.25%
660-679 $119 $308 19.20%
640-659 $149 $353 19.20%
620-639 $155 $414 19.20%* assumes a 3.5% downpayment on a $250,000 home. Conforming mortgage
rate of 3.89% and a FHA mortgage rate of 3.50%.
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
Jan-15 Mar-15 May-15 Jul-15 Sep-15 Nov-15 Jan-16 Mar-16 May-16 Jul-16 Sep-16 Nov-16 Jan-17
>=760 740-759 720-739 700-719 Total
680-699 660-679 640-659 620-639 <620
Change in PMI pricing takes effect
Fannie and Freddie guarantee mortgages with as little as 3% down
Not updated
Credit Score Distribution & MRIs, Purchase Loans*
*Data pertain to purchase loans for primary owner-occupied properties. Source: AEI Housing Center, www.AEI.org/housing.
Stark contrast between credit score distributions for FHA and GSE borrowers. FHA accounts for over 80% of scores below 660, while GSEs account for nearly 90%
above 740.
109
MRIs rise as credit scores decline – evidence of risk layering rather than compensation for risk. In a seller’s market, risk layering artificially pushes up
prices, resulting in a wealth transfer from buyers to sellers of entry-level homes.
0%
20%
40%
60%
80%
100%
580-589 600-609 620-629 640-649 660-669 680-689 700-709 720-729 740-749 760-769 780-789 800-809 820-829
Share of Total, by Credit Score Bucket, October 2018FHA share GSE share
0%
10%
20%
30%
40%
50%
580-589 600-609 620-629 640-649 660-669 680-689 700-709 720-729 740-749 760-769 780-789 800-809 820-829
MRI by FICO score, October 2018
FHA
GSEs
Purchase Loans with Down Payment of 5% or Less*
*Data pertain to purchase loans for primary owner-occupied properties.Source: AEI Housing Center, www.AEI.org/housing
58% of all primary purchase loans and 36% of such Fannie/Freddie loans have a minimal down payment. With QM silent on down payments, lots of room for these shares to rise.
In current seller’s market, this will drive up home prices more than income.
110
20%
25%
30%
35%
40%
45%
50%
55%
60%
65%
20%
25%
30%
35%
40%
45%
50%
55%
60%
65%
Sep-12 Feb-13 Jul-13 Dec-13 May-14 Oct-14 Mar-15 Aug-15 Jan-16 Jun-16 Nov-16 Apr-17 Sep-17 Feb-18 Jul-18
Fannie/Freddie
Composite
In October 2018, 90% of FHA primary owner-occupied purchaseloans had a down payment of 5% or less; the share for VA was 88%.
The October 2018 share for first-time buyers was 72%.
Composite Origination Shares and MRIs by Channel,
Purchase Loans
Source: AEI Housing Center, www.AEI.org/housing. 111
Retail and correspondent shares have stabilized at around 45% each. Broker share has remained around 10%. Correspondent and broker composite MRIs tracking higher at
levels significantly above retail MRI.
0%
10%
20%
30%
40%
50%
60%
Feb-13 Dec-13 Oct-14 Aug-15 Jun-16 Apr-17 Feb-18
Composite Origination Shares
Correspondent
Broker
Retail
6%
8%
10%
12%
14%
16%
Feb-13 Dec-13 Oct-14 Aug-15 Jun-16 Apr-17 Feb-18
Composite Mortgage Risk Indexes
Correspondent
Broker
Retail
Large Bank Origination Shares and MRIs by Channel,
Purchase Loans
*Sharp drop in MRI for broker channel is due to greatly reduced volume of GNMA loans.Source: AEI Housing Center, www.AEI.org/housing.
112
Nearly all large-bank volume comes through retail and correspondent channels; broker volume has dropped to de minimis level. MRI shows that large banks are acting to limit defaults among retail customers and reducing risk tolerance on correspondent loans.*
0%
10%
20%
30%
40%
50%
60%
70%
80%
Feb-13 Dec-13 Oct-14 Aug-15 Jun-16 Apr-17 Feb-18
Large Bank Origination Shares
Correspondent
Broker
Retail
4%
6%
8%
10%
12%
14%
16%
Feb-13 Dec-13 Oct-14 Aug-15 Jun-16 Apr-17 Feb-18
Large Bank Mortgage Risk Indexes
Correspondent
Broker
Retail
113
Change in NMRI,
Oct. 2016 to Oct. 2017
Change in standards,
2016:Q4 to 2017:Q4
Agency All banksSenior Loan Officer
Survey
GSEs Some easing Some easing
FHA, VA, and RHS Some easing Little change
Note: “Easing” denotes a rise in the NMRI of 0.25 percentage point or more, “Tightening” denotes a
decline in the NMRI of 0.25 percentage point or more, and “Little change” denotes a change in the
NMRI of less than 0.25 percentage point in either direction.
Source: AEI Housing Center, www.AEI.org/housing, and Federal Reserve Board,
http://www.federalreserve.gov/boarddocs/snloansurvey/201608/fullreport.pdf)
Fed’s Senior Loan Officer Survey is Badly Flawed
• Showed no systematic loosening in mortgage lending standards in the run-up to the 2007-08 financial crisis
• Survey design problems
– Only covers commercial bank lenders
– Based on opinions of a small number of loan officers
– Weights all responses equally
• Results over past year: some easing for GSE loans, little change for Ginnie loans. Survey misses the caution prevailing at banks revealed by NMRI (see table).
• Mortgage lending standards have eased but this is due to mix shifts not captured by the survey (from banks to nonbanks and from GSEs to FHA).
• Bottom line: don’t use the Fed survey
Not updated
Urban Myth: Tight Credit Keeping
“Creditworthy” Borrowers Out of Market
• Assertion: “Today’s lenders are simply not originating loans for borrowers with less than perfect credit.” (Urban Institute, April 2015)1
• Fact: 40% of home purchase borrowers in 2013-14 had less than perfect credit (perfect being no lates)
• Fact: Median credit score for FHA purchase loans was 674 in April 2015, well below the median for all individuals in U.S. with a score
• Assertion: “Severe” 2013 standards caused 1.25 million purchase loans to be missing relative to “cautious” 2001 standards1
• Fact: 70% of these “missing” borrowers had a credit score < 660; would have an MRI above 25% due to extensive risk layering on FHA loans2
• Fact: Urban study is fatally flawed. Credit score distribution was the same in 2005 as in 2001, so the number of “missing” loans would be the same using either year as the baseline. Because credit standards in 2005 were extremely lax, this makes the notion of “missing loans” meaningless
• Fact: Credit standards in 2001 were much looser than in the early 1990s. Thus, the early ’90s would be a more appropriate baseline for cautious standards.3
1Impact of Tight Credit Standards on 2009-2013 Lending, http://www.urban.org/publications/2000165.html.2In addition to subprime credit score, initial equity of 3% or less, 30 year loan term, average total debt ratio of 41% without use of residual income.3A 1999 Urban Institute study (http://www.urban.org/publications/1000205.html) documented the easing of standards by the GSEs through 1998 but also noted that “The GSEs’ guidelines, designed to identify creditworthy applicants, are more likely to disqualify borrowers with low incomes , limited wealth, and poor credit histories; applicants with these characteristics are disproportionately minorities.” HUD relied on this study when it greatly expanded the affordable housing goals in 2000.
114
FHA Perpetuates This Myth
• FHA charges the same mortgage insurance premium regardless of borrower credit risk. Lack of risk-based pricing:1
– Misleads high-risk borrowers into thinking they are creditworthy
– Exposes FHA to adverse selection
– Is inherently unfair
– Increases overall risk of mortgage finance system.
• AEI’s Wealth Building Home Loan offers a better solution for higher-risk borrowers
1For a detailed analysis of the value of credit scoring and risk-based pricing for promoting a fair and efficient mortgage market, see Board of Governors of the Federal Reserve System, Report to the Congress on Credit Scoring and Its Effects on the Availability and Affordability of Credit, August 2007, www.federalreserve.gov/boarddocs/rptcongress/creditscore/creditscore.pdf
115
FHA’s 2007 Loan Cohort: 90-day Delinquency Rate by Credit Score
≤ 620 620-650 650-700 700-750 >750
47% 35% 25% 14% 9%
Source: Urban Institute, VA Loans Outperform FHA Loans. Why? And What Can We Learn?, table 3, panel B
http://www.urban.org/research/publication/va-loans-outperform-fha-loans-why-and-what-can-we-learn
• FHA promotes lending to very high-risk borrowers: credit score floors of 500 with 10% down and 580 with 3.5% down
• 2007 vintage of FHA loans indicative of performance under stress
FHA Is All about Moral Hazard
Moral hazard: “A situation where one party gets involved in a risky event knowing that it is protected against the risk and the other party will incur the cost.”1
• FHA insurance presents a classic case with multiple layers of moral hazard:
– FHA insures 100% of the loss for high-risk loans, has minimal capital, and is taxpayer backed
o It neither prices for risk nor underwrites for risk layering, which is inherently unfair to borrowers and exposes FHA to adverse selection.2
o Exact opposite of the original FHA structure in the 1934 National Housing Act
– Ginnie Mae and nonbank lenders, both with minimal capital, are able to ignore borrower solvency risk since they are protected by FHA
– High-risk borrowers, misled into thinking they are creditworthy, borrow more than they should.Greater borrowing spurred by recent cut in mortgage insurance premium is a textbook example.
– Increases overall risk of mortgage finance system
o Effectively unconstrained by QM,3 increasing competition between Fannie and FHA, and eventually Freddie, will cause progressive loosening of underwriting standards
o As was the case during the last boom/bust cycle, this movement out the risk curve may take 5-10 years
• NMRI is designed to track these risks in real time
1http://economictimes.indiatimes.com/definition/moral-hazard2For a detailed analysis of the role risk-based pricing plays in promoting a fair and efficient mortgage market, see Board of Governors of the Federal Reserve System, Report to the Congress on Credit Scoring and Its Effects on the Availability and Affordability of Credit, August 2007, www.federalreserve.gov/boarddocs/rptcongress/creditscore/creditscore.pdf3QM as implemented does not constrain leverage (LTV/CLTV, credit score, or total DTI). It does constrain loan term, but at a highly levered 30 years. The rest of QM is largely window dressing (except for the current 5 year fully indexed requirement on ARMs). For example, FHA has had full doc and fully amortizing loans since its inception. This did not prevent 1 of 8 (3.4 million) of its borrowers going to claim from cohort years 1975-2013. 116
While FHA’s Capital Reached Required 2% Statutory
Level for 1st Time since 2008, It Is Insufficient Mutual Mortgage Insurance Fund at 2.07% in FY 2015 compared to 0.41% in FY 2014.
A further reduction in insurance fee is unjustified and counter productive.
117
• Impact of premium cut was minimal as most of the gain since FY 2014 projection was due to FY 2015 volume gain which was offset dollar for dollar by reductions in mortgage insurance premiums.
• Volume gain was largely due to poaching, mostly from Fannie and RHS, and an improving economy.
• Most of the increase in buying power was capitalized into the purchase of higher priced homes.
• Higher home price projection vs. FY 2014 projection also added to economic value.– Home prices are assumed to continue to increase faster than incomes for foreseeable future.
• 36% of FY 2015 volume in CA, FL and AZ (traditionally volatile states) along with TX (has high house-price risk), up from 28% in FY 2010.
• Premium cut substantially reduced FY2021 projected economic value.
• 2% capital level is insufficient.– FY 2014 report indicated a 4% capital level more appropriate given that U.S. is already
in the 7th year of an economic expansion.
– FHA not projected to hit a 4% single-family forward loan capital level until the end of FY 2020, at which point the current expansion, were it to continue, would be the longest on record.
• FHA’s MRI continues to hover near 25% and is 37% for loans with credit scores < 660.
• Extraordinarily high default rate on loans with scores below 660 is an abusive lending practice.
• These borrowers are disproportionately low-income and minority.
Share of States with Increase in SMRI for Purchase
Loans from Year-Earlier Period*
*Final value for each series based on change in each state from Aug.-Oct. 2015 average to Aug.-Oct. 2016 average. Earlier values calculated analogously.Note: SMRI applies exactly the same stress-test methodology from the NMRI to loans at the state level. Source: AEI Housing Center, www.AEI.org/housing.
Credit easing trend has stopped in majority of states – SMRI down in about two-thirds of states for agency composite. Contrast between composite and individual
agencies has re-appeared as market shares have shifted back to the GSEs after effects of FHA premium cut have worn off.
118
-10%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
-10%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49
Fannie/Freddie
FHA
Percent changes calculated fromyear-earlier three-month average.
VA
Composite
Not updated
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
FHA RHS VA Conventional
CBSA NHMI: Investor Type for Home Purchase Loans
119
FHA has greater presence in lower cost CBSAs, while the Conventional side of the market has greater presence in higher cost CBSAs. FHA, due to its highest risk rating, is
driving up risk in these lower cost CBSAs.
Investor Share by Top 25 CBSA (count), 2017:Q3 - 2018:Q2
Source: AEI Housing Center, www.AEI.org/housing, and First American Data Tree (DataTree.com).
Most volatileCBSAs
Not updated
Riverside/San Bernardino: A Case in Point
50
100
150
200
250
300
350
400
450
50
100
150
200
250
300
350
400
450
Apr-96 Apr-98 Apr-00 Apr-02 Apr-04 Apr-06 Apr-08 Apr-10 Apr-12 Apr-14
Top tier
Bottom tier
Source: AEI Housing Center, www.AEI.org/housing, based on data from Zillow.
Index = 100 in April 1996
Series show nominal house prices
A metro with very volatile house prices, especially for bottom price tier. Since the Jan. 2012 trough, bottom-tier prices up almost 70%, boosted by liberal credit
terms and low rates in a seller’s market.
120
Not updated
House Price Volatility, 51 Largest Metro Areas
Note: Each series shows the percent change from 20 quarters (5 years) earlier. Volatile metros are defined as those for which the difference between the highest and lowest annual percent changes is more than 30 percentage points. All other metros are in the “more stable” group.Source: AEI Housing Center, www.AEI.org/housing, using data from Zillow.
121
House prices most volatile in California and Florida metros, moderately volatile in 16
other metros, with 25 metros having low volatility.
-60%
-40%
-20%
0%
20%
40%
60%
80%
100%
120%
-60%
-40%
-20%
0%
20%
40%
60%
80%
100%
120%
1984 1987 1990 1993 1996 1999 2002 2005 2008 2011
California metros
Florida metros
More stable metros
Other volatilemetros
Not updated
Median Values of Risk Factors by Loan Type
• Greater riskiness of refi loans for a given credit score, total DTI, and CLTV is offset by tighter lending standards. Refis have:
– Higher credit scores
– Lower total DTIs
– Much lower CLTVs
122
Median value, October 2018
Risk Factor Purchase No-Cash-Out Refi Cash-Out Refi
Credit Score 731 733 710
Total DTI 40 38 41
CLTV 95 73 75
Note: Calculations based on loans with non-missing data for credit score, DTI, and CLTV.
Source: AEI Housing Center, www.aei.org/housing.
Risk Shares for Home Purchase Loans
Note. Risk shares pertain to the composite of all purchase loans.Source: AEI Housing Center, www.AEI.org/housing.
Loan risk greater than level conducive to long-run market stability, as low-risk loans accounted for only 37% of volume in October, far from comprising the preponderance of loans, which is necessary for long-term market stability.
123
20%
25%
30%
35%
40%
45%
50%
20%
25%
30%
35%
40%
45%
50%
Sep-12 Feb-13 Jul-13 Dec-13 May-14 Oct-14 Mar-15 Aug-15 Jan-16 Jun-16 Nov-16 Apr-17 Sep-17 Feb-18 Jul-18
Low risk
High risk
Medium risk
Low risk prime defined as stressed default rate of less than 6%, medium risk near prime is 6% to 12%, and high risk subprime is 12% or higher.
For first-time buyers, the October 2018 low-risk prime share was 21%.
DTI Distributions and MRIs, Primary Purchase Loans*
*Data pertain to purchase loans for primary owner-occupied properties. Source: AEI Housing Center, www.AEI.org/housing.
FHA has DTIs as high as 57% and GSEs have some as high as 50%. DTI limits should operate to “take the punch bowl away” before a leverage fueled boom goes too far.
But the current DTIs maximums are so high as to present no such constraint.
124
For FHA, MRIs rise with DTIs – evidence of risk layering. The same is true for the GSEs up to a DTI of 45%; they compensate for risk on only the very highest DTI loans.
0%
2%
4%
6%
8%
<20 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 >57
DTI Distribution, October 2018
FHAGSEs
0%
10%
20%
30%
40%
<20 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 >57
MRI by level of DTI, October 2018 FHA
GSEs
Cash-Out Share and Home Equity
Cash-outs accounted for 71 percent of total refis in October, more than triple the share at start of the series, owing in part to greater home equity. Temporary spike
down early last year was due to a surge in no-cash-outs from FHA premium cut and a drop in mortgage rates. Recent spike is due to a large decline in no-cash-outs from
higher mortgage rates while the demand for cash-outs has remained relatively stable.
125Source: AEI Housing Center, www.AEI.org/housing, and Financial Accounts of the United States.
15%
25%
35%
45%
55%
65%
75%
$0.0
$2.5
$5.0
$7.5
$10.0
$12.5
$15.0
$17.5
$20.0
Sep-12 Feb-13 Jul-13 Dec-13 May-14 Oct-14 Mar-15 Aug-15 Jan-16 Jun-16 Nov-16 Apr-17 Sep-17 Feb-18 Jul-18
Cash-out share of all agency refis(right axis)
Available equity by households and nonprofits in trillions (left axis)
Agency Cash-Out Share and Defaults
As cash-out share has grown, its agency composition has also changed. Compared to the series’ start, VA and FHA have tripled their share by loosening lending
standards faster than the GSEs. Today, they account more over half of the expected defaults, up from just 20%.
126Source: AEI Housing Center, www.AEI.org/housing.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Sep-12 Apr-13 Nov-13 Jun-14 Jan-15 Aug-15 Mar-16 Oct-16 May-17
Market Share
GSE
Ginnie
MRI Oct. 2012 Oct. 2017
Composite 6.3% 13.3%
Fannie Mae 5.2% 9.9%
Freddie Mac 5.5% 9.8%
FHA 19.3% 26.0%
VA 16.1% 21.4%
0
1,000
2,000
3,000
4,000
5,000
6,000
7,000
8,000
9,000
Sep-12 Apr-13 Nov-13 Jun-14 Jan-15 Aug-15 Mar-16 Oct-16 May-17
Expected Defaults under Stress
GSE
Ginnie
Not updated
Nonbank Origination Shares and MRIs by Channel,
Purchase Loans
Source: AEI Housing Center, www.AEI.org/housing. 127
Nonbank’s correspondent share has been increasing at the expense of retail and broker. While the MRIs of all three channels are increasing, the correspondent channel
has the highest MRI and has increased the most.
5%
15%
25%
35%
45%
55%
65%
Feb-13 Dec-13 Oct-14 Aug-15 Jun-16 Apr-17 Feb-18
Nonbank Origination Shares
Correspondent
Broker
Retail
4%
6%
8%
10%
12%
14%
16%
18%
20%
Feb-13 Dec-13 Oct-14 Aug-15 Jun-16 Apr-17 Feb-18
Nonbank Mortgage Risk Indexes
Correspondent
Broker
Retail
Housing Volatility Index
Today’s 21 quarters look to constitute the early part of an extended housing boom.
Sustained periods with few price declines allow market excesses to build and may lead
to a Minsky Moment.** Unsustainable increases in entry-level home prices result in
speculation in land, the more volatile part of the structure/land package.
*Only 30 metros included at beginning of series. This number grows until 1977Q4, when 81 metros are consistently reported.**A Minsky moment is a sudden major collapse of asset values which is part of the credit cycle or business cycle. Such moments occur because long periods of prosperity and increasing value of investments lead to increasing speculation using borrowed money. WikipediaSource: FHFA Quarterly House Price Index and AEI Housing Center
128
Distribution of Negative House Price Change from Four Quarters Earlier in 81 US MSAs*
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
19
76
:Q3
19
77
:Q3
19
78
:Q3
19
79
:Q3
19
80
:Q3
19
81
:Q3
19
82
:Q3
19
83
:Q3
19
84
:Q3
19
85
:Q3
19
86
:Q3
19
87
:Q3
19
88
:Q3
19
89
:Q3
19
90
:Q3
19
91
:Q3
19
92
:Q3
19
93
:Q3
19
94
:Q3
19
95
:Q3
19
96
:Q3
19
97
:Q3
19
98
:Q3
19
99
:Q3
20
00
:Q3
20
01
:Q3
20
02
:Q3
20
03
:Q3
20
04
:Q3
20
05
:Q3
20
06
:Q3
20
07
:Q3
20
08
:Q3
20
09
:Q3
20
10
:Q3
20
11
:Q3
20
12
:Q3
20
13
:Q3
20
14
:Q3
20
15
:Q3
20
16
:Q3
20
17
:Q3
20
18
:Q3
Quiescent period
Correction
21 Qtrs. 10 Qtrs. 15 Qtrs. 39 Qtrs.36 Qtrs.21 Qtrs.(through 2018:Q3)
25 Qtrs.
Unforgiving Home Price Cycles: Booms Fueled by Increasing
Leverage in a Seller’s Market, Followed by Mean Reversion
129
Fueled by growing loan leverage and tight supplies, real home prices have increased 29% since the early 2012 trough. Contrary to prevailing view, post-crisis underwriting/regulatory changes
promote rather than constrain a boom. The pattern is similar to the initial years of the price boom that began in 1998. If it continues, the risk of a serious house price correction increases.
* Calculated as FHFA's all-transaction house price index divided by BEA's price index for personal consumption expenditures.
Note: National Association of Realtors (NAR) defines a seller's market as inventory that is less than or equal to 6 months of sales. NAR data pertain to existing homes; not available
before June 1982. Data from the Census Bureau for new home inventories used before June 1982.
Source: AEI Housing Center, www.AEI.org/housing, FHFA, BEA, Census Bureau, and NAR.
80
100
120
140
160
180
200
220
80
100
120
140
160
180
200
220
1975:Q1 1979:Q2 1983:Q3 1987:Q4 1992:Q1 1996:Q2 2000:Q3 2004:Q4 2009:Q1 2013:Q2 2017:Q3
Real House Price Index (1975:Q1 = 100)*, through 2018:Q3 Predominently a buyer's market Entirely a buyer's market
Entirely a seller's market Predominently a seller's market
GSE affordable housing goals take effect for CY 1993 as mandated by the Housing Enterprises Safety and Soundness Act of 1992
2012 to date: easing loanstandards, very loose Fedpolicy, and historicallylow mortgage rates
1993-2006: period of credit easingand generally falling mortgage rates
Real average annual growth rate 1997:Q2-2003:Q2 -- 4.3%1997:Q2-2006:Q2 -- 5.1%
2012:Q2-2018:Q2 -- 4.2%
Supply-Demand Imbalance Is Greatest in the Low Price Tier
There is also a greater bifurcation on months supply in the market by price point.
From a year ago, the supply-imbalance has improved most at the upper end of the
market, which is approaching a buyer’s market nationally. Inventories remain very
tight at the lower end, continuing the strong seller’s market, which implies that house
prices will continue to increase, thereby worsening affordability.
130Source: AEI Housing Center, www.AEI.org/housing, and Zillow.
0
2
4
6
8
10
12
0
2
4
6
8
10
12
2013:Q1 2013:Q3 2014:Q1 2014:Q3 2015:Q1 2015:Q3 2016:Q1 2016:Q3 2017:Q1 2017:Q3 2018:Q1 2018:Q3
LowLow-MedMed-HighHigh
Grey bars show Q3 for each year.
Months Supply by Price Tier: 73 Metros
Comparing the Supply-Demand Imbalance: 100 Largest Metros
While the supply-demand imbalance has generally improved slightly, it remains tight in
most metros, especially at lower price tiers. For high end homes, 40 of the 100 metros have
buyer’s market conditions, up from 31 a year ago (high tier buyer’s market ≥ 8 months).
131Note: Data are for largest 100 metros using Zillow’s existing home sales. Urban Honolulu in the high price tier is outside of range shown. Source: AEI Housing Center, www.AEI.org/housing, and Zillow.
0
1
2
3
4
5
6
7
8
9
10
0 1 2 3 4 5 6 7 8 9 10
Mo
nth
s Su
pp
ly: 2
01
7:Q
3
Months Supply: 2018:Q3
Low Price Tier
Less supply than one year ago
seller's market
More supply than one year ago
0
1
2
3
4
5
6
7
8
9
10
0 1 2 3 4 5 6 7 8 9 10
Mo
nth
s Su
pp
ly: 2
01
7:Q
3
Months Supply: 2018:Q3
Low-Medium Price Tier
Less supply than one year ago
seller's market
More supply than one year ago
0
1
2
3
4
5
6
7
8
9
10
0 1 2 3 4 5 6 7 8 9 10
Mo
nth
s Su
pp
ly: 2
01
7:Q
3
Months Supply: 2018:Q3
Medium-High Price Tier
Less supply than one year ago
seller's market
More supply than one year ago
0
2
4
6
8
10
12
14
16
18
20
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Mo
nth
s Su
pp
ly: 2
01
7:Q
3
Months Supply: 2018:Q3
High Price Tier (note different axes maxima)
Less supply than one year ago
seller's market
More supply than one year ago
Home Sales, by Metro Market Size
Months’ Supply (2018:Q4) : 271 CBSAs
0.7 6 24+
High Price TierMedian of 271 metros: 15.1 months
Overall*: 8.2 months
Months’ Supply (2018:Q4) : Top 50 CBSAs
0.7 6 24
High Price TierMedian of 50 metros: 9.2 months
Overall*: 7.3 months
132
For high price tier properties, the median months’ supply for all 271 metro areas are far
higher than for the 50 largest metros. This is because the additional 221 metros have
substantially higher month’s inventory of about 16.2 months compared to 9.2 months
just for the 50 metros. On the other hand, when looked at overall, there is little
difference, since the 50 metros account for about 2/3 of the of the overall market
accounted for in the 271 metros.
*The overall months’ supply takes metros on each map as one market, and is calculated by dividing the total number of listings by total sales.Source: Zillow, AEI Housing Center, www.AEI.org/housing.
Tale of two markets: Low end entry level and high end repeat buyers
133
Months’ Supply (2018:Q4) : 271 CBSAs
0.7 6 24+
Low Price TierMedian of 271 CBSAs : 2.3 months
Overall*: 2.1 months
Months’ Supply (2018:Q4) : Top 50 CBSAs
0.7 6 24
Low Price TierMedian of 50 CBSAs: 1.6 months
Overall*: 1.8 months
For low price tier properties, the months’ supply of 221 smaller markets are only
somewhat higher at about 2.5 months, than for the top 50 metros (1.6 months),
resulting in a relatively small difference in the median.
*The overall months’ supply takes metros on each map as one market, and is calculated by dividing the total number of listings by total sales.Source: Zillow, AEI Housing Center, www.AEI.org/housing.
Tale of two markets: Low end entry level and high end repeat buyers
Comparison: Home Sale Transactions (New and Existing)
134
Over the past 2 years, the combined total of the NAR’s EHS and the Census Bureau’s NRS has become more accurate. In Q3, a wider gap has opened up again as the NAR and
Census Bureau have reported larger declines in lending volume. Given that the NAR’s EHS and the Census Bureau’s NRS are based on surveys with large gross ups, which
tend to amplify the errors, the downturn in their data may be overstated.
* Error refers to count and percent difference between the NAR’s Existing Homes Sales (EHS) plus the Census Bureau’s New Residential Sales (NRS) and NHMI. Note: The Census Bureau in its November 2018 release (https://www.census.gov/construction/nrs/index.html) estimated a 90% confidence interval of 5.8%, which equals +/-31,000 sales for all 532,000 sales from January through October 2018. The NAR does not publish a confidence interval with its Existing Home Sales (https://www.nar.realtor/news-releases/2016/12/existing-home-sales-forge-ahead-in-November) but its numbers are based on a sample of about 40 percent of Multiple Listing Services data each month. Sales outside of MLSs are not captured in the monthly series. The NAR states that it “rebenchmarks home sales periodically using other sources to assess overall home sales trends, including sales not reported by MLSs.”Source: AEI Housing Center, www.AEI.org/housing, and First American Data Tree (DataTree.com), the NAR, and Census Bureau.
800,000
1,000,000
1,200,000
1,400,000
1,600,000
1,800,000
2,000,000
800,000
1,000,000
1,200,000
1,400,000
1,600,000
1,800,000
2,000,000
2012:Q4 2013:Q2 2013:Q4 2014:Q2 2014:Q4 2015:Q2 2015:Q4 2016:Q2 2016:Q4 2017:Q2 2017:Q4 2018:Q2
NAR
NHMI
NAR & Census
Grey bars show Q3 for each year.
Period Absolute Error* % Error*
2013:Q4 - 2014:Q3 172,084 3.2%
2014:Q4 - 2015:Q3 135,656 2.4%
2015:Q4 - 2016:Q3 12,814 0.2%
2016:Q4 - 2017:Q3 (35,592) -0.6%
2017:Q4 - 2018:Q3 (239,075) -3.9%
Home Sales, by Metro Market Size
135
In 2018:Q3, home sales started to decrease significantly in the largest 50 markets, held steady in mid-size markets (top 51-100), and expanded in smaller metros/more
rural areas based on market size. This indicates further bifurcation of the market with first-time homebuyers potentially shifting to typically less expensive markets.
* Simple average of FHFA’s CBSA Annual House Price Index for all CBSAs within group.Note: Percentages in the chart refer to the respective CBSA market share by count in the most recent period.Source: AEI Housing Center, www.AEI.org/housing, First American Data Tree (DataTree.com), and FHFA.
NHMI by Metro Group (Count) (Index: 2013:Q3 = 100)
Years in the chart refer to Q3 of the labeled year.
100 100 100 100 100
97
98100
9998
105
108
112
107 107
110
112
116
112 112
110
114
120
116
113
105
108
120
123
113
90
95
100
105
110
115
120
125
Top 25 (41%) Top 26-50 (14%) Top 51-100 (13%) Rest/No CBSA (32%) All (100%)
2013 2014 2015 2016 2017 2018
Mortgage Risk Indices by Lender Type, Purchase Loans
Note: Composite includes credit unions and state housing agencies, which are not shown separately.Source: AEI Housing Center, www.AEI.org/housing. 136
A huge gap has opened up in the riskiness of purchase loans originated by banks and nonbanks. Banks have reduced risk by shifting away from subprime borrowers and
low downpayment loans. Nonbanks have increased share by taking advantage of broad Agency credit boxes and continued easing, thus making them the preferred risk
channel. While this share shift has stabilized, risk levels continue to diverge. The entire year-over-year increase in risk is attributable to nonbanks.
8%
10%
12%
14%
16%
8%
10%
12%
14%
16%
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17 Sep-18
Nonbanks
Composite
Stressed default rate
Nonbank market share, purchase loansNov12 Nov 13 Nov 14 Nov 15 Nov 16 Nov 17 Nov 1829% 41% 50% 55% 55% 56% 64%
Banks
Jumbo portfolio-GSE spreads (in bps)
Portfolio jumbo rate has been below the GSE rate since 2014, reversing prior pattern.
The reasons are an increase in the GSE guarantee fees but also lenders may be bidding
aggressively for jumbo loans to obtain low-risk assets with cross-selling opportunities.
137Source: AEI Housing Center, www.aei.org/housing, and CoreLogic.
-40
-20
0
20
40
60
80
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017
Note 1: Jumbo Portfolio minus GSE and Jumbo PMBS minus GSE spreads (in bps) between 90% and 110% of conforming limit.Note 2: Chart omits PMBS-GSE spreads for years with less than 200 jumbo PMBS loans. Inset box uses loans for all years, except as indicated by line breaks.Note 3: Data for 2017 are for January - September only.Note 4: For loans between 90 percent and 110 percent of the applicable conforming loan limit
GSEs more expensive
GSEs less expensive
<--Financial Crisis--><--Underpricing by GSEs -->
Period Spreads in bps
2001-2006 25
2007-2009t 57
2010-2013t 21
2014-2017tt -26
Homeowners Can’t Count on House Price Gains to Build Wealth
Source: AEI Housing Center, www.aei.org/housing.
A better approach would be to focus on actual wealth building through widespread adaptation of the Wealth-Building Home Loan (WBHL).
138
Using zip-level data for top 100 CBSAs to provide most complete analysis to date of risk by price tier
Zip codesShare of zips with decline in house price index
1990-1995 1995-2000 2000-2005 2005-2010 2010-2015
Top price tier 28% 0% 0% 72% 15%
Middle price tier 37% 1% 0% 83% 30%
Bottom price tier 42% 2% 0% 84% 42%
Note: Top 100 CBSAs are defined by 2010 population. Analysis uses all five-digit zip codes in these CBSAs with a FHFA house price index back to 1990 or earlier and a Zillow median house price in 2000. Zips with a median house price in 2000 in the bottom third, middle third, and top third of all the zips in its CBSA are placed in the bottom, middle, and top price tiers, respectively.
• Prices rose almost everywhere from 1995 to 2005, but many zips saw declines in other periods, especially 2005-2010
• Bottom price tier – where most first-time buyers locate – was worst-performing tier
• These results are for price indices, which average across many homes. Risk for individual homes greater than shown here. WBHL mitigates this risk.
Evaluating the GSEs 2017 Business
Principle: the only plausible reason for government to back the housing market is to
help low- or moderate income families buy homes. An evaluation of the GSEs 2017
business shows, that the GSEs fail to meet this simple test.
139
Refi Cash Out21% share
$300,000 median sales price (SP)
738 median FICO
Refi No Cash Out19% share
$286,000 med. SP
746 med. FICO
2nd home & investor7% share
$229,000 med. SP
774 med. FICO
Almost half of the GSEs’ 2017
volume wasn’t even related to
buying a primary residence.
These borrowers could be served
by the private sector
Source: AEI Housing Center. All share percentages based on dollars (YTD Aug. 2017)
Evaluating the GSEs 2017 Business (cont.)
Another 41% went to help well-to-do buyers, of which 25 percentage points went to well-to-do repeat
buyers of primary residences and 16 percentage points went to well-to-do first-time buyers.
140Source: AEI Housing Center. All share percentages based on dollars (YTD Aug. 2017)
First-time buyer (FTB) w.>85% CLTV & loan>$250,000
8% share
$353,000 med. SP
746 med. FICO
FTB w.<85% CLTV
9% share
$280,000 med. SP
752 FICO
Repeat buyer w. >85% CLTV & loan >$250,000
8% share
$365,000 med. SP
755 FICO
Repeat buyer w. <=85% CLTV
18% share
$327,000 med. SP
774 med. FICO
Unrelated to buying
a primary residence
These buyers
could be served
by the private
sector
Evaluating the GSEs 2017 Business (cont.)
Only 6.5% (1 in 16) GSE Dollars went to first-time buyers of more modest homes and only 3.7% (1 in 30)
GSE Dollars went to repeat buyers of more modest homes.
141Source: AEI Housing Center. All share percentages based on dollars (YTD Aug. 2017)
First-time buyer w. >85% CLTV & loan<=$250,0006.6% share
$168,000 median SP
736 median FICO
Repeat buyer w. >85% CLTV &
loan<=$250,0003.7% share
$189,900 median SP
755 median FICO
The private sector and a targeted and
reformed FHA could replace the GSEs
over time:
• The private sector could handle the 50% who are
not buying a primary residence and the 40%
well-to-do repeat & 1st time buyers of primary
residences
• The remaining 10% could be handled by the
FHA and the private sector
Update: John Burns Intrinsic Home Values
Over the past year the intrinsic over-valuation of the vast majority of metros has increased – the most in the metros that were already highly valued . Almost 75% of metros tracked by John Burns are overvalued today. These overvalued metros are largely concentrated in CA, NV, FL, and AZ, (the Sand States—ground zero in last
boom/bust) and CO, TX, OR, and WA (states that largely sat out the last boom/bust).
142
*Based on HMDA data for 2017.
Note: The Intrinsic Home Value Index shows current price versus intrinsic value assuming 6% mortgage rate. It tracks 131 metros in the U.S.
Source: AEI Housing Center, www.AEI.org/housing, and John Burns Real Estate Consulting.
Hartford, CT, -6%
Reno, NV, 41%
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
-10%
0%
10%
20%
30%
40%
50%
National Intrinsic Home Value Index: +18%
Aug-17
Fairly Valued: 34 Metros.These metros account for 15%
of the overall market*
Over-Valued: 97 MetrosThese metros account for nearly 53%
of the overall market*Aug-18
# of overvalued Metros
AZ, CA, FL, NV 47 (out of 49 metros)
CO, TX, OR, WA 19 (out of 19 metros)
Rest 31 (out of 63 metros)
Not updated
Large banks Large nonbanks
143
NMRI
Higher FHA risk share(relative to market share)
Lower FHA risk share(relative to market share)
Larger circle represents larger market share. Lenders shown represent the 8 largest banks and 15 largest nonbanks by origination share in 2016:Q3.
30% 20% 10% 5% 1%
Wells Fargo
JP Morgan
US Bank
SunTrust
BB&T
Flagstar
Fifth Third
Bank of America
Quicken
Amerihome
Caliber
Franklin American
Fairway
Pennymac
Stearns
Ditech
Nationstar
Guild
United Shore
Freedom Mortgage
Loan Depot
Lakeview
Plaza
GSEs: Large Lender Market Share and Relative Risk Share, Refinance Loans
2013 2014 2015 20162017/
H12017/
Q32017/Oct.
8.6% 9.7% 8.3% 7.8% 8.9% 9.1% 9.1%
25+% -25+%15 to 25% -15 to -25%-5 to -15%5 to 15% 5 to -5%
2013 2014 2015 20162017/
H12017/
Q32017/Oct.
8.6% 9.7% 8.3% 7.8% 8.9% 9.1% 9.1%
Not updated
144
NMRI
Higher GSE risk share(relative to market share)
Lower GSE risk share(relative to market share)
Large banks Large nonbanks
30% 20% 10% 5% 1%Larger circle represents larger market share. Lenders shown represent the largest 8 banks and 15 nonbanks by origination share in 2016:Q3.
Wells Fargo
US Bank
JP Morgan
SunTrust
BB&T
Amerihome
Fifth Third
FHA: Large Issuer Lender Type Market Share and Relative Risk Share, Refinance Loans
FlagstarQuicken
Plaza
Amerihome
United Shore
Stearns
Guild
Fairway
PennyMac
Lakeview
Franklin
Ditech
Freedom Mortgage
LoanDepot
Caliber
Nationstar
2013 2014 2015 20162017/
H12017/
Q32017/Oct.
18.3% 21.3% 21.8% 22.4% 23.6% 24.6% 25.0%
25+% -25+%15 to 25% -15 to -25%-5 to -15%5 to 15% 5 to -5%
2013 2014 2015 20162017/
H12017/
Q32017/Oct.
18.3% 21.3% 21.8% 22.4% 23.6% 24.6% 25.0%
Not updated
145
Even though the rate of increases have slowed over the past three years, volume is still growing from a high base. Compared to 3 years prior, October 2017 volume by
count is up 24 percent; first-time buyer volume is up 30 percent.
Source: AEI Housing Center, www.aei.org/housing.
-10%
0%
10%
20%
30%
40%
50%
-10%
0%
10%
20%
30%
40%
50%
Sep-13 Dec-13 Mar-14 Jun-14 Sep-14 Dec-14 Mar-15 Jun-15 Sep-15 Dec-15 Mar-16 Jun-16 Sep-16 Dec-16 Mar-17 Jun-17 Sep-17
Change from 3 years ago
Year-over-year change
Leverage Fueled Housing Demand Continues to Climb
Not updated
Agency Origination Shares by Risk, Purchase Loans
* We define prime loans as low-risk (with a stressed default rate of less than 6%), near prime as medium risk (with a stressed default rate of 6% to less than 12%), and subprime as high risk (with a stressed default rate of 12% or greater).Source: AEI Housing Center, www.AEI.org/housing.
Fannie and, to a lesser extent, Freddie have expanded their holdings of higher risk subprime loans, now accounting for a combined 27 percent of such loans, up from 7.5 percent in Sept. 2012. While Freddie has offset this by adding safer, prime and near-
prime loans, Fannie has shed some of its business in these categories. Over the past 6 months, Fannie appears to be changing its positioning as talks of GSE reform heat up.
146
0%
10%
20%
30%
40%
50%
60%
70%
80%
Feb-13 Dec-13 Oct-14 Aug-15 Jun-16 Apr-17 Feb-18
Prime
Fannie
Freddie
Purple: RHS
VA
FHA
Feb-13Oct-13Jun-14Feb-15Oct-15Jun-16Feb-17Oct-17Jun-18
Near Prime
Fannie
FHA
Freddie
VA
RHS
0%
10%
20%
30%
40%
50%
60%
70%
80%
Feb-13 Dec-13 Oct-14 Aug-15 Jun-16 Apr-17 Feb-18
Subprime
FHA
Fannie
RHS
Freddie
VA
What Does this Mean for the Broader Market?
147
Due to FHA’s loose lending standards, historically high loan limits and market share,
and an appraisal process focused on market price, not market value, FHA borrowers
are setting the price for a large share of the market including conventional loan buyers.
Law of Marginal Buyer: home prices will keep rising so long as the marginal buyer, who
sets price for all, has access to higher leverage. Historically, the government, has been
the most willing provider of this leverage.
* Source: John Ligon, Heritage Foundation
Borrowing at the Conforming Loan Limit, GSE Purchase Loans
148
Current policy is driving loan balances higher during a very tight market. FHFA
first raised the conforming loan limit from $417,000 to $424,100 in Jan. 2017, then
to $453,100 in Jan. 2018.* Borrowers in non-high cost areas immediately borrowed
at the new maximum. The same holds for high-cost areas (not shown).
*FHA and VA also raised its maximum guaranty amount in line with FHFA and HUD. Data for February 2018 are partial.Source: AEI Housing Center, www.AEI.org/housing.
0%
1%
2%
3%
4%
5%
6%
0%
1%
2%
3%
4%
5%
6%
Jan-16 Apr-16 Jul-16 Oct-16 Jan-17 Apr-17 Jul-17 Oct-17 Jan-18
% of Loans with Loan Amount of $417,000
% of Loans with Loan Amount over $417,000 and at or below $424,100
Share of Loans
% of Loans with Loan Amount over $424,100 and at or below $453,100
Dashed lines mark the change in the conforming loan limit from $417,000 to $424,100 in Jan. 2017 and the change in the conforming loan limit from $424,100 to $453,100 in Jan. 2018.
Not updated
The CFPB’s Qualified Mortgage Policy and GSE QM Patch Allowed for Credit Easing While Supply Is Constrained, a Direct
and Continuing Cause of the Current House Price Boom• In 1.13, “Ability-to-Repay and Qualified Mortgage Standards” rule issued, effective 1.10.14• The Bureau noted it will “protect consumers from irresponsible mortgage lending.”
• The rule effectively set a maximum debt-to-income (DTI) limit of 43% for the private sector.• GSEs and their automated underwriting systems were exempted from this provision for seven years.• Similarly, FHA, the VA and the Department of Agriculture’s Rural Housing Services (RHS) , were
exempted for up to seven years or until these agencies issued their own rules codifying their own lending practices (which all subsequently did).
• The QM rule was pursuant to the Dodd-Frank Act’s calling for minimum mortgage standards• It was to make sure “prime” loans will be made responsibly
• Yet it sets no minimum down payment, no minimum standard for credit worthiness, and no maximum debt-to-income ratio (for government agencies)
• Under this definition of “prime”, a borrower can have no down payment, a credit score of 580, and a debt ratio over 50% as long as they are approved by a government-sanctioned underwriting system.
• That this would promote an unsustainable home price boom could be foreseen:• In 2013: “Booms are fueled by excessive leverage” and “this rule does little to limit borrower leverage
and lays the foundation for the next bust.”*• In 1951: “[In transitioning] from a buyer's to a seller's market, maximum terms become so commonly
used they tend to be considered the minimum.”**
• The QM Patch does not operate counter-cyclically so as to “take the punch bowl away” so as to slow a leverage-fueled price boom.
*Pinto, “CFPB’s new ‘qualified mortgage’ rule: The devil is in the details”, http://www.aei.org/publication/cfpbs-new-qualified-mortgage-rule-the-devil-is-in-the-details/Wallison and Pinto, “New Qualified Mortgage rule setting us up for another meltdown” https://www.washingtontimes.com/news/2013/mar/3/wallison-and-pinto-new-qualified-mortgage-rule-set/**Fisher, Financing Home Ownership, NBER, 1951 149
Number of Investors Flipping Houses Creeping Up
Note: National Association of Realtors (NAR) defines a seller's market as inventory that is less than or equal to 6 months of sales.Source: ATTOM Data Solutions and the NAR.
Due to higher house prices and cash availability house flipping is making a comeback. Levels today are back to the levels seen in 2003.
150
0
10,000
20,000
30,000
40,000
50,000
60,000
70,000
80,000
90,000
100,000
0
10,000
20,000
30,000
40,000
50,000
60,000
70,000
80,000
90,000
100,000
2000:Q1 2002:Q1 2004:Q1 2006:Q1 2008:Q1 2010:Q1 2012:Q1 2014:Q1 2016:Q1
Entirely a buyer's market Entirely a seller's market
Not updated
Average House Price Change by Zip (%, annual avg.)
• Over the past quarter century, the average rate of house price appreciation has been slow and subject to substantial volatility.
• The outcomes for buyers in bottom-tier zips are worse than for buyers in top-tier zips, with lower average appreciation and greater volatility.
• The difference in volatility was especially pronounced during the housing boom (2000- 2005) and bust (2005-2010).
• The reasons: – Perhaps widening of income equality – cyclical swings in mortgage lending standards have a greater impact on FTBs than on RBs
151
Top 100 CBSAs1990-
1995
1995-
2000
2000-
2005
2005-
2010
2010-
2015
1990-
2015
All zips 1.3 4.9 9.3 -2.8 1.9 2.8
By price tier within CBSA
Top-tier zips 1.7 5.3 8.4 -1.7 2.2 3.1
Middle-tier zips 1.4 4.8 9.2 -2.8 1.9 2.8
Bottom-tier zips 0.8 4.6 10.3 -3.9 1.7 2.5
Note: Top 100 CBSAs based on 2010 population. House prices are measured using FHFA’s all-transactions HPI for five-digit zips and
are combined with Zillow’s all single-family residences median house price in 2000 for about 5,300 zips in total. Zip codes are assigned
to tiers based on the median house price in 2000. The price changes in the table are unweighted averages across the included zip
codes. For more, see https://www.aei.org/wp-content/uploads/2017/12/Wealth_Building_WP.pdf.
Source: AEI Housing Center, www.AEI.org/housing, FHFA, and Zillow.
Raising the Conventional Loan Limit – A Prediction
Raising the conforming loan limit during a seller’s market will drive up borrowing and therefore likely increase house prices. A case in point is San Diego, CA.
152
* Through November 2016. Data point for $580,000 bin in 2016 is 315 loans.Note: Data are for 1-unit properties only.Source: AEI Housing Center, www.AEI.org/housing.
0
50
100
150
200
250
$530 $535 $540 $545 $550 $555 $560 $565 $570 $575 $580 $585
# o
f lo
ans Conforming loan limit
in 2013: $546,250
San Diego, CA, MSA: Freddie Loan Distribution, 2013
0
50
100
150
200
250
$530 $535 $540 $545 $550 $555 $560 $565 $570 $575 $580 $585
# o
f lo
ans
Loan Amount Bin (in $1,000 intervals)
Conforming loan limit in 2014: $546,250
San Diego, CA, MSA: Freddie Loan Distribution, 2014
0
50
100
150
200
250
$530 $535 $540 $545 $550 $555 $560 $565 $570 $575 $580 $585
Conforming loan limit in 2015: $562,350
San Diego, CA, MSA: Freddie Loan Distribution, 2015
0
50
100
150
200
250
$530 $535 $540 $545 $550 $555 $560 $565 $570 $575 $580 $585Loan Amount Bin (in $1,000 intervals)
Conforming loan limit in 2016: $580,750
San Diego, CA, MSA: Freddie Loan Distribution, 2016*
Raising the Conventional Loan Limit – A Good Idea?
Raising the conforming loan limit would be of no value to the vast majority of FTBs. Furthermore, conforming loan limit acts as a constraint on house prices during a
seller’s market. Removing it will: 1) drive up borrowing, 2) increase house prices, and 3) shift market share away from the private lenders to the GSEs. Thus the impact on
affordability is largely overstated.
153
Note: Bar chart refers to first-time buyer loans originated from September 2015 – August 2016.Source: AEI Housing Center, www.AEI.org/housing.
Median
$417,000
$625,500
0%
1%
2%
3%
4%
5%
6%
7%
$0 $100,000 $200,000 $300,000 $400,000 $500,000 $600,000
September 2015 – August 2016
Share of Borrowers at or above $417,000
Agency First-time Repeat
GSEs 6.3% 7.6%
Fannie 6.2% 7.5%
Freddie 6.6% 7.7%
FHA 3.4% 5.2%
RHS 0.1% 0.2%
VA 6.2% 12.3%
All 4.8% 7.6%
Cash-Outs and the Economy
Cash-out refis (CO) have grown in share and absolute number. As equity is extracted from real estate, it gets recycled into the economy driving up GDP. Problem: house prices have risen faster than fundamentals can support (with no end in sight). In the long-run, this makes a house price correction likely. However, as equity is extracted
from real estate, owners have less capital to protect themselves from house price declines.
154Source: AEI Housing Center, www.AEI.org/housing and Black Knight.
Back-of the envelop calculation:
According to Black Knight, $31bn in equity was extracted via COs in 2016:Q4.Cash-out extraction was 50% higher y-o-y.
On an annualized rate, this amounts to $120bn in 2016, or a $60bn increase over 2015.
For an $18.5tr economy, this amounts to an extra annualized stimulus of ~0.3% of GDP.0%
10%
20%
30%
40%
50%
60%
-
20,000
40,000
60,000
80,000
100,000
120,000
140,000
Nov-12 Nov-13 Nov-14 Nov-15 Nov-16 Nov-17
Cash-out counts and share
Cash-out Refi Count (left axis)
Cash-out share of refis (right axis)
Cash-Out Share and Expected Defaults
As cash-out share has grown, its agency composition has also changed. Compared to the series start, VA and FHA have tripled their share by loosening lending
standards faster than the GSEs. Today, they account more over half of the expected defaults, up from just 20%.
155Source: AEI Housing Center, www.AEI.org/housing.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Sep-12 May-13 Jan-14 Sep-14 May-15 Jan-16 Sep-16 May-17 Jan-18
Market Share
GSE
Ginnie
0
1,000
2,000
3,000
4,000
5,000
6,000
7,000
8,000
9,000
Sep-12 May-13 Jan-14 Sep-14 May-15 Jan-16 Sep-16 May-17 Jan-18
Expected Defaults under Stress
GSE
Ginnie
Not updated
Punchbowl 1: Mortgage Rate Changes Applicable to
FTBs and RBs
Mortgage rates, due to Fed easing, have been near all time lows. Rates for FTBs and RBs have moved in lock-step with a small premium for RBs over FTBs. The boost
from lower rates has therefore applied equally to both buyer types – as has the increase in rates since November 2016.
156
Note: Data are for GSE primary owner-occupied 30-year fixed-rate purchase mortgages with credit scores of 720-769, CLTVs of 76-80, and DTIs of 39-43.Source: AEI Housing Center, www.AEI.org/housing.
3.50%
3.75%
4.00%
4.25%
4.50%
4.75%
5.00%
3.50%
3.75%
4.00%
4.25%
4.50%
4.75%
5.00%
Feb-13 Aug-13 Feb-14 Aug-14 Feb-15 Aug-15 Feb-16 Aug-16 Feb-17 Aug-17 Feb-18 Aug-18
Average GSE Note Rate: Primary Owner-Occupied 30-yr Fixed-Rate Purchase Mortgages
First-time buyers Repeat buyers
Punchbowl 2: Eased Underwriting Standards
Only Available to Agency First-time Buyers
157
The Agency First-time Buyer MRI (FBMRI) stood at 16.7% in August, up 0.1 ppt from a year earlier and up 2.6 ppts. from 5 Agency RBMRI is virtually unchanged since August 2013 (down 0.1 pyears earlier. The pts.). The Agency FBMRI is 7.5 ppts higher than the Agency RBMRI, 0.5 ppt. wider than the gap a year earlier. If the FBMRI trend continues,
it will reach almost 20% by August 2022.
Note: Calculated for primary owner-occupied home purchase mortgages.Source: AEI Housing Center, www.AEI.org/housing.
8%
12%
16%
20%
24%
28%
32%
36%
8%
12%
16%
20%
24%
28%
32%
36%
Feb-13 Oct-13 Jun-14 Feb-15 Oct-15 Jun-16 Feb-17 Oct-17 Jun-18 Feb-19 Oct-19 Jun-20 Feb-21 Oct-21 Jun-22
Historical Projection
FHA FTB
Agency FTB
Agency RB
Agency Purchase Loan Demand Remains Strong
158
These two punchbowls have largely driven strong growth in agency volume. However, the market has plateaued at its current high level. Agency FTB volume was unchanged compared to one year ago and up 38 percent compared to five years ago. RB volume
has pulled back slightly, but still up 21 percent from five years ago.
Note: First-time buyer volume not available before February 2013. The index is a 12 months rolling index.Source: AEI Housing Center, www.AEI.org/housing.
80
90
100
110
120
130
140
150
160
80
90
100
110
120
130
140
150
160
Jan-14 Jul-14 Jan-15 Jul-15 Jan-16 Jul-16 Jan-17 Jul-17 Jan-18 Jul-18
FTB vs RB Agency Transactions Index: Feb-2013 to Jan-2014 = 100
FTB
RB
0
50,000
100,000
150,000
200,000
250,000
300,000
350,000
0
50,000
100,000
150,000
200,000
250,000
300,000
350,000
Sep-12 Mar-13 Sep-13 Mar-14 Sep-14 Mar-15 Sep-15 Mar-16 Sep-16 Mar-17 Sep-17 Mar-18
FTB to RB median sales price ratioSept-12: 69%Aug-18: 74%
FTB
RB
Market Segmentation:
Median Sales Price for First-time and Repeat Buyers
The housing market is largely segmented by price. FTBs, or entry level buyers, traditionally buy at lower price points than RBs, or move-up buyers. Lately, FTBs have reduced the gap to RBs, an indication that recipients used added buying power from
looser lending to bid up FTB homes, ironically made more expensive by FTB leverage, as RBs have had to make downward quality adjustments.
159
Note: Data are for primary owner occupied properties only. Source: AEI Housing Center, www.AEI.org/housing.
Constant Quality Prices Outlook for the Bifurcated Market:
Slowing Price Appreciation for at the Higher End, Continued
Robust Appreciation for at the Lower End During the recent house price boom, the lower tiers of the market have experienced faster house price appreciation (HPA) due to the interaction of greater availability of leverage and extremely low inventory. The higher tiers have seen more restrained
HPA. Even as the rate punchbowl is further withdrawn, we expect lower tier HPA to be robust as available leverage continues to power prices. Significantly, our research
shows that a concentration of about 30% highly-leveraged borrowers in a census tract can raise prices for everyone in the tract. For the higher tiers, which mostly consist
of RBs, who are less reliant on leverage, the prediction is a slight moderation in HPA.
160Note: HPIs are smoothed around times of FHFA loan limit changes. Date are for 73 largest CBSAs.Source: AEI Housing Center, www.AEI.org/housing.
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
2012:Q4 2013:Q2 2013:Q4 2014:Q2 2014:Q4 2015:Q2 2015:Q4 2016:Q2 2016:Q4 2017:Q2 2017:Q4 2018:Q2
Cumulative Constant-Quality HPI, by Price Tier (2012:Q4 = 0%)
Low
Low-Med
Med-High
High
Outlook for Bifurcation of Market – Quality Changes
So far in the boom, borrowers in the higher price tiers have offset higher constant-quality (CQ) prices by reducing the quality of their home purchases. This quality
adjustment has kept the market transaction price nearly flat. As the rate punchbowl is withdrawn further, we expect additional quality offsets in higher price tiers to
compensate for higher rates. We expect borrowers in lower price tiers to continue to use the leverage punchbowl to largely offset both higher CQ prices and interest rates.
161
Note: HPIs are smoothed around times of FHFA loan limit changes. Date are for 73 largest CBSAs.Source: AEI Housing Center, www.AEI.org/housing.
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%Low Price Tier
Constant-quality
Market Expenditure
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%High Price Tier
Constant-quality
Market Expenditure
Quality offset
Quality offset
Outlook for a Bifurcated Market –
Transaction Prices by State & Changes in Transaction Volume
There is evidence for this market bifurcation at the state level. States with lower median prices, which also tend to be states with higher risk scores, have seen their volume flatten or increase, while higher priced states, which tend to be states with
lower risk scores (and which tend to be mostly featured in the media), have tended to see declines. This trend will likely continue. Yet, there is a second component to this
story as the next slide shows.
162
Source: AEI Housing Center, www.AEI.org/housing.
AK
AL
AR
AZ
CA
CO
CT
DC
DE
FLGAHI
IA
IDIL
IN
KSKY
LA
MA
MD
ME
MI MN
MOMS
MT
NC
ND
NE
NHNJ
NM
NV
NY
OH
OK
OR
PA
RI
SC
SD TN
TX
UT
VA
VT
WA
WI
WVWY
R² = 0.3278
-20%
-15%
-10%
-5%
0%
5%
10%
15%
$100,000 $150,000 $200,000 $250,000 $300,000 $350,000 $400,000 $450,000 $500,000 $550,000 $600,000
Ch
ange
in A
gen
cy P
urc
has
e V
olu
me
fro
m O
ne
Year
ago
Median Agency Transaction Price
June-August 2018
Outlook for a Bifurcated Market – Transaction Prices by State
& Changes in Median Transaction Prices
There is no correlation between the level of state transaction prices and changes in said transaction prices. Transaction prices are still rising in virtually every state, with the few exceptions having more or less flat transaction prices over the last year. This
is further evidence of the bifurcation of the market, where slowing house price appreciation (HPA) at the top is offset by continuing rapid HPA at the bottom.
163
Source: AEI Housing Center, www.AEI.org/housing.
AK
AL
AR
AZ
CA
CO
CT
DC
DE
FLGA
HIIA
ID
IL
IN
KS
KY
LA
MA
MD
MEMI
MN
MO
MS
MT
NC
ND
NENH
NJ
NM
NV
NY
OHOK
OR
PA RI
SCSD
TN
TX
UT
VAVT
WA
WI
WV
WY
R² = 0.0071
-2%
0%
2%
4%
6%
8%
10%
12%
14%
16%
$100,000 $150,000 $200,000 $250,000 $300,000 $350,000 $400,000 $450,000 $500,000 $550,000 $600,000
Ch
ange
in M
edia
n A
gen
cy T
ran
sact
ion
Pri
ce
fro
m O
ne
Year
ago
Median Agency Transaction Price
June-August 2018
While FHA’s Forward Program Capital Is at 3.9%, in
Excess of Statutory Minimum of 2%, It Should be 7%
164
• FHA’s 2014 Annual Report to Congress provides a useful starting point and methodology for evaluating an appropriate level of capital today for FHA’s forward program.
– The 2014 report concludes that a 8.5% capital buffer on outstanding insurance in force is needed– Our analysis adjusts this ratio upward to 9% to account for the growing risk of FHA’s portfolio. – Our analysis also excludes the substantial negative impact of the HECM program on the capital level of the
Mutual Mortgage Insurance Fund.
• Applying the 2014 framework to the 2018 book, we find that FHA had a Capital Resource shortfall of 1.1% or about $13 billion
– At end of FY 2018, FHA had $1.2 trillion in outstanding insurance in force (IIF).– FHA reported at 9.30.18, Capital Resources of 3.9% or $46.8 billion on $1.2 trillion outstanding– Given the 9% assumptions from the 2014 report, this suggests that FHA would need $108 billion in NPV
Claims-Paying Capacity in the next crisis, similar to the Great Recession.• This assumes that steady state MIP income would provide a stream of 4% or $48 billion• Therefore, the Capital Resources portion would need to equal 5.0% or $60 billion, of which only $46.8 billion are currently covered
• Yet, FHA’s 2014 approach does not adequately address the current risks. While the 2014 Report noted “capital [in the form of house price appreciation] disappears in times of stress,” it misses:
– The current home price appreciation (HPA) for entry level homes has again been inflated by excess leverage, most of which has been provided by FHA.
– Entry-level homes’ faster HPA vs the slower HPA of non-highly leveraged higher priced homes– FHA’s sizable market share, its geographic concentration, and how its underwriting policies are exacerbating
the current house price cycle.
• When taking these factors into account, FHA’s Capital Resources shortfall rises to 3.1% or about $37 billion– We think a buffer of an additional 2% in Capital Resources should be provided for each 10% that home
prices in the low price tier have increased faster than prices in the med-high and high price tiers (currently a difference of +16%)
• Today, this would require $24 billion in additional Capital Resources for a total of $84 billion to support $1.2 trillion in IIF• Thus, Capital Resources would need to total 7% or $84 billion (forward program only) compared to the 3.9% or $46.8 billion at
9.30.18 for the forward program
FHA Cash Out Count and MRI
FHA commissioner Brian Montgomery stated that the agency was closely monitoring “the exponential rise in cash-out refinance transactions.” FHA’s CO volume has
tripled from the beginning of the series and risk has increased from 19% to 27%. COs do nothing to promote homeownership for lower-income and minority buyers.
165
Source: AEI Housing Center, www.AEI.org/housing.
0
2,500
5,000
7,500
10,000
12,500
15,000
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17
FHA Cash Out Count
16%
18%
20%
22%
24%
26%
28%
30%
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17
FHA Cash Out MRI
Average Credit Score and DTI: FHA Purchase Loans
FHA commissioner Brian Montgomery also stated that the agency was closely monitoring “a continuing increase in the average FHA-insured borrower’s debt-to-income ratio, and declining average credit scores.” FHA’s average credit score for
purchase loans has dropped from 697 in September 2012 to 671 in August 2018, while it’s average DTI has risen from 40 to 44.1 over the same time period. Lower credit
scores are often combined with higher DTIs, a process known as risk-layering.
166
Source: AEI Housing Center, www.AEI.org/housing.
655
660
665
670
675
680
685
690
695
700
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17
Average Credit Score
37
38
39
40
41
42
43
44
45
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17
Average DTI
Which Risk Factors Have Driven Up the FTB NMRI?
• Since 2012, all the key risk factors have contributed, which has magnified the effect on the NMRI through risk layering.
• Over the past 3 years DTIs have contributed the most to a higher FTB NMRI, with about one-third of FTBs having a DTI in excess of the QM “limit” of 43 percent
167
Share of first-time buyer home purchase loans with
Risk factorAug
2013
Aug
2014
Aug
2015
Aug
2016
Aug
2017
Aug
2018
Credit score < 660 15% 19% 21% 21% 21% 23%
DTI > 43% 24% 24% 27% 27% 31% 38%
CLTV ≥ 95% 64% 67% 71% 71% 71% 71%
30-year term 95% 96% 97% 97% 97% 97%
Risk Layering 27% 30% 34% 35% 37% 41%
Note: Calculated for primary owner-occupied home purchase loans with a government guarantee and reported risk factor.
Risk layering is defined as having at least 3 of the 4 risk features presented in the table above present in a loan.
Source: AEI Housing Center, www.AEI.org/housing.
Share of GSE FTB Purchase Loans w. DTIs of 46-50%
DTI limits should “take the punch bowl away” so as to slow a leverage-fueled price boom. The GSEs had standard DTIs as high as 45%, but traditionally allowed DTIs up to 50% with compensating factors. In this 46-50% DTI range, Freddie has historically outpaced Fannie. Fannie responded in Aug. 2017 by eliminating the requirement for compensating factors. The GSEs’ competition on income leverage, combined with
FHA’s even looser DTI standards, will help fuel the ongoing price boom.
168Source: AEI Housing Center, www.AEI.org/housing.
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
FTB Purchase Loans, by Level of Downpayment
169
Note: Calculated for primary home purchase loans with a government guarantee and reported CLTV. Borrowers with downpayment assistance and CLTVs of 95% or greater are assumed to have no downpayment. Terms and conditions of the downpayment assistance programs vary by program, but in most cases they allow borrowers to offset the downpayment entirely.Source: FHA and AEI Housing Center, www.AEI.org/housing.
It is often reported that down payments of 20% are an impediment to homeownership today. The truth is that VA and RHS don’t require any downpayment at all. And down
payment or closing cost assistance is available through State Housing Finance Agencies, with about 15% of FTBs taking advantage of these programs, which often
lowers their downpayment to $0. FHA purchasers have an average CLTV of 98%. Only 17% of FTBs put down 15% or more.
0%
5%
10%
15%
20%
25%
30%
35%
40%
0%
5%
10%
15%
20%
25%
30%
35%
40%
Feb-13 Aug-13 Feb-14 Aug-14 Feb-15 Aug-15 Feb-16 Aug-16 Feb-17 Aug-17 Feb-18 Aug-18
3-3.5%
No downpayment
15% or more
5%
10%
170
Note: First-time buyer volume not available before February 2013. Source: AEI Housing Center, www.AEI.org/housing.
Agency First-time Buyer Purchase Loan Share
Agency FTB share for August stood at 57.8%, up 0.3 ppt from a year ago. FTB share has likely reached saturation with tight inventory holding back buyers. An expanding
economy and further credit easing will help maintain current levels as they offset higher prices and higher mortgage rates.
53%
54%
55%
56%
57%
58%
59%
60%
61%
53%
54%
55%
56%
57%
58%
59%
60%
61%
Feb-13 Jul-13 Dec-13 May-14 Oct-14 Mar-15 Aug-15 Jan-16 Jun-16 Nov-16 Apr-17 Sep-17 Feb-18 Jul-18
Red markers show August share in each year.
Government Housing Policy Creates an Economics Free Zone
• Law of the Marginal Buyer: In a seller’s market, prices rise faster than
incomes as long as marginal buyer, who sets the price for all, has access
to higher leverage. Determines not only price level, but also degree of
stability, as price is not necessarily equal to value.
• Fisher’s Law: [I]n a seller's market, when choice is restricted and the seller
virtually dictates sales terms, more liberal credit is likely to be capitalized in
price.*
• Law of Ignorance: Policy makers ignore principles of supply, demand, and
housing finance, resulting in an economics free zone. Cross-subsidies and
expanded access to credit push up demand against a regulation-
constrained supply.
* Fisher, Financing Home Ownership, NBER, 1951 (FHA’s first chief
economist)
171
Definition of Low-Risk / Prime Loans
• We define low-risk / prime loans as those with a stressed default rate of less than 6%. Why?
• Low-risk / prime definition calibrated from two sources
– Original QRM proposal to implement Dodd-Frank
– FHA underwriting standards over 1935-55
– Both yield an average stressed default rate of ≈ 3%
• This is consistent with a maximum stressed default rate of ≈ 6% on individual loans, assuming a uniform distribution starting near 0%
• Hence the use of 6% as the highest stressed default rate for a low-risk / prime loan
172
Agency First-Time Buyer Loan Count
Agency FTB volume remained unchanged and up 38 percent compared to one and five years ago, respectively.
173
Note: For primary owner-occupied home purchase mortgages with a government guarantee. November 2017 count is a preliminary estimate.Source: AEI Housing Center, www.AEI.org/housing.
60,000
80,000
100,000
120,000
140,000
160,000
180,000
200,000
60,000
80,000
100,000
120,000
140,000
160,000
180,000
200,000
Feb-13 Jul-13 Dec-13 May-14 Oct-14 Mar-15 Aug-15 Jan-16 Jun-16 Nov-16 Apr-17 Sep-17 Feb-18 Jul-18
Red markers show August count in each year.
Agency Origination Shares, FTB Purchase Loans
Source: AEI Housing Center, www.AEI.org/housing.
FHA FTB origination market share, which jumped after its cut in mortgage insurance premium (MIP) in January 2015, has been gradually trending down over the last two
years. Since then, the GSEs started clawing back some of the market share they had lost. In August, FHA’s FTB share was near its pre-MIP cut level.
174
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
Feb-13 Jul-13 Dec-13 May-14 Oct-14 Mar-15 Aug-15 Jan-16 Jun-16 Nov-16 Apr-17 Sep-17 Feb-18 Jul-18
Freddie
RHS
Fannie
FHA
VA
Agency Origination Shares, FTB Purchase Loans by
Market Segment
* We define prime loans as low-risk (with a stressed default rate of less than 6%), near prime as medium risk (with a stressed default rate of 6% to less than 12%), and subprime as high risk (with a stressed default rate of 12% or greater).Source: AEI Housing Center, www.AEI.org/housing.
GSEs dominate prime segment accounting for 86% of that market. FHA has consolidated most of the subprime segment. Competition is greatest in near-prime segment.
175
0%
10%
20%
30%
40%
50%
60%
70%
80%
Feb-13 Dec-13 Oct-14 Aug-15 Jun-16 Apr-17 Feb-18
Prime
Fannie
Freddie
Purple: RHS
VA
FHA
Feb-13 Dec-13 Oct-14 Aug-15 Jun-16 Apr-17 Feb-18
Near Prime
Fannie
FHA
Freddie
VA
RHS
0%
10%
20%
30%
40%
50%
60%
70%
80%
Feb-13 Dec-13 Oct-14 Aug-15 Jun-16 Apr-17 Feb-18
Subprime
FHA
VA
RHS
Freddie
Fannie
Originations by Market Segment, FTB Purchase Loans
* We define prime loans as low-risk (with a stressed default rate of less than 6%), near prime as medium risk (with a stressed default rate of 6% to less than 12%), and subprime as high risk (with a stressed default rate of 12% or greater).Source: AEI Housing Center, www.AEI.org/housing.
The high-risk subprime market segment continues to outpace the growth in the lower-risk segments. The near-prime segment now accounts for equal number of loans as low-
risk prime segment.
176
0
10,000
20,000
30,000
40,000
50,000
60,000
70,000
80,000
90,000
100,000
0
10,000
20,000
30,000
40,000
50,000
60,000
70,000
80,000
90,000
100,000
Feb-13 Jun-13 Oct-13 Feb-14 Jun-14 Oct-14 Feb-15 Jun-15 Oct-15 Feb-16 Jun-16 Oct-16 Feb-17 Jun-17 Oct-17 Feb-18 Jun-18
Subprime
Prime
Near-prime
Combined First-Time Buyer Mortgage Share Index
Combined first-time buyer share at 54.4% in August, up 0.3 ppt from a year earlier. The NAR’s monthly realtor survey is badly flawed, providing a much noisier picture, and as of recently, perhaps the wrong trend. NAR Sep ‘18 down 3 ppt. from Sep ’17.
177
Note: Calculated as a share of primary owner-occupied home purchase mortgages (both government guaranteed and private-sector mortgages). The NAR’s monthly survey (http://www.realtor.org/reports/realtors-confidence-index) is sent to more than 50,000 realtors (out of a total of 1.3 million members), but has a low response rate; only 7,605 responses were received for the March 2018 survey. Source: AEI Housing Center, www.AEI.org/housing, and the NAR.
25%
27%
29%
31%
33%
35%
46%
48%
50%
52%
54%
56%
58%
60%
Feb-13 Jul-13 Dec-13 May-14 Oct-14 Mar-15 Aug-15 Jan-16 Jun-16 Nov-16 Apr-17 Sep-17 Feb-18 Jul-18
NAR realtor survey (movedahead 2 months), right scale
AEI measure, left scale
Red markers show August share in each year.
The NAR’s first time buyer series is fatally flawed. After
removing seasonality, most of what remains is noise
178
As a result, the NAR series yields little real trend information. AEI’s First-time Buyer Market Share Index (FBMSI) conveys real trend information.
Bottom line: don’t use the NAR survey.
Note: Calculated as a share of primary owner-occupied home purchase mortgages (both government guaranteed and private-sector mortgages). The NAR’s monthly survey (http://www.realtor.org/reports/realtors-confidence-index) is sent to more than 50,000 realtors (out of a total of 1.3 million members), but has a low response rate; only 4,555 responses were received for the April 2018 survey. Source: AEI Housing Center, www.AEI.org/housing, and the NAR.
R² = 0.364
-12%
-10%
-8%
-6%
-4%
-2%
0%
2%
4%
6%
Feb-14 Oct-14 Jun-15 Feb-16 Oct-16 Jun-17 Feb-18
Based on monthly survey with around 2,000responses for closed sales.
Y-o-Y Change in NAR FTB Share
R² = 0.4996
-4%
-2%
0%
2%
4%
6%
Feb-14 Oct-14 Jun-15 Feb-16 Oct-16 Jun-17 Feb-18
Based on census of agency loans: April figure based on 275,000 loans, over 100 times more than the NAR.
Y-o-Y Change in AEI FBMSI
Share of States with Rise in First-time Buyer Loan
Volume and Share from Year-Earlier Period*
*Final value for each series based on change in each state from December 2015-Febraruy 2016 average to December 2016-February 2017 average. Earlier values calculated analogously.Source: AEI Housing Center, www.AEI.org/housing.
First-time buyer loan volume is trending up in vast majority of states. First-time buyer share is also trending up in two-thirds of states due to faster growth than
repeat buyers.
179
20%
30%
40%
50%
60%
70%
80%
90%
100%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Feb '14 -Apr '14
May '14 -Jul '14
Aug '14 -Oct '14
Nov '14 -Jan '15
Feb '15 -Apr '15
May '15 -Jul '15
Aug '15 -Oct '15
Nov '15 -Jan '16
Feb '16 -Apr '16
May '16 -Jul '16
Aug '16 -Oct '16
Nov '16 -Jan '17
Feb '17 -Apr '17
Share
Percent changes calculated fromyear-earlier three-month average.
Volume
NOT UPDATED
180
Profiles of GSE and FHA First-time Buyers with >95% CLTV
The GSEs are primarily expanding their high CLTV business to higher credit score borrowers. These borrowers mainly profited from lower PMI capital requirements which resulted in lower insurance fees. FHA has expanded further down the credit distribution. With high credit scores and relatively low DTIs, GSE risk index (14.8%) is about half of
FHA’s (29.1%).
Source: AEI Housing Center, www.AEI.org/housing.
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
March 2016
August 2018
Share of GSE and FHA FTB Loans with CLTV > 95 by Credit Score Bin
Characteristics of Mortgages Taken Out by First-Time and
Repeat Homebuyers
• The higher risk of the mortgages taken out by first-time buyers is largely due to risk layering.
• Given the combination of little money down and slow amortization, these buyers will have very little home equity for a number of years unless their house appreciates substantially.
• The mortgages taken out by repeat buyers are less risky along two dimensions in particular: – a much smaller share had a CLTV of 95 percent or higher and
– a smaller share had a credit score below 660.
• Bottom line: the supply of mortgage credit to first-time buyers is not tight.
181
Source: AEI Housing Center, www.AEI.org/housing.
August 2018
30-year
Term
CLTV ≥
95%
Credit
Score < 660DTI > 43%
Risk
Layering
First-time Buyers 97% 71% 23% 38% 41%
Repeat Buyers 93% 39% 10% 36% 22%
Note: Calculated for primary owner-occupied home purchase loans with a government guarantee and reported risk factor.
Risk layering is defined as having at least 3 of the 4 risk features presented in the table above present in a loan.
Source: AEI Housing Center, www.AEI.org/housing.
Rising Prices Have Disparate Effects on Buyers
Repeat buyers profiting from higher prices have managed to lower their CLTVs, while FTBs have to stretch further. Recently we are also seeing greater separation in DTIs.
182
Note: Includes all types of NMRI purchase loans (primary owner-occupied, second home, and investor loans).Source: AEI Housing Center, www.AEI.org/housing.
82
84
86
88
90
92
94
96
Feb-13 Oct-13 Jun-14 Feb-15 Oct-15 Jun-16 Feb-17 Oct-17 Jun-18
Chart Title
First-timebuyers
Repeat buyers
Average CLTV
34
35
36
37
38
39
40
Feb-13 Oct-13 Jun-14 Feb-15 Oct-15 Jun-16 Feb-17 Oct-17 Jun-18
Chart Title
First-timebuyers
Repeat buyers
Average DTI
Agency-Specific First-Time Buyer Mortgage
Share Indices
183
Share varies widely across agencies. FHA and RHS are at the high end with a share of around 83 percent, while Freddie Mac is at the low end with a share around 45
percent.
Note: Calculated as a share of primary owner-occupied home purchase mortgages. RHS is Rural Housing Service. FHA share is taken directly from FHA’smonthly production report, due to concerns about the accuracy of the first-time buyer classification in the NMRI dataset. Source: AEI Housing Center, www.AEI.org/housing and FHA.
30%
40%
50%
60%
70%
80%
90%
30%
40%
50%
60%
70%
80%
90%
Feb-13 Jul-13 Dec-13 May-14 Oct-14 Mar-15 Aug-15 Jan-16 Jun-16 Nov-16 Apr-17 Sep-17 Feb-18 Jul-18
VA
Fannie
Freddie
FHA
RHS
DTI Distributions, Agency FTB Purchase Loans*
*Data pertain to all first-time buyer agency purchase loans for primary owner-occupied properties. Source: AEI Housing Center, www.AEI.org/housing.
DTIs have been shifting higher as the rise in house prices has been outpacing income gains. The share of DTIs below 34% has declined, offset by a greater share of DTIs above 40%. While bullish for home prices in the near term, this
presents long-term sustainability problems for both homeowners and the FHA.
184
California shows how the shift could intensify as affordability worsens.
0%
1%
2%
3%
4%
5%
6%
<20 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58
FTB DTI Distribution for Entire U.S.
February 2013
GSE August 2018
FHA August 2018
0%
2%
4%
6%
8%
10%
<20 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 >57
FTB DTI Distribution, August 2018
U.S. excluding CA
FHA CA
GSE California
The Effect of FHA Mortgage Insurance Premium Cut
• FHA’s Jan. 2015 MIP failed to live up to its billing because it was undertaken during a seller’s market. FHA’s recently announced (and since suspended) MIP cut, during an even stronger seller’s market, likely would have had a similar outcome.
• Our research addresses two questions:
• Question 1: How did FHA borrowers use the 6% in additional buying power?
– Have analyzed this question with data from ATTOM Data Solutions, which allowed a robust comparison of FHA and Conventional buyers
– Because prices rose by 3% for FHA financed homes vis-à-vis conventionally financed, borrowers only saved half of the MIP cut
– The other half was capitalized into higher prices:
o The median price paid by ALL FHA borrowers amounted to $1,300 more for the exact same house
o The rest (around $3,600) was used to go up-market, as FHA buyers opted to purchase larger or more expensively appointed homes or opted for more expensive neighborhoods
• Question 2: How accurate was FHA’s prediction that the cut would spur 250,000 first-time buyer (FTB) home purchases over the coming 3 years (≈ 83,000/year)?
– FHA’s first-time buyer volume increased about 180,000 in 1st year after MIP cut. Using the NMRI data, we estimate that roughly:
o 35,000 (20%) went to new entrant FTB brought in by the MIP cut, only 42% of projection
o 85,000 of these loans (nearly half) were poached from the other Agencies
o 60,000 (33%) represented market trend growth unrelated to the MIP cut
– Upshot: FHA fell far short of goal despite big rise in [largely poached] total FTB volume185
186
Over the last 6 years FHA-insured home buyers have seen a growth in housing debt that has greatly outpaced income growth. Lower income borrowers have had the
largest increase in debt burden relative to incomes (+14 ppts. differential).
Income and Debt Growth by Income Group: FHA
Purchase Loans
Note: Data are for largest 73 metros.
Source: AEI Housing Center, www.AEI.org/housing.
18%
11%
4%
32%
22%
14%
0%
5%
10%
15%
20%
25%
30%
35%
40%
25th Percentile 50th Percentile / Median 75th PercentileHousehold Income
Income Growth
Housing Debt Payment Growth
Cumulative change, 2013:Q1 – 2017:Q4
187
House Price Appreciation (HPA) by Price Tier
House price appreciation (HPA) has been greater for entry-level homes (low and low-medium price tiers) than for move-up homes (medium-high and high price tiers). Since October 2012, house prices in the low tier have risen 49% but only 29% in the high tier.
This trend of divergent growth rates is continuing and even accelerating.
Source: AEI Housing Center www.AEI.org/housing.
0%
10%
20%
30%
40%
50%
0%
10%
20%
30%
40%
50%
Oct-12 Apr-13 Oct-13 Apr-14 Oct-14 Apr-15 Oct-15 Apr-16 Oct-16 Apr-17 Oct-17 Apr-18 Oct-18
Low
Low-Med
Med-High
High
Cumulative HPA (Oct. 2012 - Jan. 2019): by Price Tier
House Price Appreciation (HPA) by Price Tier: 73 Metros
188
HPA remained strongly bifurcated. In January 2019, house prices in the low and low-medium
price tiers appreciated at a faster pace than in December, with the biggest rebound in prices
coming in the low price tier, where access to credit is most prevalent. Compared to a year ago,
house prices in the low price tier appreciated 5.3% and 4.0% in the low-medium tier, while
house prices in the medium-high tier appreciated 3.1% and only 1.5% in the high tier.
Note: Data for October 2018 to January 2019 are preliminary. Price tiers are set at the metro level and are defined as follows: Low: all sales at or below the 40th
percentile of FHA sales prices; Low-Medium: all sales at or below the 80th percentile of FHA sales prices; Medium-High: all sales at or below the 125% of the GSE loan
limit; and High: Rest. HPAs are smoothed around the times of FHFA loan limit changes.
Source: AEI Housing Center, www.AEI.org/housing.
Red markers show November HPA in each year.
5.7%
5.3%
0%
2%
4%
6%
8%
10%
12%
Oct
-13
Oct
-14
Oct
-15
Oct
-16
Oct
-17
Oct
-18
Low
5.2%
4.0%
0%
2%
4%
6%
8%
10%
12%
Oct
-13
Oct
-14
Oct
-15
Oct
-16
Oct
-17
Oct
-18
Low-Med
5.0%
3.1%
0%
2%
4%
6%
8%
10%
12%
Oct
-13
Oct
-14
Oct
-15
Oct
-16
Oct
-17
Oct
-18
Med-High
3.7%
1.5%
0%
2%
4%
6%
8%
10%
12%
Oct
-13
Oct
-14
Oct
-15
Oct
-16
Oct
-17
Oct
-18
High
-40%
-32%
-24%
-16%
-8%
0%
8%
16%
24%0
2
4
6
8
10
12
14
16
Jan-99 Jan-01 Jan-03 Jan-05 Jan-07 Jan-09 Jan-11 Jan-13 Jan-15 Jan-17 Jan-19
Y-o
-Y P
rice
Ch
ange
(sc
ale
inve
rte
d)
Mo
nth
's S
up
ply
Months Supply* (left axis)
FHFA House Price Index, smoothed**(right axis)
Buyer's Market, Prices Falling
Seller's Market, Prices Rising
Dashed line: Price equilibrium point
Supply-Demand Imbalance in the Market Is Driving Prices Up
The supply-demand imbalance persists. The NAR’s not-seasonally adjusted months (mo.) inventory in January, which is traditionally the month with the greatest inventory and
lowest sales, stood at 5.6 mo., up 0.7 mo. from a year ago. While this metric has started to increase over the past 5 mo., it is still averaging below 6 mo., the demarcation between a buyer’s and seller’s market, and it will fall back with the beginning of the spring buying season. Thus, it is too soon to project a return of a buyer’s market. Instead, we expect
the seller’s market to modestly strengthen. This means further credit easing will continue to be capitalized into higher home prices. According to the FHFA, not-seasonally adjusted home prices rose 5.8% in November year-over-year, down from 6.8% a year ago.
The chart below shows the strong inverse relationship between supply and prices.
189
* National Association of Realtors (NAR) “Number of homes available for sale (NSA) divided by NAR’s “Existing Homes Sales (NSA)”. The NAR defines a seller’s market to exist when
the inventory of existing homes for sale would be exhausted in six months or less at the current sales pace. Conversely, a buyer’s market exists when the inventory of existing homes
for sale exceeds six months at the current sales pace. (http://www.realtor.org/news-releases/2013/04/march-existing-home-sales-slip-due-to-limited-inventory-prices-maintain-uptrend).
** FHFA Monthly Purchase-Only Not Seasonally Adjusted house price index. The series is a 6 month trailing average.
Source: National Association of Realtors, FHFA, and AEI Housing Center, www.AEI.org/housing.
National Month’s Inventory & Changes in Nominal House Prices*
Length of current seller’s market: 77 months
Affordability Worsens in a Seller’s Market
Nominal Price-to-Income Ratio* has retraced 53% of the drop from the 2006 peak to the 2012 trough. Combination of a continued highly accommodative monetary policy and
easier lending promotes further capital flows into real estate, increasing the potential for economic damage as highly leveraged lending fuels a cyclically volatile housing sector.
190
* Calculated as median house price divided by median household income.
Note: The National Association of Realtors (NAR) defines a seller’s market as inventory that is less than or equal to 6 months of sales. NAR data pertain to existing homes; not
available before June 1982. Data from the Census Bureau for new home inventories used before June 1982.
Source: AEI Housing Center, www.AEI.org/housing, Zillow, Census Bureau, and the NAR.
2.2
2.4
2.6
2.8
3.0
3.2
3.4
3.6
3.8
4.0
4.2
4.4
2.2
2.4
2.6
2.8
3.0
3.2
3.4
3.6
3.8
4.0
4.2
4.4
1975:Q1 1979:Q2 1983:Q3 1987:Q4 1992:Q1 1996:Q2 2000:Q3 2004:Q4 2009:Q1 2013:Q2 2017:Q3
Nominal Price-to-Income Ratio, through 2018:Q4* Predominantly a buyer's market Entirely a buyer's market
Entirely a seller's market Predominantly a seller's market
GSE affordable housing goals take effect for CY 1993 as mandated by the Housing Enterprises Safety and Soundness Act of 1992
2012 to date: easing loanstandards, very loose Fedpolicy, and historicallylow mortgage rates
* Calculated as median house price divided by median household income.Source: Zillow.
1993-2006: period of credit easingand generally falling mortgage rates
Fannie vs. Freddie Risk Index, GSE Purchase Loans
Source: AEI Housing Center, www.AEI.org/housing.
Fannie Mae’s risk index continues to outpace Freddie Mac’s. Fannie’s purchase MRI in Nov. 2018 was 1.5 ppts (or 22%) higher than Freddie’s. Interestingly, this risk pick-up has not translated into any meaningful market share gains for Fannie. While its share
has been volatile, it has averaged around 58-61% for each November since 2013.
191
4%
5%
6%
7%
8%
9%
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17 Sep-18
Mortgage Risk Index
Stressed default rate
Freddie
Fannie
Rough contribution to higher MRIover last 2 years from higher…
Credit Scores CLTVs DTIs
Freddie -24% 36% 88%
Fannie 20% 56% 24%
50%
55%
60%
65%
70%
75%
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17 Sep-18
Fannie Share of GSEs
GSE Market Share
History Repeats Itself: the “Quiet” Battle for Subprime (High Risk
>95 CLTV Purchase Loans) among Fannie, Freddie & FHA
Note: Data are for high risk primary owner-occupied home purchase loans with a CLTV > 95%. High risk are loans with a stressed default rate of 12% or greater.
Source: AEI Housing Center, www.AEI.org/housing.
As this segment of the GSE’s business has grown, Fannie has been holding the clear advantage over Freddie with a market share of 75-95% (Fannie’s share of all GSE
purchase business is currently less than 60%). Interestingly, with Fannie in the driver seat, it has been charging higher loan rates on a risk-adjusted basis. As the prior slide shows, Fannie is able to poach these loans from FHA, since FHA does not price for risk.
192
-10
-5
0
5
10
15
20
25
30
35
Sep
-12
Mar
-13
Sep
-13
Mar
-14
Sep
-14
Mar
-15
Sep
-15
Mar
-16
Sep
-16
Mar
-17
Sep
-17
Mar
-18
Sep
-18
Fannie-Freddie risk-adjusted note rate spread (in bps.)
Fannie w/ higher rate
Fannie w/ lower rate
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0
2,000
4,000
6,000
8,000
10,000
12,000
14,000
16,000
18,000
20,000
Sep
-12
Mar
-13
Sep
-13
Mar
-14
Sep
-14
Mar
-15
Sep
-15
Mar
-16
Sep
-16
Mar
-17
Sep
-17
Mar
-18
Sep
-18
High Risk GSE Purchase loans with a CLTV > 95%
# of Loans(left axis)Fannie Share(right axis)Fannie Share(all purchase, right axis)
Leverage Fueled Housing Demand Pauses Due to Higher Rates
193
While still being up 25 percent from 5 years ago, purchase volume in November 2018 declined 5.1 percent from a year earlier. First-time buyer volume was down 3.6 percent,
while repeat buyer volume was down 7.0%. Greater access to credit is allowing first-time buyers to offset higher mortgage rates and higher house prices, while repeat
buyers, with less access to credit, are electing to drop out of the market in larger numbers.
Note: October 2018 count is a preliminary estimate. First-time buyer volume not available before February 2013.
Source: AEI Housing Center, www.AEI.org/housing.
60,000
110,000
160,000
210,000
260,000
310,000
360,000
60,000
110,000
160,000
210,000
260,000
310,000
360,000
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17 Sep-18
Red markers show November count in each year.Composite
First-timebuyers
RepeatBuyers
Purchase loans
Agency First-time and Repeat Buyer Mortgage Risk Indices
194
The November Agency FBMRI stood at 17.0%, up 0.6 ppt from a year earlier. It is also 7.4 ppts. higher than the mortgage risk index for repeat buyers, which is 0.4
ppt. wider than the gap a year earlier. Given supply constraints and absent a triggering event, we expect house prices and leverage to continue to rise for FTBs.
Note: Calculated for primary owner-occupied home purchase mortgages.Source: AEI Housing Center, www.aei.org/housing.
7%
9%
11%
13%
15%
17%
19%
21%
7%
9%
11%
13%
15%
17%
19%
21%
Feb-13 Feb-14 Feb-15 Feb-16 Feb-17 Feb-18 Feb-19 Feb-20 Feb-21 Feb-22
First-time buyer (FBMRI)
Repeat buyer MRI
Historical Projection
Origination Shares and MRIs by Seller Lender Type,
GSE Purchase Loans
Note: Data for most recent months may understate large-bank share by perhaps 2 percentage points, as large banks are slower to move recent originations to the guarantee agencies for securitization and our market shares are based on securitized loans. MRIs for credit unions and state housing agencies are not shown because of low loan volumes.*Origination shares do not show shares for State Housing Finance Agencies or Credit Unions which account for about 5% of the GSE Purchase market.Source: AEI Housing Center, www.AEI.org/housing.
195
The shift in GSE market share from large banks to nonbanks appears to be resuming. The large-bank share has dropped to a series’ low of around 25% in November 2018, down from 34% a year ago. Other banks are also losing share.
Banks (both large and other) have a lower GSE risk profile than nonbanks.
0%
10%
20%
30%
40%
50%
60%
70%
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17 Sep-18
Origination Shares*
Other banks
Large banks
Nonbanks
4.5%
5.0%
5.5%
6.0%
6.5%
7.0%
7.5%
8.0%
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17 Sep-18
Mortgage Risk Indexes
Large banks
Nonbanks
Other banks
Composite shown by blue line
Origination Shares and MRIs by Issuer Lender Type,
FHA Purchase Loans
196
The dramatic market shift from large banks to nonbanks for FHA loans appears to be continuing. In November, the large bank share dropped below 10%, a new series’
low. Migration to nonbanks has boosted overall risk levels, as nonbanks are willing to originate riskier FHA loans than large banks.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17 Sep-18
Origination Shares*
Large banks
Nonbanks
Other banks
18%
20%
22%
24%
26%
28%
30%
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17 Sep-18
Mortgage Risk Indexes
Large banks
Nonbanks
Other banks
Composite shown by blue line
Note: Data for most recent months may understate large-bank share by perhaps 2 percentage points, as large banks are slower to move recent originations to the guarantee agencies for securitization and our market shares are based on securitized loans. MRIs for credit unions and state housing agencies are not shown because of low loan volumes.*Origination shares do not show shares for State Housing Finance Agencies and Credit Unions which account for about 4% of the FHA Purchase market.Source: AEI Housing Center, www.AEI.org/housing.
Mortgage Risk Indices by Lender Type, Refi Loans
Note: Composite includes credit unions and state housing agencies, which are not shown separately.Source: AEI Housing Center, www.AEI.org/housing. 197
The same share shift from banks to nonbanks applies to agency refi loans. The large bank share dropped below 20% for the first time in the series. Similar to purchase
loans, the gap in riskiness between banks and nonbanks has also widened over time.
0%
10%
20%
30%
40%
50%
60%
70%
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17 Sep-18
Refi Origination Shares
Other banks
Large banks
Nonbanks
4%
6%
8%
10%
12%
14%
16%
18%
20%
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17 Sep-18
Refi Mortgage Risk Index
Other banks
Large banks
Nonbanks
Composite
FHA’s March 2018 Action to Address Excessive Loan Risk Is a Positive Step, but Too Early to Tell How Significant
We’ll be looking to confirm three positive outcomes, note however rates are down, wages and jobs are up, and these will spur demand:• Entry-level homes may be more affordable.
• FHA's excessively high level of loan leverage (example: DTIs >50% leads to increased entry level homes house price appreciation due to the long-standing sellers' market. This change will slow the rate of house price appreciation (all things equal)
• But impossible to say by how much, as definitive data until mid-June or mid-July
• A small reduction in imbalance between supply and demand*• Nationally an extremely strong sellers' market continues in the entry-level price range comprising
57% of sales. • 96 of the 100 largest metros are experiencing a moderate to extreme sellers’ market for the
low price tier (28% of sales); • 98 of the 100 largest metros for the low-med tier (29% of sales).
• This high demand (and no change in supply) will result in about the same number of homes selling as before (all things being equal)
• Those who remain as renters will avoid jumping into a 7 year old boom market.• We do not know when the boom will end, just that it will.• High risk borrowers entering the market late in the cycle tend to get foreclosed on at high rates.
The FHA loans these renters avoid have an extremely high risk of default under stress: >40%.With the recent rate drop, this step may prove to little to avoid unsustainable entry-level price increases.*Data are for 2018:Q4. Months' inventory is for 912 counties (weighted by volume) and represents well over 90% of all sales. Share by price range is for the entire US. Source AEI Housing Center and Zillow
Credit Easing = Punchbowl Spiking Continues, Led by FHA
Note: Includes all types of NMRI purchase loans (primary owner-occupied, second home, and investor loans).
Source: AEI Housing Center, www.AEI.org/housing.
The Composite NMRI for purchase loans increased from already elevated levels a year ago. For FHA, the index is rising at a rate of 1.5% year-over-year. While this rate has
slowed, it is coming off very high levels of increase. First-time buyers have consistently been taking on greater leverage and default risk, continues to fuel accelerating house price growth for entry-level homes. Higher default risk combined with unsustainable
home price increases will lead to unnecessarily high default rates during the eventual market correction.
199
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
Sep-13 Jan-14 May-14 Sep-14 Jan-15 May-15 Sep-15 Jan-16 May-16 Sep-16 Jan-17 May-17 Sep-17 Jan-18 May-18 Sep-18 Jan-19
Repeat buyersComposite
First-time buyers
Change from 12 months earlier, in percentage points
Easing
Tightening
FHA
Supply-Demand Imbalance in the Market Is Driving Prices Up
Given the strong relationship between the level of supply and price movements, today’s housing market has too much highly leveraged demand chasing too little supply.
According to the NAR, monthly inventory for February was at 3.9 months, 0.4 months higher than last year’s reading. While house price appreciation (HPA) at 5.3% has slowed from a year ago, it is still much higher than the rates of wage and inflation
growth. Given the previously noted drop in rates, we expect supply to tighten and HPA rate to increase.
200
*Month’s supply updated through March 2019; FHFA House Price Index updated through February 2019.
** The NAR defines a seller’s market to exist when the inventory of existing homes for sale would be exhausted in six months or less at the current sales pace. Conversely, a
buyer’s market exists when the inventory of existing homes for sale exceeds six months at the current sales pace. (http://www.realtor.org/news-releases/2013/04/march-
existing-home-sales-slip-due-to-limited-inventory-prices-maintain-uptrend).
*** FHFA Monthly Purchase-Only Seasonally Adjusted house price index. The series is a 6 month trailing average.
Source: National Association of Realtors, FHFA, and AEI Housing Center, www.AEI.org/housing.
National Month’s Inventory & Changes in Nominal House Prices*
-18.0%
-12.0%
-6.0%
0.0%
6.0%
12.0%2
4
6
8
10
12
Jan-99 Jan-01 Jan-03 Jan-05 Jan-07 Jan-09 Jan-11 Jan-13 Jan-15 Jan-17 Jan-19
YoY
Pri
ce C
han
ge (
scal
e in
vert
ed)
Mo
nth
's S
up
ply
Month's Supply*
FHFA Price Index,
Buyer's Market, Prices Falling
Seller's Market, Prices Rising
Dashed line: Price equilibrium point
Cash Out Refi Volume by Agency
Agency cash out volume (by count) for January 2019 was down 33% from January 2017. Cash-Out refi volume has declined sharply as rates have risen (albeit
declined less than no-cash out refis.) The decline has been particularly sharp for the GSEs, with FHA and VA having more stable volumes.
201
Note: Data are for Agency loan market only.Source: AEI Housing Center, www.AEI.org/housing.
0
20,000
40,000
60,000
80,000
100,000
120,000
140,000
0
20,000
40,000
60,000
80,000
100,000
120,000
140,000
Sep-12 Mar-13 Sep-13 Mar-14 Sep-14 Mar-15 Sep-15 Mar-16 Sep-16 Mar-17 Sep-17 Mar-18 Sep-18
FHA VA Freddie Fannie
Red markers show January counts in each year.
Compositional Change of Cash-Out Refis
Note: Includes all types of NMRI purchase loans (primary owner-occupied, second home, and investor loans).
Source: AEI Housing Center, www.AEI.org/housing.
A closer look at the risk distribution reveals that lower risk borrowers, mostly served by the GSEs, have been exiting the market as rates have risen, while higher risk borrowers – mostly
served by FHA and VA - have been less sensitive to rate changes.
202
0%
10%
20%
30%
40%
50%
0 4 8 12 16 20 24 28 32
Mortgage Risk Index
January 2019 DistributionFannie Freddie FHA VA
Percent of loans
-20%
-15%
-10%
-5%
0%
5%
10%
15%
0 4 8 12 16 20 24 28 32Mortgage Risk Index
Change in Distribution, January 2017 to January 2019
Fannie Freddie FHA VA
Percentage points
Unforgiving Home Price Cycles: Booms Fueled by Increasing
Leverage in a Seller’s Market, Followed by Mean Reversion
203
Fueled by growing loan leverage and tight supplies, real home prices have increased 29% since the early 2012 trough. Contrary to prevailing view, post-crisis underwriting/regulatory changes
promote rather than constrain a boom. The pattern is similar to the initial years of the price boom that began in 1998. If it continues, the risk of a serious house price correction increases.
* Calculated as FHFA's all-transaction house price index divided by BEA's price index for personal consumption expenditures.
Note: National Association of Realtors (NAR) defines a seller's market as inventory that is less than or equal to 6 months of sales. NAR data pertain to existing homes; not available
before June 1982. Data from the Census Bureau for new home inventories used before June 1982.
Source: AEI Housing Center, www.AEI.org/housing, FHFA, BEA, Census Bureau, and NAR.
80
100
120
140
160
180
200
220
80
100
120
140
160
180
200
220
1975:Q1 1979:Q2 1983:Q3 1987:Q4 1992:Q1 1996:Q2 2000:Q3 2004:Q4 2009:Q1 2013:Q2 2017:Q3
Real House Price Index (1975:Q1 = 100)*, through 2018:Q4 Predominantly a buyer's market Entirely a buyer's market
Entirely a seller's market Predominantly a seller's market
GSE affordable housing goals take effect for CY 1993 as mandated by the Housing Enterprises Safety and Soundness Act of 1992
2012 to date: easing loanstandards, very loose Fedpolicy, and historicallylow mortgage rates
1993-2006: period of credit easingand generally falling mortgage rates
Real average annual growth rate 1997:Q2-2003:Q2 -- 4.3%1997:Q2-2006:Q2 -- 5.1%
2012:Q2-2018:Q2 -- 4.2%
Housing Volatility Index
204
Today’s 23 quarters constitute an extended housing boom, only exceeded by the last
boom. Sustained periods with few price declines allow market excesses to build and
may lead to a Minsky Moment.** Unsustainable increases in entry-level home prices
result in speculation in land, the more volatile part of the structure/land package.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
19
76
:Q3
19
77
:Q3
19
78
:Q3
19
79
:Q3
19
80
:Q3
19
81
:Q3
19
82
:Q3
19
83
:Q3
19
84
:Q3
19
85
:Q3
19
86
:Q3
19
87
:Q3
19
88
:Q3
19
89
:Q3
19
90
:Q3
19
91
:Q3
19
92
:Q3
19
93
:Q3
19
94
:Q3
19
95
:Q3
19
96
:Q3
19
97
:Q3
19
98
:Q3
19
99
:Q3
20
00
:Q3
20
01
:Q3
20
02
:Q3
20
03
:Q3
20
04
:Q3
20
05
:Q3
20
06
:Q3
20
07
:Q3
20
08
:Q3
20
09
:Q3
20
10
:Q3
20
11
:Q3
20
12
:Q3
20
13
:Q3
20
14
:Q3
20
15
:Q3
20
16
:Q3
20
17
:Q3
20
18
:Q3
Quiescent period
Correction
21 Qtrs. 10 Qtrs. 15 Qtrs. 39 Qtrs.36 Qtrs.23 Qtrs.(through 2018:Q4)
25 Qtrs.
Distribution of Negative House Price Change from Four Quarters Earlier in 81 US MSAs*
*Only 30 metros included at beginning of series. This number grows until 1977:Q4, when 81 metros are consistently reported.
**A Minsky moment occurs when a market collapses following a prolonged period of growth due to speculation and borrowing.
Source: FHFA Quarterly House Price Index and AEI Housing Center
Supply-Demand Imbalance Is Greatest in the Low Price Tier
There is a growing bifurcation on months’ supply in the market by entry-level (low and
low-med) vs. move-up (med-high and high). From a year ago, the supply-imbalance
has improved at all price points, but most at the upper end of the market. Inventories
remain historically tight at the lower end, continuing the strong seller’s market, which
implies that house prices will continue to increase, thereby worsening affordability.
205Note: Data are for 918 counties representing approximately 90% of sales.Source: AEI Housing Center, www.AEI.org/housing, and Zillow.
0
1
2
3
4
5
6
7
8
9
10
0
1
2
3
4
5
6
7
8
9
10
2013:Q1 2013:Q3 2014:Q1 2014:Q3 2015:Q1 2015:Q3 2016:Q1 2016:Q3 2017:Q1 2017:Q3 2018:Q1 2018:Q3
LowLow-MedMed-HighHigh
Grey bars show Q4 for each year.
Months’ Supply by Price Tier
Comparing the Supply-Demand Imbalance: 100 Largest Metros
While supply-demand imbalance has improved slightly, it remains tight in most metros,
especially at lower price tiers. For low and low-med tiers, 96% and 98% of metros are
sellers’ markets. For high tier, 69% are buyers’ markets, up from 48% a year ago.*
206* The demarcation point between buyer’s and seller’s market likely varies by price point. We have estimated them at 5, 6, 7, and 8 months respectively.
Note: Data for largest 100 metros using Zillow’s existing home sales. Source: AEI Housing Center, www.AEI.org/housing, and Zillow.
0
1
2
3
4
5
6
7
8
9
10
0 1 2 3 4 5 6 7 8 9 10
Mo
nth
s Su
pp
ly: 2
01
7:Q
4
Months Supply: 2018:Q4
Low Price Tier
Less supply than one year ago
seller's market
More supply than one year ago
Two out of range
0
1
2
3
4
5
6
7
8
9
10
0 1 2 3 4 5 6 7 8 9 10
Mo
nth
s Su
pp
ly: 2
01
7:Q
4
Months Supply: 2018:Q4
Medium-High Price Tier
Less supply than one year ago
seller's market
More supply than one year ago
0
1
2
3
4
5
6
7
8
9
10
0 1 2 3 4 5 6 7 8 9 10
Mo
nth
s Su
pp
ly: 2
01
7:Q
4
Months Supply: 2018:Q4
Low-Medium Price Tier
Less supply than one year ago
seller's market
More supply than one year ago
Two out of range
0
2
4
6
8
10
12
14
16
18
20
0 2 4 6 8 10 12 14 16 18 20
Mo
nth
s Su
pp
ly: 2
01
7:Q
4
Months Supply: 2018:Q4
High Price Tier (note different axes maxima)
Less supply than one year ago
seller's market
More supply than one year ago
Two out of range
Thirteen out of range
Months’ Supply (2018:Q4) : 271 CBSAs
0.7 6 24+
Low Price TierMedian of 271 metros : 2.3 months
Overall*: 2.1 months
High Price TierMedian of 271 metros: 15.1 months
Overall*: 8.2 months
* Both maps show 271 metros. The overall months’ supply takes 271 metros as one market, and is calculated by dividing the total number oflistings by total sales.Source: Zillow, AEI Housing Center, www.AEI.org/housing. 207
Months’ Supply: A Tale of Two Markets
Levels of inventories remained strongly bifurcated. In the low price tier, 2.1 months’
supply on average, while it was 8.2 months in the high price tier. This relationship
holds pretty much across the country as the maps show.
Months’ Supply (2018:Q4) : Top 50 CBSAs
0.7 6 24
Low Price TierMedian of 50 metros: 1.6 months
Overall*: 1.8 months
High Price TierMedian of 50 metros: 9.2 months
Overall*: 7.3 months
* Both maps show the top 50 metros, ranked by 2017 HMDA loan originations. The overall months’ supply takes 50 metros as one market, and is calculated by dividing thetotal number of listings by total sales.Source: Zillow, AEI Housing Center, www.AEI.org/housing.
208
Months’ Supply: A Tale of Two Markets (Cont.)
Levels of inventories are even tighter for the largest metros, 1.8 months vs. 7.3 months
on average overall.
House Price Appreciation (2012:Q4 vs. 2018: Q4): Top 50 CBSAs
10% 25% 74%
Low Price TierMedian House Price Appreciation: 43%
High Price TierMedian House Price Appreciation: 19%
209
Note: Both maps show the top 50 metros, ranked by 2017 HMDA loan originations.Source: Zillow, AEI Housing Center, www.AEI.org/housing.
House Price Appreciation: A Tale of Two Markets
As levels of inventories have remained tight for the largest metros, additional leverage
that is mostly applied to the lower price tiers has stimulated even more demand. With
supply limited, this leverage has been capitalized into higher house price increases at
the lower end.
Annualized Home Sales (New vs Existing)
210Source: AEI Housing Center, www.AEI.org/housing, and First American Data Tree (DataTree.com).
The national housing market retreated slightly in 2018. In 2018, 6.0 million sales transactions were reported, down 90,000 transactions, or 1.5 percent, from 2017. New
home sales accounted for 10.7% of sales in 2018:Q4, which is up 0.9 ppt from 2013:Q4.
4.0
4.5
5.0
5.5
6.0
6.5
4.0
4.5
5.0
5.5
6.0
6.5
2012:Q4 2013:Q2 2013:Q4 2014:Q2 2014:Q4 2015:Q2 2015:Q4 2016:Q2 2016:Q4 2017:Q2 2017:Q4 2018:Q2 2018:Q4
New construction
Existing
Sales Transactions in millions
Data are for the 4 quarter period ending with respective quarter.
New construction share of all sales:2013:Q4: 9.8%2014:Q4: 10.0%2015:Q4: 10.3%2016:Q4: 10.6%2017:Q4: 10.9%2018:Q4: 10.7%
Quarterly Home Sales (New and Existing): by Type
211
Sales for 2018:Q4 were down (-7.9%) from 2017:Q4, with institutionally financed sales down (-7.5%). Compared to the 2012:Q4, sales by count have grown 13%. As predicted, over-all 2018 volume has come in about flat over 2017 volume. Over the next months, we
expect a rebound in sales due to lower mortgage rates and availability of leverage.
Source: AEI Housing Center, www.AEI.org/housing, and First American Data Tree (DataTree.com).
0
200,000
400,000
600,000
800,000
1,000,000
1,200,000
1,400,000
1,600,000
1,800,000
2,000,000
0
200,000
400,000
600,000
800,000
1,000,000
1,200,000
1,400,000
1,600,000
1,800,000
2,000,000
2012:Q4 2013:Q2 2013:Q4 2014:Q2 2014:Q4 2015:Q2 2015:Q4 2016:Q2 2016:Q4 2017:Q2 2017:Q4 2018:Q2 2018:Q4
All home sales
Institutionally financed sales
Cash sales
Other financed sales
Red markers show Q4 counts in each year.
Home purchase transactions
Origination Shares Based on Purchase Loan Counts
The GSEs, FHA, VA, and RHS continue to account for about 80% of institutionally financed home sales. The GSEs accounted for nearly half of all mortgage lending in 2018:Q4. They have more than regained market share lost to FHA after its mortgage insurance premium (MIP) cut in January 2015. FHA’s market share is now over 1 ppt. below its pre-MIP cut level. FHA and the GSEs have all had substantial increases in
mortgage risk.
212Note: Data are for institutionally financed sales only.Source: AEI Housing Center, www.AEI.org/housing, and First American Data Tree (DataTree.com).
0%
10%
20%
30%
40%
50%
60%
0%
10%
20%
30%
40%
50%
60%
2012:Q4 2013:Q2 2013:Q4 2014:Q2 2014:Q4 2015:Q2 2015:Q4 2016:Q2 2016:Q4 2017:Q2 2017:Q4 2018:Q2
GSE
FHA
Private
VA
RHS
Agency Refi and Purchase Loan Counts
Agency refi volume (by count) for December 2018 was down 46% from December 2017. The majority of refi lending is now Cash-Out refis, which accounted for 70% of all refis
during the current month. No Cash-Out refi volume has declined sharply with the increase in mortgage rates in 2017.
213
Note: Data are for Agency loan market only.Source: AEI Housing Center, www.AEI.org/housing.
0
100,000
200,000
300,000
400,000
500,000
600,000
700,000
800,000
0
100,000
200,000
300,000
400,000
500,000
600,000
700,000
800,000
Sep-12 Sep-13 Sep-14 Sep-15 Sep-16 Sep-17 Sep-18
Refis
Purchase loansNo-Cash-Out Refis
Cash-Out Refis
Total Red markers show December count in each year.Cash-out share of refi
Dec 2012 - 20%Dec 2013 - 31%Dec 2014- 34%Dec 2015 - 40%Dec 2016 - 41%Dec 2017 - 56%Dec 2018 - 70%
Agency First-Time Buyer Mortgage Share
214
The Agency First-Time Buyer Mortgage Share Index (FBMSI) for December 2018 stood at 59.2%, up 0.8 ppt and setting a new series’ high for the month of December.
Compared to five years ago, the FBMSI is up 3.8 ppts. from 55.4%. It appears that the index has increased from its already high level due to repeat buyers’ greater sensitivity
to higher rates.
Source: AEI Housing Center, www.AEI.org/housing.
53%
54%
55%
56%
57%
58%
59%
60%
61%
53%
54%
55%
56%
57%
58%
59%
60%
61%
Feb-13 Feb-14 Feb-15 Feb-16 Feb-17 Feb-18
Red markers show December share in each year.
First-time buyer mortgage share
Ratio of Sales Price for First-time to Repeat Buyers
215
The trend upward is towards higher first-time buyer (FTB) prices relative to repeat buyers (RBs). FTBs have access to the leverage punchbowl, thereby greatly
reducing the tendency to make downward quality adjustments to offset rapid home price appreciation. RBs without access to this punchbowl, tend to make downward
quality adjustments to offset home price appreciation. This adds to demand at lower price tiers.
Source: AEI Housing Center, www.AEI.org/housing.
70%
71%
72%
73%
74%
75%
76%
70%
71%
72%
73%
74%
75%
76%
Feb-13 Feb-14 Feb-15 Feb-16 Feb-17 Feb-18
Red markers show ratio in December of each year
Ratio of FTB to RB sale price
Median Sale Price by Risk Segment*, FTB Purchase Loans
216
Higher risk borrowers are being provided additional leverage which is fueling rapidly increasing home prices. Market prices for subprime borrowers have increased 27
percent since Feb-2013, while market prices for prime borrowers have only increased 12 percent.
95
100
105
110
115
120
125
130
95
100
105
110
115
120
125
130
Feb-13 Aug-13 Feb-14 Aug-14 Feb-15 Aug-15 Feb-16 Aug-16 Feb-17 Aug-17 Feb-18 Aug-18
Subprime (4.0% avg. annual growth)
Dec-18: $221,000
Prime(2.0% avg. annual growth)
Dec-18: $307,000
Prime 270,000$
Subprime 170,000$
Mean price Feb-13
Prime Subprime
Feb-13 3.0% 20.4%
Dec-18 3.2% 22.7%
NMRI
Index: Feb-13 == 100
* We define prime loans as low-risk (with a stressed default rate of less than 6%), and subprime as high risk (with a stressed default rate of 12% or greater).Source: AEI Housing Center, www.AEI.org/housing.