structural reforms in a debt overhangstructural reforms in a debt overhang javier andrØs1, Óscar...
TRANSCRIPT
Structural Reforms in a Debt Overhang
Javier Andrés1, Óscar Arce2 and Carlos Thomas3
National Bank of Belgium, June 18 2014
1Universidad de Valencia, Banco de España2Banco de España3Banco de EspañaAndrés, Arce & Thomas () Reforms in a Debt Overhang
National Bank of Belgium, June 18 2014 1/ 34
Motivation (I)
High levels of private debt and need to deleverage act as a drag ongrowth in the EMU periphery
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 2
/ 34
Motivation (II)
In the short term, little room for
fiscal policy (large deficits)(conventional) monetary policy (ZLB).
Much of the focus is on structural reforms, mainly in product andlabor markets.
Most offi cial views (e.g. OECD, IMF, ECB) support reforms.
Reforms are clearly positive in the long run, but their short/mediumterm impact is less well understood.
EMU periphery conditioned by high debt cum slow private-sectordeleveragingThis paper: study impact of structural reforms in an environment ofslow deleveraging
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 3
/ 34
Framework
DSGE model, small open economy inside monetary union
Lenders & borrowers, collateral constraints à la Kiyotaki & Moore(1997). As in Iacoviello (2005), real estate is the only collateral.
Key point of departure: long-term debt ⇒ double debt regime:
a) In normal times: debt restricted by value of collateralb) In crisis times: no new credit, debt amortized slowly
Baseline deleveraging scenario: negative shock to LTV ratios (’creditcrunch’)⇒ economy enters regime (b): slow and protracted deleveraging
Time of change from regime (b) to (a), i.e. end of deleveragingphase, is endogenous
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 4
/ 34
Preview of results
Structural reforms (reductions in desired price & wage markups)boost output in long run (as expected), but also in short run.
Particularly true for product market reform
Brings forward the (endogenous) end of deleveraging phase/recession
Labor market reform creates modest short-run gains
Double layer of nominal rigidities (wages and prices) delaysimprovement in price competitivenessBroader reform (including higher wage flexibility) generates sizableshort-run gains
Long-run debt weakens negative Fisherian debt deflation effect(becomes second order)
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 5
/ 34
Recent literature
Some recent work on the impact of reforms:
Eggertsson, Ferrero & Raffo (2014):
if monetary policy is at ZLB, deflationary structural reforms increasereal interest rate → depress aggregate demandthis channel may dominate positive income effect (from long-run gains)in the short run
Galí & Monacelli (2013): short-run effects of wage moderation(through lower payroll taxes) is small if no monetary accommodation
Fernández-Villaverde, Guerrón-Quintana & Rubio-Ramírez (2012):
credible announcement of future structural reforms triggers gainsalready in the short-run (positive income effect)BUT no deflationary effect on impact
None of these papers study effects of reforms in a scenario of slowdeleveraging
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 6
/ 34
Model structure
Small open economy in a monetary union⇒ monetary policy exogenous ≈ ZLBThree consumer types
Patient households (lenders)Impatient households (borrowers)(Impatient) entrepreneurs (borrowers)
Three production sectors
Consumption goods (entrepreneurs + retailers)Equipment capital producersConstruction
Trade with rest of world: consumption goods and foreign debt
Standard real and nominal frictions: investment adjustment costs,nominal price and wage rigidities
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 7
/ 34
Impatient households
Maximize
E0∞
∑t=0
βt
{log (ct ) + ϑ log ht − χ
∫ 1
0
nCt (i)1+ϕ
1+ ϕdi
},
subject to
ct + pht [ht − (1− δh) ht−1] = bt −Rt−1πt
bt−1 +∫ 1
0
Wt (i)Pt
nCt (i) di .
and an asymmetric debt constraint...
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 8
/ 34
Asymmetric debt constraint
We assume long run debt
A constant fraction 1− γ of nominal outstanding principal isamortized each period (Woodford, 2001)
Dynamics of real outstanding debt,
bt =bt−1πt
+ bnewt − 1− γ
πtbt−1 =
γ
πtbt−1 + bnewt .
bnewt : gross new credit
If collateral value < γ bt−1πt, setting bt = collateral value would require
bnewt < 0 ...
... but debtor cannot be forced to pay back faster than 1− γ; hencebnewt = 0.
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 9
/ 34
Asymmetric debt constraint (cont’d)
This implies a double debt regime:
in ’normal’times, borrowing is restricted by expected discounted valueof collateral,
1RtmtEtπt+1p
ht+1ht ,
mt : exogenous loan-to-value (LTV) ratiowhen collateral values fall below contractual amortization path,γbt−1/πt , the latter becomes the effective debt limit
Formally,
bt ≤{
1RtmtEtπt+1pht+1ht ,
1RtmtEtπt+1pht+1ht ≥ γ bt−1πt
γ bt−1πt, 1
RtmtEtπt+1pht+1ht < γ bt−1πt
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 10
/ 34
Entrepreneurs
Maximize
E0∞
∑t=0
βt log cet ,
subject to
cet + pht [h
et − (1− δh) h
et−1] + qt [kt − (1− δk ) kt−1]
= mcty et −Wt
Ptnet + b
et −
Rt−1πt
bet−1 + ∑s=r ,h,k
Πrt ,
y et = Atkαkt−1 (h
et−1)
αh (net )1−α−αk ,
bet ≤{
1Rtmet Etπt+1p
ht+1h
et ,
1Rtmet Etπt+1p
ht+1h
et ≥ γe
bet−1πt
γebet−1πt, 1
Rtmet Etπt+1p
ht+1h
et < γe
bet−1πt
.
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 11
/ 34
Calibration
We target key ratios of the Spain in 2007:
Ratio Data (%) Model (%)construction share of GDP 12.45 15.11construction share of employment 13.39 15.44labor share of GDP 61.59 64.84corporate debt / annual GDP 125.36 128.85household debt / annual GDP 80.22 79.94net foreign debt / annual GDP 79.3 79.3gross exports / GDP 26.9 26.9
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 12
/ 34
Calibration (2)
Parameters not pinned down by targets are set to standard valueswithin NK-DSGE literature
Parameters affecting debt constraints
LTV ratios: households m = 0.85, entrepreneurs me = 0.71Amortization rates: households 1− γ = 0.02, entrepreneurs1− γe = 0.04⇒ average debt maturity: 1/ (1− γ) = 50, 1/ (1− γe ) = 25 qrts
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 13
/ 34
Baseline scenario: a deleveraging shock
We simulate a deleveraging shock for entrepreneurs and constrainedhouseholds:
Gradual, permanent fall (10pp) in loan-to-value (LTV) ratios: mt , met
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 14
/ 34
Deleveraging shock: LTV ratios
0 10 20 30 40 50 600.75
0.76
0.77
0.78
0.79
0.8
0.81
0.82
0.83
0.84
0.85LTV households (m)
quarters0 10 20 30 40 50 60
0.6
0.62
0.64
0.66
0.68
0.7
0.72LTV entrepreneurs (me)
quarters
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 15
/ 34
Deleveraging shock: regime changes
0 10 20 30 40 50 60
8
10
12
14
16
18
quarters
Entrepreneur debt
γebet1/πt
met ph
t+1πt+1het /Rt
bet
T*
0 10 20 30 40 50 602
4
6
8
10
12
14
quarters
Household debt
γbt1/πt
mtpht+1πt+1ht/Rt
btT**
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 16
/ 34
Baseline scenario: a deleveraging shock
Large initial shock and asymmetric debt limits produce a doubleregime change:
For t = 1, ...,T ∗, value of entrepreneurs’collateral falls belowγebet−1/πtFor t = 1, ...,T ∗∗, value of households’collateral falls below γbt−1/πt
γe < γ⇒ T ∗ < T ∗∗: faster amortization of entrepreneurial debt
Dates of regime change T ∗ and T ∗∗ are solved endogenously
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 17
/ 34
Deleveraging shock: macroeconomic effects
0 20 40 6010
5
0
5GDP
quarters
%
0 20 40 6010
5
0
5employment
quarters%
0 20 40 6040
60
80
100household debt / GDP
quarters
pp
0 20 40 6080
100
120
140entrepreneurial debt / GDP
quarters
pp
0 20 40 6010
5
0
5real estate prices
quarters
%
0 20 40 6030
20
10
0total investment
quarters
%
0 20 40 606
4
2
0
2total consumption
quarters
%
0 20 40 602
1
0
1CPI inflation
quarters
annu
aliz
ed p
p
0 20 40 606
4
2
0real wage
quarters
%
0 20 40 601
0
1
2exante real interest rate
quarters
annu
aliz
ed p
p
0 20 40 602
1
0
1
2terms of trade
quarters
%
0 20 40 6050
0
50
100net exports
quarters
%Andrés, Arce & Thomas () Reforms in a Debt Overhang
National Bank of Belgium, June 18 2014 18/ 34
Deleveraging shock: long vs short-term debt
Long run debt produces a more realistic deleveraging path and (critically)allows for endogenous regime change
0 20 40 6020
10
0
10GDP
quarters
%
0 20 40 6030
20
10
0
10employment
quarters
%
0 20 40 6020
40
60
80
100household debt / GDP
quarters
pp
0 20 40 6080
100
120
140entrepreneurial debt / GDP
quarters
pp
0 20 40 6010
5
0
5real estate prices
quarters
%
0 20 40 6030
20
10
0total investment
quarters
%
0 20 40 6030
20
10
0
10total consumption
quarters%
0 20 40 604
2
0
2CPI inflation
quarters
annu
aliz
ed p
p
0 20 40 606
4
2
0real wage
quarters
%
0 20 40 601
0
1
2exante real interest rate
quarters
annu
aliz
ed p
p
0 20 40 602
1
0
1
2terms of trade
quarters
%
0 20 40 60200
0
200
400net exports
quarters
%
longrun debtshortrun debt
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 19
/ 34
Deleveraging shockDebt, consumption and investment
Two phases in the dynamics of debt:
Until T ∗ (T ∗∗), smooth deleveraging at rate γe/πt (γ/πt )After T ∗ (T ∗∗), debt picks up quickly: real estate is again valuable ascollateral ⇒ asset prices, credit and investment "virtuous circle"
Consumption follows a similar pattern to debt
Investment recovers somewhat earlier than consumption and debt(’creditless recovery’)
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 20
/ 34
Product market reform
We simulate a sudden, permanent fall in desired price markups (5%)
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 21
/ 34
Product market reform
0 20 40 6010
5
0
5GDP
quarters
%
0 20 40 6010
5
0
5employment
quarters%
0 20 40 6040
60
80
100household debt / GDP
quarters
pp
0 20 40 6080
100
120
140entrepreneurial debt / GDP
quarters
pp
0 20 40 6010
5
0
5real estate prices
quarters
%
0 20 40 6030
20
10
0
10total investment
quarters
%
0 20 40 606
4
2
0
2total consumption
quarters
%
0 20 40 606
4
2
0
2CPI inflation
quarters
annu
aliz
ed p
p
0 20 40 605
0
5real wage
quarters
%
0 20 40 601
0
1
2
3exante real interest rate
quarters
annu
aliz
ed p
p
0 20 40 604
2
0
2terms of trade
quarters
%
0 20 40 6050
0
50
100net exports
quarters
%
baselinereform
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 22
/ 34
Product market reform: macro effects
Long run:
GDP goes up, employment remains stable (real wages and labourshare go up)
Short/medium run:
GDP and employment fall by less than in the baseline
Investment behaves significantly better, anticipating higher futuredemand.
Consumption falls slightly below the baseline
Additional terms of trade depreciation fuels gross exports, though netexports worsen due to stronger domestic demand
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 23
/ 34
Product market reform: positive effect on investment
Key question: How is the additional investment financed in the short term?- On the one hand,
Entrepreneurs current unit profits drop as markups fall
Deflationary effect of reform raises the real value of debt repayments
- On the other hand,
Higher asset prices → entrepreneurs’net worth is higher in the reformscenario
Entrepreneurs cut down their consumption significantly
Total demand goes up, pushing up profits
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 24
/ 34
Product market reform: deleveraging ends earlier
Reform brings forward the end of the deleveraging phase: T ∗ and T ∗∗
both go down.
Focus on T ∗ (entrepreneurs):
Higher initial net worth allows for more investment in the short termHigher investment today implies higher net worth and investmenttomorrow, and so onFaster recovery of net worth leads ceteris paribus to an earlier T ∗
Anticipation of earlier recovery of credit leads to higher asset pricestoday, higher net worth and investment, etc.
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 25
/ 34
Labor market reform
We simulate a sudden, permanent fall in desired wage markups (5%).
Model proxy for unions’bargaining power.
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 26
/ 34
Labor market reform
0 20 40 6010
5
0
5GDP
quarters
%
0 20 40 6010
5
0
5employment
quarters%
0 20 40 6040
60
80
100household debt / GDP
quarters
pp
0 20 40 6080
100
120
140entrepreneurial debt / GDP
quarters
pp
0 20 40 6010
5
0
5real estate prices
quarters
%
0 20 40 6030
20
10
0
10total investment
quarters
%
0 20 40 606
4
2
0
2total consumption
quarters
%
0 20 40 602
1
0
1CPI inflation
quarters
annu
aliz
ed p
p
0 20 40 606
4
2
0real wage
quarters
%
0 20 40 601
0
1
2exante real interest rate
quarters
annu
aliz
ed p
p
0 20 40 604
2
0
2terms of trade
quarters
%
0 20 40 6050
0
50
100net exports
quarters
%
baselinereform
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 27
/ 34
Labor market reform (cont’d)
Long-run gains in GDP and employment
Short/medium-run effects:
No effect on GDP on impact, then gradual improvementSimilar effect on employment (main variable targeted by such a reform)
Positive short/medium-run effects smaller than those of productmarket reform:
Investment does not respond positively: entrepreneurs meet higherdemand by hiring more (cheaper) laborEntrepreneur consumption slightly increases⇒ forces that brought T ∗’s forward with product market reform arenot active now
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 28
/ 34
Broader labor market reform
Reduction in desired wage markups must overcome a double layer ofnominal rigidities (wages and prices) before affecting pricecompetitiveness
Typically, labor market reforms affect not only markups, but alsospeed of nominal wage adjustment
Spain’s 2012 reform a clear example!
Consider a broader labor market reform that also reduces nominalwage rigidity
Reduce Calvo parameter from 3/4 to 2/3 (average wage duration from4 to 3 qrts)
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 29
/ 34
Broader labor market reform
0 20 40 6010
5
0
5GDP
quarters
%
0 20 40 6010
5
0
5employment
quarters%
0 20 40 6040
60
80
100household debt / GDP
quarters
pp
0 20 40 6080
100
120
140entrepreneurial debt / GDP
quarters
pp
0 20 40 6010
5
0
5real estate prices
quarters
%
0 20 40 6030
20
10
0total investment
quarters
%
0 20 40 606
4
2
0
2total consumption
quarters
%
0 20 40 603
2
1
0
1CPI inflation
quarters
annu
aliz
ed p
p
0 20 40 606
4
2
0real wage
quarters
%
0 20 40 601
0
1
2
3exante real interest rate
quarters
annu
aliz
ed p
p
0 20 40 604
2
0
2terms of trade
quarters
%
0 20 40 6050
0
50
100net exports
quarters
%
baselinereform
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 30
/ 34
Further analysis
Two important channels for understanding the positive short-runeffects of reforms:
The role of the external sectorThe role of long-term debt
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 31
/ 34
The role of the external sector
Responsiveness of net exports to reform-driven depreciation in terms oftrade is key
0 5 10 15 20 25 30 35 400.5
0
0.5
1
1.5
2
2.5
3
3.5
4Product market reform, export elasticity
quarters
%
εF = 1 (baseline)εF = 0.5εF = 1.5
0 5 10 15 20 25 30 35 400.4
0.2
0
0.2
0.4
0.6
0.8
1
1.2Labor market reform, export elasticity
quarters
%
εF = 1 (baseline)εF = 0.5εF = 1.5
0 5 10 15 20 25 30 35 400.4
0.2
0
0.2
0.4
0.6
0.8
1
1.2Labor market reform, import elasticity
quarters
%εH = 1 (baseline)εH = 0.5εH = 1.5
0 5 10 15 20 25 30 35 400.5
0
0.5
1
1.5
2
2.5
3
3.5
4Product market reform, import elasticity
quarters
%
εH = 1 (baseline)εH = 0.5εH = 1.5
Differential effect of reform on GDP
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 32
/ 34
The role of long-run debt
Entrepreneurial net debt flows (gross of interest payments) duringdeleveraging phase (bet = γebet−1/πt , t ≤ T ∗):
Rt−1πt
bet−1 − bet =Rt−1 − γe
πtbet−1
=
net interest rate︷ ︸︸ ︷(Rt−1 − 1) +
amortization rate︷ ︸︸ ︷(1− γe )
πtbet−1.
Long-run debt ⇒ amortization rate 1− γe is small ⇒ debt deflationeffect (↓ πt) from reform is small!
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 33
/ 34
Concluding remarks
Structural reforms may boost GDP and employment already in theshort run...
... even without monetary accommodation
Especially true for product market reform (brings forward end ofdeleveraging/recession)
Also true for a broad labor market reform that includes higher wageflexibility
Long-run debt buffers short-term costs of reforms
Andrés, Arce & Thomas () Reforms in a Debt OverhangNational Bank of Belgium, June 18 2014 34
/ 34