Robert J. GordonRobert J. Gordon

For Presentation at For Presentation at

Advanced Workshop for Central Advanced Workshop for Central Bankers,Bankers,

Northwestern, September 28, 2004Northwestern, September 28, 2004

Productivity Growth:Productivity Growth:Concepts, Causes, and theConcepts, Causes, and the

Diverging U. S. PerformanceDiverging U. S. Performance

Ultimate Measure of Ultimate Measure of Economic SuccessEconomic Success

Standard of Living = Income per capitaStandard of Living = Income per capita– 1.3% growth, doubles every 53 years 1.3% growth, doubles every 53 years

(Philippines)(Philippines)– 5.6% growth, doubles every 12 years 5.6% growth, doubles every 12 years

(Korea)(Korea) For very long-term growth or comparing For very long-term growth or comparing

rich and poor nations, Income per capita rich and poor nations, Income per capita and productivity are the same thingand productivity are the same thing

Not the same thing for short-term or Not the same thing for short-term or comparisons among rich nationscomparisons among rich nations

How Productivity is Related to Total Output

Output (Q) Equal to the product of: Productivity (Q/A) Hours per Employee (A/E) Employment Rate (E/L),

that’s just (1 – U/L) Labor-force Participation

Rate (L/N) Working-age Population

(N)

NN

L

L

E

E

A

A

QQ

How Productivity is Related to Output per Capita

Output (Q) Equal to the product of: Productivity (Q/A) Hours per Employee (A/E) Employment Rate (E/L),

that’s just (1 – U/L) Labor-force Participation

Rate (L/N) Working-age Population

(N)

N

L

L

E

E

A

A

Q

N

Q

How Could Europe be So Productive Yet So Poor

Output per Capita (Q/N)In Europe 75% of U. S.Productivity 95% of U. S.The Difference:

Hours per Employee (A/E) Employment Rate (E/L) Labor-force Participation

Rate (L/N)

N

L

L

E

E

A

A

Q

N

Q

Determinants of Determinants of Productivity GrowthProductivity Growth Simplest Production FunctionSimplest Production Function

Q = F(K,N)Q = F(K,N) Add TechnologyAdd Technology

Q = F(K,N,T)Q = F(K,N,T) This is the Solow Growth ModelThis is the Solow Growth Model Predictions:Predictions:

– Raising saving rate affects growth rate Raising saving rate affects growth rate only temporarilyonly temporarily

– Universal convergence Universal convergence

Puzzles the Solow Puzzles the Solow ModelModel

Does Not ExplainDoes Not Explain Lack of Convergence (East Asia Lack of Convergence (East Asia

vs. Africa and Latin America)vs. Africa and Latin America) To explain a 10-1 ratio of Qpc in To explain a 10-1 ratio of Qpc in

rich vs. poor countries, rich vs. poor countries, unrealistically requires:unrealistically requires:– 10000 times as much capital per 10000 times as much capital per

capitacapita– 1/1000 the rate of return on capital1/1000 the rate of return on capital

What is Missing?What is Missing?

Lack of Human CapitalLack of Human Capital Q = F(K,N,T,H)Q = F(K,N,T,H) Instead of receiving 25% of national Instead of receiving 25% of national

income, total capital (human and income, total capital (human and physical) receives 90%physical) receives 90%

Labor’s 75% income share is Labor’s 75% income share is interpreted as 10% “brute force labor” interpreted as 10% “brute force labor” and 65% as reward to education and and 65% as reward to education and experienceexperience

What Else is Missing?What Else is Missing?

Human capital resolves the problem Human capital resolves the problem about unrealistic ratios of rich-to-about unrealistic ratios of rich-to-poor K/N and return to capitalpoor K/N and return to capital

But it leaves the “Rio Grande Puzzle”But it leaves the “Rio Grande Puzzle” How to go from 40c per hour to $10 How to go from 40c per hour to $10

per hour in one easy wadeper hour in one easy wade Human capital is not changed, must Human capital is not changed, must

be something elsebe something else

Technology is not FreeTechnology is not Free

The “New” Growth Theory Makes The “New” Growth Theory Makes Technology part of economicsTechnology part of economics– Must pull resources from production to Must pull resources from production to

work in R&D labswork in R&D labs– Must provide an incentive for innovationMust provide an incentive for innovation– Patent ProtectionPatent Protection– Technology requires H, K to be usedTechnology requires H, K to be used

Case of the drug companies, AIDS in Case of the drug companies, AIDS in Africa, importing drugs from CanadaAfrica, importing drugs from Canada

The New Comparative The New Comparative Economics: What Else is Economics: What Else is

Missing?Missing? Geography: The Tragedy of the Geography: The Tragedy of the

Tropics (soil, diseases, enervation)Tropics (soil, diseases, enervation) Crime, corruptionCrime, corruption Infrastructure Infrastructure

– Lack of phones, electricity, roadsLack of phones, electricity, roads

Putting it TogetherPutting it Together

The augmented production function, it The augmented production function, it starts to explain rich vs. poorstarts to explain rich vs. poor

How this solves the Rio Grande puzzleHow this solves the Rio Grande puzzle

P = political capital (legal system etc)P = political capital (legal system etc)

G = geographyG = geography

R = infrastructureR = infrastructure

Q = F(K,N,T,H,P,G,R)Q = F(K,N,T,H,P,G,R)

Application #1: How Can Application #1: How Can Europe be So Productive Europe be So Productive Yet So PoorYet So Poor The History: Europe falls back 1870-The History: Europe falls back 1870-

1950 and then catches up1950 and then catches up The catch-up is almost complete in The catch-up is almost complete in

productivity (Q/A)productivity (Q/A) The catch-up is incomplete in output The catch-up is incomplete in output

per capita (Q/N)per capita (Q/N) Why?Why?

– Must be that Europe’s A/N is lowerMust be that Europe’s A/N is lower– Why?Why?

Per Capita Real GDPPer Capita Real GDPper Capita Real GDP, Europe and the United States,

Selected Years, 1820-2000

1000

10000

100000

1820 1830 1840 1850 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Co

nst

ant

1990

Gea

ry-K

ham

is D

olla

rs

United States

Europe

Essential Features of Essential Features of Income per Capita since Income per Capita since 18701870 Steady rate of real GDP per capita growth Steady rate of real GDP per capita growth

in the USin the US– 1.81% per year growth between 1870-20001.81% per year growth between 1870-2000– Huge acceleration between 1963-73Huge acceleration between 1963-73

Slower growth in EuropeSlower growth in Europe– 1.67% per year growth between 1870-20001.67% per year growth between 1870-2000– Downward dislocations due to the World WarsDownward dislocations due to the World Wars– Golden years of catch-up between 1950-1973Golden years of catch-up between 1950-1973

Since 1973 catch-up is completeSince 1973 catch-up is complete

Real GDP per HourReal GDP per HourReal GDP per Hour, Europe and the United States,

Selected Years, 1870-2000

1

10

100

1820 1830 1840 1850 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Co

nst

ant

1990

Gea

ry-K

ham

is D

olla

rs p

er H

ou

r

United States

Europe

Summarizing the Summarizing the Productivity RecordProductivity Record U.S. record of productivity growth is U.S. record of productivity growth is

not as steady as for output-per-capitanot as steady as for output-per-capita– Strongest performance between 1938-50Strongest performance between 1938-50– Slowdown between 1973-92Slowdown between 1973-92

Europe plays catch-upEurope plays catch-up– Much slower growth than the U.S. between Much slower growth than the U.S. between

1870-1950 (1.50% vs 2.15% for the US1870-1950 (1.50% vs 2.15% for the US– Nearly closes the gap by 2000Nearly closes the gap by 2000

In this section we’re ignoring the new In this section we’re ignoring the new divergence after 2000divergence after 2000

Output per CapitaOutput per Capitaand Output per Hourand Output per Hour

Ratio of Europe to the United States, Output per Capita and Output per Hour,

selected years, 1820-2000

40

50

60

70

80

90

100

110

1820 1830 1840 1850 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Per

cen

t

Output per Capita

Output/Hour

Features of the Output Features of the Output per Captia, per Captia, Europe/U.S. ratioEurope/U.S. ratio The Europe/U.S. ratio of output The Europe/U.S. ratio of output

per capita declines steadily from per capita declines steadily from 1829 to 1950.1829 to 1950.

Upsurge from 1950-1973Upsurge from 1950-1973 Stagnation between 1973-2000Stagnation between 1973-2000

Europe/U.S. ratio for Europe/U.S. ratio for productivity growthproductivity growth The same downward slide The same downward slide

between 1870 and 1950between 1870 and 1950 Europe has a higher level of hours Europe has a higher level of hours

per capitaper capita After 1950 much faster growth in After 1950 much faster growth in

the productivity ratiothe productivity ratio

Real GDP per Capita and Real Real GDP per Capita and Real GDP per HourGDP per Hour

  

 1820-70

 1870-1913

 1913-1929

 1929-1950

 1950-1973

 1973-2000

  Output per Capita

 

  

  

  

  

  

  Europe

 1.05

 1.25

 0.97

 0.79

 3.61

 1.77

  U. S.

 1.29

 1.79

 1.65

 1.55

 2.40

 1.64

  Europe - U. S.

 -0.24

 -0.54

 -0.68

 -0.76

 1.21

 0.13

  Output per Hour

  

  

  

  

  

  

  Europe

 

 1.49

 1.76

 1.35

 4.44

 2.40

  U. S.

 

 1.90

 2.40

 2.48

 2.68

 1.37

  Europe - U. S.

 

 -0.41

 -0.64

 -1.13

 1.76

 1.03

  YpC / YpH

  

  

  

  

  

  

  Europe

  

 -0.24

 -0.79

 -0.56

 -0.83

 -0.63

  U. S.

  

 -0.11

 -0.75

 -0.93

 -0.28

  0.27

  Europe - U. S.

  

 -0.13

 -0.04

 0.37

 -0.55

 -0.90

The Post-1950 The Post-1950 ReversalReversal Sharp turn of Europe/U.S. ratios of Sharp turn of Europe/U.S. ratios of

output per capita and productivity output per capita and productivity after 1950.after 1950.

Sharp retardation in growth of Sharp retardation in growth of output per capita in Europe relative output per capita in Europe relative to productivity growth after 1950.to productivity growth after 1950.– Longer vacations contribute to few Longer vacations contribute to few

hours worked per employeehours worked per employee

The Contributions of E/N The Contributions of E/N and H/Eand H/E

Ratio of Europe to the United States, Ratio of Output per Capita to Output per Hour, Decomposed into Hours/ Employee and Employee/Population Ratios, selected

years, 1870-2000

80

90

100

110

120

1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Pe

rce

nt

Output PC/Output PH

Hours/Employee

Employees/Population

Standard of living: Standard of living: held down by held down by vacations (H/E)vacations (H/E) Have citizens chosen to use their Have citizens chosen to use their

prosperity to take longer prosperity to take longer vacations in contrast to vacations in contrast to Americans? Americans?

Have Europeans been forced to Have Europeans been forced to take vacations because of union take vacations because of union or parlimentary politics?or parlimentary politics?

Ian on Work HoursIan on Work Hours

““To call long work hours in To call long work hours in America a bad thing seems odd”America a bad thing seems odd”

““People here have the choice to People here have the choice to work as long as they want”work as long as they want”– ““Europeans would work longer if Europeans would work longer if

they could”they could”– ““France wouldn’t need labor police if France wouldn’t need labor police if

nobody wanted to work more than nobody wanted to work more than 35 hours”35 hours”

Europe’s Low E/N Europe’s Low E/N Matters as much as Low Matters as much as Low H/EH/E High UnemploymentHigh Unemployment

– High Youth UnemploymentHigh Youth Unemployment– High long-term UnemploymentHigh long-term Unemployment

Low Labor-force ParticipationLow Labor-force Participation– Of YouthOf Youth– Of ElderlyOf Elderly

Causes of Low E/NCauses of Low E/N

Lack of Job Opportunities for Lack of Job Opportunities for Youth:Youth:– Late Marriage AgesLate Marriage Ages– Late Development of IndependenceLate Development of Independence

U. S. Youths working in High School and U. S. Youths working in High School and CollegeCollege

– Low Fertility RatesLow Fertility Rates– Italy: Living at Home with MamaItaly: Living at Home with Mama

Poor Labor-Market Poor Labor-Market Performance in EuropePerformance in Europe

Why is Average EU Why is Average EU Unemployment Rate Higher than Unemployment Rate Higher than US, LFPR Lower?US, LFPR Lower?

Minimum Wages, U BenefitsMinimum Wages, U Benefits Regulations on Hiring, Firing, Plant Regulations on Hiring, Firing, Plant

Closings, Plant OpeningsClosings, Plant Openings This is an old Story, still validThis is an old Story, still valid

Phelps’ Refreshing Phelps’ Refreshing departure from departure from VaguenessVagueness Too little competition, too much Too little competition, too much

corporatismcorporatism ““penalties, impediments, prohibitions, penalties, impediments, prohibitions,

mandates” that dampen “creative mandates” that dampen “creative destruction”destruction”

Youth in America vs. Europe, culture of Youth in America vs. Europe, culture of “dependency”“dependency”

American teens work at McDonalds, American teens work at McDonalds, pay part of their college expensespay part of their college expenses

Those Italian men!Those Italian men!

Other Big IssuesOther Big Issues

GDP Exaggerates U. S. GDP per CapitaGDP Exaggerates U. S. GDP per Capita– This has nothing to do with CompetitionThis has nothing to do with Competition– Extreme climate, lots of air conditioning, Extreme climate, lots of air conditioning,

low petrol prices, huge excess energy uselow petrol prices, huge excess energy use– Crime, excessive urban density impose Crime, excessive urban density impose

costscosts U. S. Medical Care Inefficiency Creates U. S. Medical Care Inefficiency Creates

Medicare CrisisMedicare Crisis U. S. Social Security Crisis can be put U. S. Social Security Crisis can be put

off forever through open immigrationoff forever through open immigration

This is not black vs. This is not black vs. white. It reflects white. It reflects different valuesdifferent values U. S. Low-density metro areas U. S. Low-density metro areas

dependent on auto, high dependent on auto, high unmeasured cost of traffic unmeasured cost of traffic congestion, subsidies to auto congestion, subsidies to auto transit, starvation of public transittransit, starvation of public transit

Europe high-density metro areas, Europe high-density metro areas, unmeasured time cost of public unmeasured time cost of public transit, subsidies to public transittransit, subsidies to public transit

Ian on Urban DensityIan on Urban Density

““We overspend on highways, they overspend We overspend on highways, they overspend on trains”on trains”

““We live in suburbs and have long commutes, We live in suburbs and have long commutes, they live in cramped homes and are closer to they live in cramped homes and are closer to work”work”

““We have options: in Chicago I can live in a We have options: in Chicago I can live in a suburb and drive OR live in an apartment and suburb and drive OR live in an apartment and walk to work”walk to work”

Contra Ian, many Americans lack such optionsContra Ian, many Americans lack such options– Inner city African Americans seeking suburban Inner city African Americans seeking suburban

jobsjobs– Many medium and small cities have virtually no Many medium and small cities have virtually no

public transit options, and there are few jobs public transit options, and there are few jobs where you can “walk to work”where you can “walk to work”

A Solid Reason why the A Solid Reason why the U. S.U. S.

Welfare Level is Truly Welfare Level is Truly HigherHigher Hedonic regressions show: people Hedonic regressions show: people

value square feet of housing and value square feet of housing and exterior landexterior land

The average American housing unit is The average American housing unit is more than double the average more than double the average European unitEuropean unit

The land area is at least 4x, maybe The land area is at least 4x, maybe moremore

The time cost of commuting may be The time cost of commuting may be less when all the delays of public less when all the delays of public transit are taken into accounttransit are taken into account

Summarizing Welfare Summarizing Welfare ComparisonComparison

Started with Europe/ US RatiosStarted with Europe/ US RatiosQ/N 77Q/N 77 Q/A 93Q/A 93

One-third of A/N is voluntaryOne-third of A/N is voluntary

Q/N 82Q/N 82 Q/A 93Q/A 93 One-half of remaining YPC One-half of remaining YPC

difference disappears because U. difference disappears because U. S. GDP is overstatedS. GDP is overstated

Q/N 91Q/N 91 Q/A 102Q/A 102

The New Productivity The New Productivity Divergence Divergence

Focus on 1995-2003Focus on 1995-2003 Growth rates of GDP per Hour Growth rates of GDP per Hour

WorkedWorked– U. S. 2.33U. S. 2.33– Europe 1.15Europe 1.15– Difference 1.18Difference 1.18

Over eight years, causes Europe/US Over eight years, causes Europe/US to fall back from 94 to 85 percentto fall back from 94 to 85 percent

The U. S. Productivity The U. S. Productivity Growth “Explosion”Growth “Explosion”

LP Actual vs Trend

-2.00

-1.00

0.00

1.00

2.00

3.00

4.00

5.00

6.00

1948 1953 1958 1963 1968 1973 1978 1983 1988 1993 1998 2003

133 Years: Falling Behind, Catching Up, Now Falling Behind

Annual Growth Rate of GDP per Hour, EU minus US, 1870-2003

-1.50

-1.00

-0.50

0.00

0.50

1.00

1.50

2.00

2.50

1870-1913 1913-1950 1950-1973 1973-1995 1995-2003

The Reversal Shown in Levels

GDP per Hour, EU as a percent of US, 1870-2003

0

20

40

60

80

100

120

1870 1890 1910 1930 1950 1970 1990

A Closer Look at the Last A Closer Look at the Last DecadeDecade

Annual Growth Rate of GDP per Hour, EU and US, 1990-2003

0.0

0.5

1.0

1.5

2.0

2.5

3.0

1990-95 1995-2000

U.S.

E.U.

Basic Paradox about ITBasic Paradox about IT

Both Europe and U. S. Rapidly Both Europe and U. S. Rapidly Adopted New Economy TechnologyAdopted New Economy Technology– Personal ComputersPersonal Computers– Web AccessWeb Access– Mobile PhonesMobile Phones

But Europe hasn’t taken offBut Europe hasn’t taken off Conclusion: Role of IT in U. S. revival Conclusion: Role of IT in U. S. revival

must have been exaggeratedmust have been exaggerated

Output per Hour by Industry Group, EU and US, 1990-2003

-2.0

0.0

2.0

4.0

6.0

8.0

10.0

12.0

1990-1995 1995-2001

US ICT Pro

EU ICT Pro

US ICT Using

EU ICT Using

US Non-ICT

EU Non-ICT

Finding the Culprit Industries

Where is the Difference? Where is the Difference?

The Van-Ark The Van-Ark DecomposionDecomposion 55% retail trade55% retail trade 24% wholesale trade24% wholesale trade 20% securities20% securities Rest of the economy: ZERORest of the economy: ZERO U. S. negative in telecom, U. S. negative in telecom,

backwardness of mobile phonesbackwardness of mobile phones

Europe in RetailingEurope in Retailing

Not uniform – Carrefour, IkeaNot uniform – Carrefour, Ikea U. S. “Big Boxes” (Wal-Mart, Home U. S. “Big Boxes” (Wal-Mart, Home

Depot, Best Buy, Target)Depot, Best Buy, Target) Europe: Europe:

– Land-use regulation, planning approvalLand-use regulation, planning approval– Shop-closing restrictionsShop-closing restrictions– Central-city congestion, protection of Central-city congestion, protection of

central-city shopping precinctscentral-city shopping precincts

Not enough emphasis on Not enough emphasis on new vs. oldnew vs. old

It’s not just that land-use planning It’s not just that land-use planning prevents Wal-mart from setting prevents Wal-mart from setting up a new big box on every up a new big box on every highway interchange in Europehighway interchange in Europe

It’s that the MIX of retailing in It’s that the MIX of retailing in Europe is heavily composed of Europe is heavily composed of small, old-fashioned firmssmall, old-fashioned firms

Let’s Walk down a street Let’s Walk down a street in Paris on the Left Bankin Paris on the Left Bank

Every few blocks, a green cross Every few blocks, a green cross indicating a pharmacyindicating a pharmacy

To American eyes, these are antique To American eyes, these are antique anachronismsanachronisms– One-by-one service at the counter, no One-by-one service at the counter, no

check out stationscheck out stations– Tiny, small, don’t carry any of the obvious Tiny, small, don’t carry any of the obvious

things that a pharmacy should carry. things that a pharmacy should carry. Walgreens.Walgreens.

University FundingUniversity Funding

Block grants for ugrad tuition Block grants for ugrad tuition subsidiessubsidies

U. S. peer reviewed grants to U. S. peer reviewed grants to young professors, not young young professors, not young studentsstudents

NSF, NIHNSF, NIH

Explanations of Rapid U. Explanations of Rapid U. S. Productivity Growth: S. Productivity Growth:

2000-20032000-2003 Unusual degree of downward pressure on Unusual degree of downward pressure on

profitsprofits Intangible capital became important after Intangible capital became important after

ICT boomICT boom– Productivity benefits of ICT investment could Productivity benefits of ICT investment could

have been delayedhave been delayed– Mismeasurement of timing of productivity Mismeasurement of timing of productivity

growthgrowth Outsourcing and changes in labor marketsOutsourcing and changes in labor markets Are payroll employment or real GDP Are payroll employment or real GDP

underestimated?underestimated?

Cost Cutting and the Cost Cutting and the Profit SqueezeProfit Squeeze

Productivity growth leads outputProductivity growth leads output Income shares reveal effect of productivity Income shares reveal effect of productivity

cycles on profitscycles on profits NIPA says profits doubled between ’92 and NIPA says profits doubled between ’92 and

’97, then declined through ‘00’97, then declined through ‘00 S&P reported profits grew by 70% between S&P reported profits grew by 70% between

’98 and ’00’98 and ’00– Shady accountingShady accounting– Low ratio of reported to operating earningsLow ratio of reported to operating earnings– Write-offs to correct for accounting and business Write-offs to correct for accounting and business

mistakesmistakes

Delay and Hidden Delay and Hidden CapitalCapital

O-S requires full productivity payoff O-S requires full productivity payoff occurs at moment computer is producedoccurs at moment computer is produced

David argues for delayDavid argues for delay– O-S overstates productivity post-’95 and O-S overstates productivity post-’95 and

understates ’01-’03understates ’01-’03– Comparison to electricity, 1880-1920Comparison to electricity, 1880-1920

Intangible capital complements ICT Intangible capital complements ICT capitalcapital– Measurement effectsMeasurement effects

Other Substantive Other Substantive ExplanationsExplanations

OutsourcingOutsourcing– Reduced cost benefitsReduced cost benefits– Some productivity benefitSome productivity benefit

Actual number of jobs outsourced unclearActual number of jobs outsourced unclear

Labor Market FlexibilityLabor Market Flexibility– Share of part- time and self-employed has Share of part- time and self-employed has

stayed constantstayed constant– Erosion of union membership and rise of Erosion of union membership and rise of

temp agencies is nothing newtemp agencies is nothing new

Four Reasons Why 2000-Four Reasons Why 2000-03 Productivity Growth 03 Productivity Growth

Should not be Should not be ExtrapolatedExtrapolated

#1 Profit Squeeze has been reversed#1 Profit Squeeze has been reversed #2 Intangible Capital Hypothesis; #2 Intangible Capital Hypothesis;

disequilibrium is being correcteddisequilibrium is being corrected #3 Diminishing returns: geometric growth of #3 Diminishing returns: geometric growth of

Moore’s law vs. limits of human brain and Moore’s law vs. limits of human brain and fingersfingers

#4 Jorgenson-Ho-Stiroh on Labor Quality#4 Jorgenson-Ho-Stiroh on Labor Quality– 1995-2001 0.38 percent contribution1995-2001 0.38 percent contribution– 2001-2011 0.162001-2011 0.16– 2011-2021 0.022011-2021 0.02

#5 What is the right time horizon for #5 What is the right time horizon for forecasting 10 years, 20 years, 75 years?forecasting 10 years, 20 years, 75 years?