The Financial Crisis and Macroeconomic Policy: Four Years On

 John B. Taylor

Stanford University

MONFISPOL Conference   

September 19, 2011

The Revival of Keynesian Discretionary Fiscal Policy in the 2000s

• Economic Growth and Tax Relief Reconciliation Act of 2001 – Refund checks; first installment of 2001 tax rate cuts

• Economic  Stimulus Act of 2008  (February)– Rebate checks and credits

• American Recovery and Reinvestment Act of 2009 (February)– One-time payments, withholding change, refunds– More government spending too

• Miscellaneous interventions in 2009-10– Cash for clunkers program – First time home buyers program

• Tax Relief , Unemployment Insurance Reauthorization, and Job Creation Act of 2010 (December))– Temporary cut in payroll tax

• American Jobs Act of 2011? 

Keynesian Discretionary Fiscal Policy First Became Popular in the ‘60s & ‘70s

• First in academia in the 1950s and 1960s – Arguments appeared in the major textbooks (Samuelson).  – Keynesian econometric models

• Then in practice: 1962 Economic Report of the President– “The task of economic stabilization cannot be left entirely to built-in stabilizers,” the report warned. “Discretionary budget policy, e.g. changes in tax rates or expenditure programs, is indispensible—sometimes to reinforce, sometimes to offset, the effects of the stabilizers.” 

– investment tax credit (1962), tax surcharge (1968)– tax rebates (1975)– countercyclical grants to states for infrastructure (1977-78)

• Keynesian discretionary policy continued to the late 1970s. 

 

Keynesian Policy Fell Out of Favor in ‘80s & ‘90s• Research raised doubts about discretionary policy

– Lucas and Sargent “After Keynesian Economics”– Gramlich “the general idea of stimulating the economy through state and local governments is probably not a very good one”

• Soon automatic stabilizers dominated the budget cycle• Bush 41: proposed tiny stimulus package in 1992

– Shift $10 billion in G from future to the present – did not pass the Congress  

• Clinton: proposed tiny stimulus in 1993 – $16 billion more G– did not pass the Congress. 

• Eichenbaum (1997) “there is now widespread agreement that countercyclical discretionary fiscal policy is neither desirable nor politically feasible.”  

The Basic Model: A decline in I causes the aggregate expenditure line to shift down 

45-degree line

New E line

Original E line

Original income

level

New income

level

INCOME OR REAL GDP

SPENDING

Income or real GDP falls by this amount (more than by amount I falls ).

I falls by this amount

Original point of spending

balance

New point of spending balance

25_10

Countercyclical discretionary fiscal policy:Increase in G raises GDP depending on size of the 

multiplier and amount of crowding out 

45-degree line

INCOME OR REAL GDP

SPENDING

G rises

25_10

But Macro Models Differ Greatly• Romer and Bernstein (Jan 2009) used estimated old Keynesian models (without RE) to predict ARRA effect – Large multipliers, around 1.5. 

• Cogan, Cwik, Taylor and Wieland (Feb 2009) used estimated New Keynesian model to predict ARRA effect– Much smaller multipliers, around 0.5. 

•  What not use these existing macro models for the evaluation of actual packages? – Because they simply repeat the same prediction story over again. 

• So you learn virtually nothing

Consider two models relating stimulus S to output Y. Model A is Y= αS +Z Model B is Y = Zwhere Z is a shock and α =1.5

Now, suppose that a stimulus is enacted: S = 2 and Y decreases by -1 According to Model A , Z = - 4 According to Model B, Z= -1

Now consider policy evaluation with counterfactual: S=0

Economists using Model A say: Just as we predicted, the stimulus package worked. Without it, Y would have fallen to -4 rather than -1. The decline in

output would have been 4 times as deep, a Great Depression 2.0.

Economists using Model B say Just as we predicted the stimulus package did not work.

A Stylized Illustration

-1

0

1

2

3

4

5

6

2009 2010

With stimulus

If nostimulus

-1

0

1

2

3

4

5

6

2009 2010

With stimulus

If nostimulus

“The accumulation of hard data and real-life experience has allowed more dispassionate analysts to reach a consensus that the stimulus package, messy as it is, is working” New York Times November 12, 2009

New KeynesianSmets - ECB

Robert BarroHarvard

A Less Stylized Illustration

Use a Direct Approach

• Micro data (used in 2001, 2008)– Shapiro and Slemrod (2003, 2009), – Johnson, Parker, Souleles (2006)– Parker, Souleles, Johnson, and McClelland (2009). 

• Macro data– Special BEA satellite account– “Personal Income and Output” (monthly to mid ’09) – “Effect of the ARRA on Selected Federal Government Sector Transactions”  (quarterly)

Monthly Data on Rebate Payments in 2001 and 2008 ($ billions, annual rates)

2001 2008April 0 23.3May 0 577.1June 0 334.4July 95.1 164.1August 223.1 12.4September 144.9 0October 2.5 0

Temporary stimulus meets permanent income hypothesis

Temporary  stimulus meetspermanent income hypothesis again

Cash for clunkers: incentives really matter

Based on Mian and Sufi (2010)

0

40

80

120

160

200

240

280

320

01 02 03 04 05 06 07 08 09 10

Effects of Three Stimulus Packages on Disposable Personal Income

2008

20092001

Billions of dollars

Quarterly Data

9,600

10,000

10,400

10,800

11,200

11,600

12,000

07Q1 07Q3 08Q1 08Q3 09Q1 09Q3 10Q1 10Q3 11Q1

Billions of dollars

Disposable personal income with stimulus

and without stimulus

Personal consumption expenditures

Quaterly disposable personal income, with andwithout stimulus, and personal consumption

Quarterly PCE Regressions With and Without Stimulus Payments   (1)   (2)   (3)

Disposable Personal  .817 ---- -----Income (40.9)

Disposable PersonalIncome--Without Stimulus ---- .857 .851

(73.0) (60.4)

Stimulus Payments ----- ----- 0.128(0.81)

Oil Price ($/bbl lagged 2 quarters) -2.41 -2.55 -2.55(-4.71) (-4.14) (-4.61)

Net Worth (lagged 2 quarters) .021 .017 .018(8.53) (7.32) (7.97)

Standard error of regression 76.9 65.8 66.3

0

40

80

120

160

200

240

280

09Q1 09Q2 09Q3 09Q4 10Q1 10Q2 10Q3 10Q4 11Q1

Billions of dollars(annual rates)

Major Federal Budget Categories of ARRA

- Temporary transfers and tax credits to persons

- Grants to state and local governments

- Federal government consumption- Federal government investment

500

600

700

800

900

1,000

1,100

1,200

1,300

2000 2002 2004 2006 2008 2010

Billions of dollars

With ARRA

Without ARRA

Federal Government Purchases

Effect of ARRA on Federal Government Purchases

1,200

1,400

1,600

1,800

2,000

2,200

2,400

2000 2002 2004 2006 2008 2010

Billions of dollars

Total Receipts ofState and LocalGovernments

With ARRA grants

Without ARRA grants

Effect of ARRA on Receipts of State and Local Governments

1,100

1,200

1,300

1,400

1,500

1,600

1,700

1,800

1,900

2000 2002 2004 2006 2008 2010

Billions of dollars

State and Local Government Purchases: 2000.1 - 2011.1

200

250

300

350

400

450

500

00 01 02 03 04 05 06 07 08 09 10

Billions of dollars

State and Local Government ExpendituresOther Than for Purchases of Good and Services

20

40

60

80

100

120

140

160

180

2000 2002 2004 2006 2008 2010

Net Borrowing By State and Local Governments

Billions of dollars

-150

-100

-50

0

50

100

150

2009Q1 2009Q3 2010Q1 2010Q3 2011Q1

Billions of dollars(annual rates)

ARRA Grants

Government purchases

Borrowing (net)

Other expenditures

ARRA Grants and State and Local Budgets(change from 2008.4 when ARRA grants were zero)

State and local government budget constraint Gt + Et + Lt = Rt + At

whereG = Government purchases of goods and servicesE = Expenditures other than for the purchasesL = Lending or borrowing (-), netA= ARRA grants (exogenous)R = Revenues excluding ARRA grants (exogenous)

Estimated 3-equation system (1969Q1- 2011Q1)

Gt = 3.86 + 0.864Gt-1 + 0.124Rt - 0.114At

Et = -3.83+ 0.818Et-1 + 0.0398Rt + 0.113At

Lt = .0321 - 0.864Gt-1 - 0.818Et-1 + .836Rt + 1.001At

0

20

40

60

80

100

120

140

07Q1 07Q3 08Q1 08Q3 09Q1 09Q3 10Q1 10Q3 11Q1

Billions of dollars

ARRA grants

Counterfactual

Actual and Counterfactual ARRA grantsto State and Local Governments

0

40

80

120

160

200

240

07Q1 07Q3 08Q1 08Q3 09Q1 09Q3 10Q1 10Q3 11Q1

Billions of dollars

Counterfactualsimulation

Dynamic simulation

Historical data

Borrowing (net) by State and Local Governments:Historical and Counterfactual without ARRA

2,000

2,050

2,100

2,150

2,200

2,250

2,300

2,350

07Q1 07Q3 08Q1 08Q3 09Q1 09Q3 10Q1 10Q3 11Q1

Counterfactual simulationDynamic simulationHistorical data

Billions of dollars

Total Expenditures by State and Local Governments:Historical and Counterfactual without ARRA

360

380

400

420

440

460

480

500

07Q1 07Q3 08Q1 08Q3 09Q1 09Q3 10Q1 10Q3 11Q1

Billions of dollars

Dynamic simulation Historical

data

Counterfactualsimulation

Other Expenditures by State and Local Governments:Historical and Counterfactual without ARRA

1,640

1,680

1,720

1,760

1,800

1,840

1,880

07Q1 07Q3 08Q1 08Q3 09Q1 09Q3 10Q1 10Q3 11Q1

Billions of dollars

Counterfactualsimulation

Dynamic simulation

Historical data

Purchases of Goods and Services by State and Local Governments:Historical and Counterfactual without ARRA

The Plausibility of the Counterfactual• Weren’t many states borrowing constrained after the crisis?• Not clear but in any case Fed’s Flow of Funds data show that that net borrowing would have increased even with such borrowing constraints.  

• Net borrowing = net increase in liabilities - net acquisition of financial assets 

• Net borrowing = - $118 billion from 2008 to 2010. Net increase in liabilities = $53 billionNet acquisition of financial assets = $171 billion.

• Thus state and local government added significantly to their financial assets as ARRA grants came in.  

• With no ARRA they would not have done so.

Why the Negative Effect on Purchases?• “Other expenditures” consist largely of Medicaid, TANF, and other transfer programs

• ARRA conditioned states’ receipt of additional Medicaid grants on their not reducing benefits or restricting eligibility rules

• In some states, this meant undoing benefit reductions or eligibility restrictions that were implemented in the previous 6 months  – July 1, 2008 is the date in Section 5001 of ARRA

• This “hold-harmless” provision, may have forced states to reallocate funds that would have been used for purchases

Test By Splitting ARRA Grants into Medicaid and Other

0

20

40

60

80

100

120

140

07Q1 07Q3 08Q1 08Q3 09Q1 09Q3 10Q1 10Q3 11Q1

Total ARRA grants

Medicaid

Other

Billions of dollars

Dependent VariablesG E L

Constant 3.356 -2.471 2.691(3.5) (-2.1) (1.3)

G(-1) 0.882 ------- -0.877(17.8) (-12.8)

E(-1) ------ 0.875 -0.734(20.2) (-11.0)

R 0.108 0.028 0.829(2.6) (3.2) (13.8)

M -0.318 0.129 1.200(-2.3) (2.6) (6.7)

N -0.002 0.076 0.851(-0.03) (1.6) (13.0)

R² 0.99 0.99 0.95

Regressions with ARRA grants split into Medicaid (M) and other (N)

Net Effect on Federal, State& Local Government Purchases

• If government purchases have a greater impact on GDP than temporary transfers—which the permanent income theory predicts—then ARRA could have had a negative effect 

• According to the simulations the cumulative negative effect on state and local government purchases was $85 billion (341/4). Larger than the $30 billion (119/4) cumulative positive effect of ARRA on federal government purchases.

-8

-6

-4

-2

0

2

4

6

07Q1 07Q3 08Q1 08Q3 09Q1 09Q3 10Q1 10Q3

Percent, annual rate

Growth rate of real GDP

Contribution ofgovernment purchases

Cross check on GDP growth and G-contribution 

Parrallel Developments in Monetary Policy

Toward Discretion: ‘60s & ‘70s• Fed did not follow Milton Friedman’s rules message “of setting itself a steady course and sticking to it.” (AEA 1968)

• 1965-1970s saw a series of boom-bust cycles in monetary policy with inflation rising steadily higher at each cycle. 

• The wage and price controls of the 1970s – Epitome of interventionist policy– Defended by Fed Chair Burns: “wage rates and prices no longer respond as they once did to the play of market forces.”  

• Insider reports showed little strategy or systematic thinking• Econometric studies show that Fed’s responses to inflation were unstable over time in the 1970s – Rising implicit inflation target

• Not rule-like as it would become in the 1980s and 1990s. 

0

2

4

6

8

10

12

14

56 58 60 62 64 66 68 70 72 74 76 78 80 82 84

CPI InflationLivingston Survey

U.S. Inflation and Expected Inflation 1965-1980

From “Has the Fed Gotten Tougher on Inflation?” The FRBSF Weekly  Letter, March 31, 1995, by John P Judd and Bharat Trehan  of the San Francisco Fed 

1965-79

1965-1980: monetary policy not well described by a rules-based price stability objective

Toward Rules in the ‘80s –’90s• Dramatic shift of policy under Volcker in 1979-87.   

– Volcker (1983)  “We have…gone a long way toward changing the trends of the past decade and more.”  

• Policy continued in ‘80s and ‘90s under Greenspan.    • Additional evidence of more rules-based policy 

– more predictable and transparent decision-making process– focus on expectations of future policy actions. – announcing interest rate decisions when making them. 

• Transcripts of the FOMC in the 1990s, many references to rules.  

• Actual monetary policy closer to simple policy rule   

From “Has the Fed Gotten Tougher on Inflation?” The FRBSF Weekly  Letter, March 31, 1995, by John P Judd and Bharat Trehan  of the San Francisco Fed 

1987-92

1993-94

1965-79

Monetary policy gets more predictable, inflation targets, rules-based

Illustrative monetary policy chart from St Louis FedFebruary 2007, Bill Poole  

 • Interest rates too low for too long in 2003-05• On-again, off-again bailouts financed by central bank’s balance sheet– on for BSC creditors’ bailout, off for Lehman creditors’ bailout, on for AIG creditors’ bailout, off for TARP role out

• Government regulators and supervisors deviated from sound regulatory rules, especially at large banks

The Swing Away from Rules in Recent Years

Illustrative monetary policy chart from St Louis FedFebruary 2007, Bill Poole  

Illustrative monetary policy chart from St Louis FedFebruary 2007, Bill Poole  

(2000-2009)

Chart from Kansas City Fed, 2009, Tom Hoenig

0

2

4

6

8

10 2009Q3 - 2011Q1 1983Q1 - 1984Q3

PercentReal GDP growth

1 2 3 4 5 6 7

The Recovery in Historical Context

Quarter since recession ended

2

3

4

5

6

7

8

9

10

11

50 55 60 65 70 75 80 85 90 95 00 05 10

Percent

The Unemployment Rate                                                   

2

3

4

5

6

7

8

9

10

11

50 55 60 65 70 75 80 85 90 95 00 05 10

Percent More intervention,  

The Unemployment Rate                                                    

2

3

4

5

6

7

8

9

10

11

50 55 60 65 70 75 80 85 90 95 00 05 10

Percent More intervention, more unemployment 

The Unemployment Rate                                                    

2

3

4

5

6

7

8

9

10

11

50 55 60 65 70 75 80 85 90 95 00 05 10

Percent More intervention, more unemployment 

LessIntervention,

The Unemployment Rate                                                    

2

3

4

5

6

7

8

9

10

11

50 55 60 65 70 75 80 85 90 95 00 05 10

Percent More intervention, more unemployment 

Lessintervention,lessunemployment

The Unemployment Rate                                                    

2

3

4

5

6

7

8

9

10

11

50 55 60 65 70 75 80 85 90 95 00 05 10

Percent More intervention, more unemployment 

Lessintervention,lessunemployment

Moreintervention, 

The Unemployment Rate                                                 

2

3

4

5

6

7

8

9

10

11

50 55 60 65 70 75 80 85 90 95 00 05 10

Percent More intervention, more unemployment 

Lessintervention,lessunemployment

Moreintervention, more unemployment

The Unemployment Rate and Unintended Consequences

Conclusion• Revival of discretionary policy has been ineffective• Fiscal stimulus packages

– People largely saved the transfers and tax rebates  – Federal government increased purchases by a tiny amount– State and local governments used stimulus grants to reduce borrowing rather than increase expenditures, and they shifted expenditures away from purchases 

– The results do not support the view that things would have been worse

• Provide evidence against “Model A”• Plus counterfactual simulations

• Parallel developments for monetary policy• Results are consistent with consensus prior to the revival, based on experience of 30 years ago