Fiscal Sentiment and the Weak Recovery from the Great ... · the Great Recession: A Quantitative Exploration Finn E. Kydland University of California, Santa Barbara Carlos Zarazaga

Introduction Measurement Issues Model Calibration Numerical Experiments Results

Fiscal Sentiment and the Weak Recovery fromthe Great Recession: A Quantitative Exploration

Finn E. KydlandUniversity of California, Santa Barbara

Carlos ZarazagaFederal Reserve Bank of Dallas

Growth, Rebalancing, and Macroeconomic Adjustments after Large ShocksWorshop at Magyar Nemzeti Bank, Budapest, Hungary

September 20, 2013


Disclaimer

• The views expressed in this presentation are those of theauthors and do not necessarily reflect those of the FederalReserve Bank of Dallas, or the Federal Reserve System.


Motivation

0.7

0.9

1.1

1.3

1.5

1.7

1.9

2.1

2.3

2.5

19

74

- Q

11

97

5 -

Q1

19

76

- Q

11

97

7 -

Q1

19

78

- Q

11

97

9 -

Q1

19

80

- Q

11

98

1 -

Q1

19

82

- Q

11

98

3 -

Q1

19

84

- Q

11

98

5 -

Q1

19

86

- Q

11

98

7 -

Q1

19

88

- Q

11

98

9 -

Q1

19

90

- Q

11

99

1 -

Q1

19

92

- Q

11

99

3 -

Q1

19

94

- Q

11

99

5 -

Q1

19

96

- Q

11

99

7 -

Q1

19

98

- Q

11

99

9 -

Q1

20

00

- Q

12

00

1 -

Q1

20

02

- Q

12

00

3 -

Q1

20

04

- Q

12

00

5 -

Q1

20

06

- Q

12

00

7 -

Q1

20

08

- Q

12

00

9 -

Q1

20

10

- Q

12

01

1 -

Q1

UNITED STATES Output Stuck Below Its Pre-Great-Recession Trend

Real GDP

Log scale

12% gap


Motivation

Different accounts of the weak recovery:

• Abnormally large and persistent frictions in intermediation ofcapital.

• "Sunspots" or self-fulfilling loss of confidence.• "The Stock Market Crash of 2008 Caused the GreatRecession" (Roger Farmer, 2012).

• "Fiscal sentiment hypothesis":• Loss of confidence induced by prospect of higher taxes.• Fears justified by:

• Pre-existing structural U.S. fiscal imbalances aggravated bycrisis.

• Reinhart-Rogoff’s famous (infamous?) finding of negativecorrelation between growth and government debt.


Motivation








Motivation




• "Fiscal sentiment hypothesis":• Loss of confidence induced by prospect of higher taxes.

• Fears justified by:• Pre-existing structural U.S. fiscal imbalances aggravated bycrisis.



Motivation








Motivation








Motivation








US Rising Federal Noninterest Spending

0.0

5.0

10.0

15.0

20.0

25.0

30.0

1970 1980 1990 2000 2010 2020 2030 2040 2050

UNITED STATES Federal Government Noninterest Spending

in % of GDP

Total Noninterest Spending

Spending on Health Programs (Medicare and Medicaid)

Executed Projected


MotivationThe "fiscal sentiment" conjecture

Summarized by Robert E. Lucas, Jr. in Spring 2011 Wall StreetJournal interview:

"A healthy economy that falls into recession hashigher than average growth for a while and gets back tothe old trend line. We haven’t done that. I have plentyof suspicions but little evidence. I think people areconcerned about high tax rates... But none of this hashappened yet. You can’t look at evidence. The taxeshaven’t really been raised yet."


MotivationThe Fiscal Sentiment Conjecture

• What did Lucas mean by "You can’t look at the evidence"?

• Methodological challenge: Higher taxes not in place...yet.• "Peso problem" in interpreting the evidence.• Lucas himself helped develop "policy experiment" tools toovercome this diffi culty!

• No one has used them yet to explore the quantitativerelevance of the fiscal sentiment hypothesis.

• This is exactly what the paper sets out to do.



• What did Lucas mean by "You can’t look at the evidence"?• Methodological challenge: Higher taxes not in place...yet.

• "Peso problem" in interpreting the evidence.• Lucas himself helped develop "policy experiment" tools toovercome this diffi culty!





• What did Lucas mean by "You can’t look at the evidence"?• Methodological challenge: Higher taxes not in place...yet.• "Peso problem" in interpreting the evidence.

• Lucas himself helped develop "policy experiment" tools toovercome this diffi culty!





• What did Lucas mean by "You can’t look at the evidence"?• Methodological challenge: Higher taxes not in place...yet.• "Peso problem" in interpreting the evidence.• Lucas himself helped develop "policy experiment" tools toovercome this diffi culty!














Goal of the paper

Contribute to the debate on the causes behind the disappointingrecovery from the Great Recession by exploring the fiscal sentimenthypothesis quantitatively.


Preview of the results

Prospects of higher taxes matter more than critics of thehypothesis typically concede, but less than what its advocatestypically believe.


Analytical framework

• Neoclassical growth model.

• Why?

• "A healthy economy that falls into recession has higher thanaverage growth for a while and gets back to the old trend line."

• No reference to financial frictions in Lucas’s characterizationof the weak recovery.

• How far can fiscal sentiment hypothesis go without distortionsother than future higher taxes?

• Size of the "residual" potentially useful to infer the potentialquantitative role of the "missing" frictions (financial amongthem) in the weakness of the recovery.



• Neoclassical growth model.• Why?


































Measurement issues

• Skeptics of fiscal sentiment hypothesis argue no tax increasesof plausible magnitude can account for a 12% decline ofoutput from its pre-recession trend.

• They are right!• But has output declined from trend as much as 12%?


Measurement issues


• They are right!

• But has output declined from trend as much as 12%?


Measurement issues


• They are right!• But has output declined from trend as much as 12%?


Measurement issues

• Problem with measure of labor input consistent with theneoclassical growth model:

• It hasn’t been stationary, as it’s supposed to be along a"balanced-growth" path.


Measurement issues

• Problem with measure of labor input consistent with theneoclassical growth model:

• It hasn’t been stationary, as it’s supposed to be along a"balanced-growth" path.


0.23

0.24

0.25

0.26

0.27

0.28

0.29

0.3

19

59

19

61

19

63

19

65

19

67

19

69

19

71

19

73

19

75

19

77

19

79

19

81

19

83

19

85

19

87

19

89

19

91

19

93

19

95

19

97

19

99

20

01

20

03

20

05

20

07

20

09

LABOR INPUT Fraction of available time average household devoted to work

ht


Measurement issues

• Removal of non-stationarity reduces decline of output relativeto trend by 2/3!

• Fiscal sentiment hypothesis has a shot at accounting for thissmaller decline from trend.


Measurement issues

• Removal of non-stationarity reduces decline of output relativeto trend by 2/3!

• Fiscal sentiment hypothesis has a shot at accounting for thissmaller decline from trend.


0.7

0.9

1.1

1.3

1.5

1.7

1.9

2.1

2.3

2.5

19

74

- Q

1

19

75

- Q

1

19

76

- Q

1

19

77

- Q

1

19

78

- Q

1

19

79

- Q

1

19

80

- Q

1

19

81

- Q

1

19

82

- Q

1

19

83

- Q

1

19

84

- Q

1

19

85

- Q

1

19

86

- Q

1

19

87

- Q

1

19

88

- Q

1

19

89

- Q

1

19

90

- Q

1

19

91

- Q

1

19

92

- Q

1

19

93

- Q

1

19

94

- Q

1

19

95

- Q

1

19

96

- Q

1

19

97

- Q

1

19

98

- Q

1

19

99

- Q

1

20

00

- Q

1

20

01

- Q

1

20

02

- Q

1

20

03

- Q

1

20

04

- Q

1

20

05

- Q

1

20

06

- Q

1

20

07

- Q

1

20

08

- Q

1

20

09

- Q

1

20

10

- Q

1

20

11

- Q

1

UNITED STATES "A healthy economy that falls into recession has higher than average growth for a

while and gets back to the old trend line. We haven't done that..."

Real GDP

Log scale

12% gap


0.88

0.92

0.96

1

1.04

1.08

1.121

97

7

19

79

19

81

19

83

19

85

19

87

19

89

19

91

19

93

19

95

19

97

19

99

20

01

20

03

20

05

20

07

20

09

20

11

Detrended Privately Produced Output

- 4.0%

Privately produced output and TFP steady-state level


Measurement issues

• Discrepancy between "historical" trend and model-consistenttrend suggested need to be careful about mapping betweenvariables in the model and their empirical counterparts.

• "Private sector economy" approach in Gomme-Rupert (2000)particularly suitable to that end.

• Paper updates approach to incorporate latest NIPAmethodological changes.

• Conference participants will be spared the tedious steps,critical nevertheless for trusting the quantitative results of themodel.


Measurement issues






Measurement issues






Measurement issues






Government Budget Constraint

• Question of interest:• Can anticipated switch to a higher taxes regime account forweakness of the recovery?

• Can be answered on a first pass abstracting from governmentdebt dynamics:

• Balanced budget, additional revenues rebated as lump-sumtransfers.

• Quantitative discipline needed to limit size of expected taxincreases.




















Government Policies

• Trivial ones:• stochastic public sector labor input demand• iid shocks to share of value added by the public sector (used toinfer private sector output).

• Important one: anticipated switch to a higher taxes regime:{{τht+i , τkt+i}ji=0, {τht+j+n, τkt+j+n}∞

n=1

}t=s

;

τht+j+n > τht+i and/or τkt+j+n > τkt+i , for all i and n.


Government Policies

• Trivial ones:• stochastic public sector labor input demand• iid shocks to share of value added by the public sector (used toinfer private sector output).

• Important one: anticipated switch to a higher taxes regime:{{τht+i , τkt+i}ji=0, {τht+j+n, τkt+j+n}∞

n=1

}t=s

;

τht+j+n > τht+i and/or τkt+j+n > τkt+i , for all i and n.


Model

Stand-in household’s choice problem:

Max{ct ,lt ,kt+1}

E∞

∑t=s[β(1+ n)(1+ γ)α(1−σ)]t

[cαt (1− ht )1−α]1−σ − 1

1− σ

subject to:

ct + xt = (1− τht )wt (hprt + h

get + h

gct ) +

[rt − τkt (rt − δ)]kt + ckget + τt

(1+ n)(1+ γ)kt+1 = xt + (1− δ)kt

1 = lt + hprt + h

get + h

gct

government policies


Model

Representative firm’s choice problem:

Maxhprt , kt

[yprt − wthprt − rtkt ]

subject to:

yprt =1

e(1−θ)γtAeztkθ

t [eγthprt ]

1−θ,

wherezt = ρzt−1 + εt

• TFP long-run growth rate γ assumed deterministic (tocapture "rubber band" growth effect implied by Lucas.)


• Nominal variables deflated by implicit price deflator fornondurable consumption goods and services.

• Deflating procedure and Cobb-Douglas technology incorporateinvestment-specific technological progress in mannerconsistent with balanced growth.

• Depreciation rate should be interpreted as economicdepreciation rate.


Calibration• Parameter values calibrated using steady-state relationshipsand relevant averages for U.S. economy over period1977-2007.

• Important step, as decision rules will be computed withperturbation methods around the steady-state.

• Parameter calibrated using historical averages:

•

x/y (private sector investment-output ratio) 0.19δ (economic depreciation rate) 0.05gy (general government output absorption) 0.086vy (value added by government enterprises) 0.013τkt (capital income tax rate) 0.40τht (labor income tax rate) 0.23γ (private sector TFP annual growth rate) 0.7 %ky pr (private sector capital-output ratio) 2.7θ (private sector capital income share) 0.35





•






•



Calibration

• Utility parameter α particularly diffi cult to calibrate.

• Typical approach uses steady-state version of intratemporalmarginal rate of substitution between consumption and leisure:

α =1

1−hpr−hpuhpr

(1−τh)(1−θ)1+ vy − gy − x

ypr+ 1

• But... what is the stationary value of hpr ?


0.23

0.24

0.25

0.26

0.27

0.28

0.29

0.3

19

59

19

61

19

63

19

65

19

67

19

69

19

71

19

73

19

75

19

77

19

79

19

81

19

83

19

85

19

87

19

89

19

91

19

93

19

95

19

97

19

99

20

01

20

03

20

05

20

07

20

09

Fraction of available time average household devoted to work Population in working age (16-65 years old)

Hours actually worked (not paid)

HP filter

ht

Source: own calculations and Cociuba, Prescott, and Ueberfeldt (2012)


Calibration

Heroic decision:

h = h(HPF )2007 = 0.28045

hpr = hpr (HPF )2007 = 0.24519

hpu = h− hpr = 0.03526


CalibrationEffects of heroic decision:

0.2

0.21

0.22

0.23

0.24

0.25

0.26

19

77

19

79

19

81

19

83

19

85

19

87

19

89

19

91

19

93

19

95

19

97

19

99

20

01

20

03

20

05

20

07

20

09

20

11

Fraction of Time Spent Working in the Private Sector Observed and Adjusted

htpr

Data

Adjusted

calibrated value


CalibrationEffects of heroic decision:

0.88

0.92

0.96

1

1.04

1.08

1.12

19

77

19

79

19

81

19

83

19

85

19

87

19

89

19

91

19

93

19

95

19

97

19

99

20

01

20

03

20

05

20

07

20

09

20

11

Detrended Private Sector Output and Stochastic Technology Levels

TFP

ŷpr

- 4.0%

Private Sector Output and TFP steady-state level


Contrast with 12% output decline from trend without correctingfor non-stationarity of labor input

0.7

0.9

1.1

1.3

1.5

1.7

1.9

2.1

2.3

2.5

19

74

- Q

11

97

5 -

Q1

19

76

- Q

11

97

7 -

Q1

19

78

- Q

11

97

9 -

Q1

19

80

- Q

11

98

1 -

Q1

19

82

- Q

11

98

3 -

Q1

19

84

- Q

11

98

5 -

Q1

19

86

- Q

11

98

7 -

Q1

19

88

- Q

11

98

9 -

Q1

19

90

- Q

11

99

1 -

Q1

19

92

- Q

11

99

3 -

Q1

19

94

- Q

11

99

5 -

Q1

19

96

- Q

11

99

7 -

Q1

19

98

- Q

11

99

9 -

Q1

20

00

- Q

12

00

1 -

Q1

20

02

- Q

12

00

3 -

Q1

20

04

- Q

12

00

5 -

Q1

20

06

- Q

12

00

7 -

Q1

20

08

- Q

12

00

9 -

Q1

20

10

- Q

12

01

1 -

Q1

UNITED STATES Output Stuck Below Its Pre-Great-Recession Trend

Real GDP

Log scale

12% gap


The Productivity Puzzle

• TFP above trend while output below trend a rarity:

• Subject of "The Labor Productivity Puzzle," by McGrattanand Prescott.

• They would question TFP measure obtained in this paper(abstracts from intangible capital).

• RBC critics have always questioned fluctuations in Solowresiduals as measuring fluctuations in technology level.

• Paper agnostic on this issue: reports results with and withouttechnology shocks.



• TFP above trend while output below trend a rarity:• Subject of "The Labor Productivity Puzzle," by McGrattanand Prescott.























Calibration

• Benchmark: σ = 2 =⇒ Intertemporal Elasticity ofSubstitution = 0.5,

• less than the larger value of 1 proposed in the typicalcalibration of RBC models.

• In combination with calibrated value for α =⇒ Frischelasticity = 1.7,

• intermediate value between larger value of at least 3 proposedin the RBC literature and smaller value of 0.5 suggested bymicroeconomic studies for the intensive margin of labor supply.

• Results sensitive to the choice of these parameter values.


Numerical ExperimentsRestrictions on Tax Regime Change

• What higher tax regime quantitatively plausible to consider?

• Controversial issue.• Paper takes at face value assessment of Congressional BudgetOffi ce, an allegedly non-partisan agency.

• CBO Director testimony to Congress on September 2011:

• The U.S. must reduce fiscal deficits by at least $3.8 trillionover next decade.

• Paper takes this to mean: higher taxes must generateadditional annual revenues of $0.38 trillion, or 2.5% of currentGDP, for next ten years.

• Modest extra revenues of 0.3% of GDP thereafter (to coverrising costs of government-sponsored health care programs.)



• What higher tax regime quantitatively plausible to consider?• Controversial issue.

• Paper takes at face value assessment of Congressional BudgetOffi ce, an allegedly non-partisan agency.







• What higher tax regime quantitatively plausible to consider?• Controversial issue.• Paper takes at face value assessment of Congressional BudgetOffi ce, an allegedly non-partisan agency.















• CBO Director testimony to Congress on September 2011:• The U.S. must reduce fiscal deficits by at least $3.8 trillionover next decade.

















Search over tax rates that can deliver targeted additional revenues,within the following class:{{τht+i , τkt+i}3i=0, {τht+3+i , τkt+3+i}10i=1, {τht+13+i , τkt+13+i}∞

i=1

}t=2009

τh2009+i = 0.23; τk2009+i = 0.40 for 0 ≤ i ≤ 3,τh2013+i = τh2013; τk2013+i = τk2013 for 0 ≤ i ≤ 9,τh2023+i = τh2023; τk2023+i = τk2023 for all i > 0.

• Higher labor income taxes and higher capital income taxesconsidered one at a time:

• As in Christiano, Eichenbaum, and Rebelo 2011 JPE paper onthe size of fiscal multipliers.


Numerical ExperimentsComputational approach

• Innovations (iid shocks) to all stochastic processes are setequal to zero

• Computation uses second order perturbation approximationaround the steady state.

• Why?• Paper compares data and model predictions for level ofvariables.

• Ignoring precautionary savings could bias results in favor of thefiscal sentiment hypothesis.


Numerical ExperimentsHigher labor income tax regime

• Standard arguments suggest anticipated higher labor incometaxes regime cannot do the job:

• Higher taxes on labor income tomorrow should inducehouseholds to work harder today.

• Output should be above trend before the regime changematerializes.

• For the sake of completion, analyze this regime anyway.





















Labor tax rates implied by additional revenues target:

τh2009+i = 0.23; τk2009+i = 0.40 for 0 ≤ i ≤ 3,τh2013+i = 0.27; τk2013+i = 0.40 for 0 ≤ i ≤ 9,τh2023+i = 0.24; τk2023+i = 0.40 for all i > 0.


-2.10

-2.00

-1.90

-1.80

-1.70

-1.60

-1.50

20

03

20

05

20

07

20

09

20

11

20

13

20

15

20

17

20

19

20

21

20

23

20

25

20

27

20

29

20

31

20

33

20

35

PRIVATE GROSS DOMESTIC INVESTMENT Data and Model Predictions

Fully Anticipated Switch to Higher Labor Income Tax Regime (detrended levels)

IES = 0.5

steady-state level for low labor income tax rate regime

Ln(xt)

Data 2nd order perturbation without technology shocks


-1.52

-1.50

-1.48

-1.46

-1.44

-1.42

-1.40

-1.38

20

03

20

05

20

07

20

09

20

11

20

13

20

15

20

17

20

19

20

21

20

23

20

25

20

27

20

29

20

31

20

33

20

35

LABOR INPUT (fraction of time spent working)

Data and Model Predictions Fully Anticipated Switch to Higher Labor Income Tax Regime

IES = 0.5

Ln(htpr)


2nd order perturbation without technology shocks

Data


-0.37

-0.35

-0.33

-0.31

-0.29

-0.27

-0.25

-0.23

-0.21

-0.19

-0.17

20

03

20

05

20

07

20

09

20

11

20

13

20

15

20

17

20

19

20

21

20

23

20

25

20

27

20

29

20

31

20

33

20

35

CONSUMPTION Data and Model Predictions

Fully Anticipated Switch to Higher Labor Income Tax Regime (detrended levels)

IES = 0.5

Ln(ct)

Data

Data: C + NX




Numerical ExperimentsHigher capital income tax regime

• Future higher tax rates on capital income can do the job intheory.

• Why fear higher taxes on just capital income?• Because incentives of democratically elected offi cials is tocorrect structural fiscal imbalances with unanticipated taxationof capital (time inconsistency) rather than with entitlementreforms.

• Do these fears matter quantitatively?




• Why fear higher taxes on just capital income?

• Because incentives of democratically elected offi cials is tocorrect structural fiscal imbalances with unanticipated taxationof capital (time inconsistency) rather than with entitlementreforms.













23.0

23.5

24.0

24.5

25.0

25.5

26.0

26.5

20

13

20

15

20

17

20

19

20

21

20

23

20

25

20

27

20

29

20

31

20

33

20

35

20

37

20

39

20

41

20

43

20

45

20

47

20

49

Model Economy Government Revenues

Fully Anticipated Switch to Higher Capital Income Tax Regime (without technology shocks)

IES = 0.5

Note: total output corresponds to the model economy total output under the initial tax regime. All calculations correspond to the 2nd order perturbation.

Revenues under initial tax regime: τk = 0.40, τh = 0.23

τk2013-2022 = 0.61

Revenues after switch to higher capital income tax regime τk

2023 on = 0.45

Targeted revenues

% GDP


How plausible are tax hikes of this magnitude?

• Higher capital income tax regime implies a 20 percentagepoints jump in the tax rate (from 40% to 61%).

• Similar jump in 2013 under current law for:

• top dividend tax rate (from 15% to 43.4%).• estate tax rate (from 35% to 55%).




• Similar jump in 2013 under current law for:

• top dividend tax rate (from 15% to 43.4%).• estate tax rate (from 35% to 55%).




• Similar jump in 2013 under current law for:• top dividend tax rate (from 15% to 43.4%).

• estate tax rate (from 35% to 55%).




• Similar jump in 2013 under current law for:• top dividend tax rate (from 15% to 43.4%).• estate tax rate (from 35% to 55%).


-2.20

-2.10

-2.00

-1.90

-1.80

-1.70

-1.60

-1.50

20

03

20

05

20

07

20

09

20

11

20

13

20

15

20

17

20

19

20

21

20

23

20

25

20

27

20

29

20

31

20

33

20

35


Fully Anticipated Switch to Higher Capital Income Tax Regime (detrended levels)

IES = 0.5

Ln(xt)

steady-state level for low capital income tax rate regime

2nd order perturbation solution

2nd order perturbation without technology shocks Perfect foresight solution

without technology shocks

Data


-1.52

-1.50

-1.48

-1.46

-1.44

-1.42

-1.40

-1.38

20

03

20

05

20

07

20

09

20

11

20

13

20

15

20

17

20

19

20

21

20

23

20

25

20

27

20

29

20

31

20

33

20

35


Data and Model Predictions Fully Anticipated Switch to Higher Capital Income Tax Regime

IES = 0.5 Ln(htpr)


Data 2nd order perturbation solution 2nd order perturbation

without technology shocks

Perfect foresight solution without technology shocks


Results for labor input better than in latest generation of complex financial

frictions models, such as Jermann and Quadrini (AER, February 2012):


• Generic problem of models with financial frictions:

• hard time accounting for weakness of the recovery becausewidely used indicators of financial stress are back to normallevels.


• Generic problem of models with financial frictions:

• hard time accounting for weakness of the recovery becausewidely used indicators of financial stress are back to normallevels.


-200

-100

0

100

200

300

400

2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

Daily,

basis pts.

Financial frictions indicators back to normal levels in the recovery

3-month LIBOR minus

3-month OIS

3-month T-Bill minus

3-month OIS

TED

spread

Lehman

collapse

4/19/2013


-0.08

-0.06

-0.04

-0.02

0

0.02

0.04

20

03

20

05

20

07

20

09

20

11

20

13

20

15

20

17

20

19

20

21

20

23

20

25

20

27

20

29

20

31

20

33

20

35

20

37

20

39

20

41

20

43

20

45

20

47

20

49

PRIVATE SECTOR OUTPUT Data and Model Predictions


IES = 0.5



2nd order perturbation solution with technology shocks

Data

Ln(ytpr)


-0.37

-0.35

-0.33

-0.31

-0.29

-0.27

-0.25

-0.23

-0.21

-0.19

-0.17

20

03

20

05

20

07

20

09

20

11

20

13

20

15

20

17

20

19

20

21

20

23

20

25

20

27

20

29

20

31

20

33

20

35

20

37

20

39

20

41

20

43

20

45

20

47

20

49



IES = 0.5

Ln(ct)

Data

Data: C + NX





Above-trend consumption prediction

• Consumption above steady-state prior to the regime change.

• Most other interpretations of the weak recovery predict belowsteady-state consumption.

• Dynamics of consumption potentially critical to discriminatebetween alternative interpretations of the weak recovery.



• Consumption above steady-state prior to the regime change.• Most other interpretations of the weak recovery predict belowsteady-state consumption.




• Consumption above steady-state prior to the regime change.• Most other interpretations of the weak recovery predict belowsteady-state consumption.



Above-trend consumption predictionEconomic intuition

• Households made prior decisions assuming continuation of lowcapital income tax regime.

• Suddenly fear switch to higher capital income tax regime:• Have too much capital!• Let capital stock depreciate faster, before taxman gets to it.

• No need to produce as many investment goods: work less.• Devote more output to consumption, less to investment.• Consumption and leisure shifted from future to present.• How much?

• Depends on Intertemporal Elasticity of Substitution. Resultsare sensitive to this parameter value.




• Suddenly fear switch to higher capital income tax regime:

• Have too much capital!• Let capital stock depreciate faster, before taxman gets to it.













• No need to produce as many investment goods: work less.• Devote more output to consumption, less to investment.

• Consumption and leisure shifted from future to present.• How much?






• No need to produce as many investment goods: work less.• Devote more output to consumption, less to investment.• Consumption and leisure shifted from future to present.

• How much?















Summary of findings

• Quantitative exploration of fiscal sentiment hypothesis withfrictionless neoclassical growth model delivers mixed(confusing?) results.

• Cannot account for weakness of the recovery if higher taxesanticipated to fall on labor income.

• Can account for weakness of the recovery if higher taxesanticipated to fall on capital income.

• How much depends on whether technology level fluctuate asmuch as suggested by Solow residuals:


Summary of findings






Summary of findings






Summary of findings






Summary of findingsHigher capital income tax rates scenario:

• If true TFP relatively unchanged over the cycle, as RBCcritics maintain, fiscal sentiment hypothesis accounts for:

• more than all of the investment decline from pre-recessiontrend.

• at least half of the labor input decline from pre-recession trend.• almost all of the output decline from pre-recession trend.

• If TFP fluctuates over the cycle as much as suggested bySolow residuals, fiscal sentiment hypothesis accounts for:

• three-fourths of investment decline from pre-recession trend.• one third of labor input decline from pre-recession trend.• not much of output decline from pre-recession trend.

• In both cases, fiscal sentiment hypothesis prediction ofabove trend consumption during the recovery seeminglyvalidated by the data.


















Next steps

• Given importance of the dynamics of consumption, improvecorrespondence between consumption in the model and itsempirical counterpart in the data.

• Results for the higher capital income tax case "buried" inChristiano, Eichenbaum, and Rebelo’s 2011 paper suggestincorporation of fiscal sentiment hypothesis in their modelwith financial frictions could account for a non-negligiblefraction of the labor input gap "remainder"...

• ... perhaps preserving critical prediction of above trendconsumption.


Next steps





Next steps





WINE TASTING COMING SOON


22.5

23.5

24.5

25.5

26.5

27.5

20

13

20

15

20

17

20

19

20

21

20

23

20

25

20

27

20

29

20

31

20

33

20

35

20

37

20

39

20

41

20

43

20

45

20

47

20

49

Model Economy Government Revenues

Fully Anticipated Switch to Higher Capital Income Tax Regime (without technology shocks)

IES = 1.0

Note: total output corresponds to the model economy total output under the initial tax regime. All calculations correspond to the 2nd order perturbation.

Targeted revenues

% GDP

τk2013-2022 = 0.72

Revenues after switch to higher capital income tax regime τk

2023 on = 0.45

Revenues under initial tax regime: τk = 0.4, τh = 0.23


-1.58

-1.56

-1.54

-1.52

-1.50

-1.48

-1.46

-1.44

-1.42

-1.40

-1.38

-1.36

20

03

20

05

20

07

20

09

20

11

20

13

20

15

20

17

20

19

20

21

20

23

20

25

20

27

20

29

20

31

20

33

20

35



IES = 1.0 Ln(htpr)


Data





-0.18

-0.16

-0.14

-0.12

-0.1

-0.08

-0.06

-0.04

-0.02

0

0.02

0.04

20

03

20

05

20

07

20

09

20

11

20

13

20

15

20

17

20

19

20

21

20

23

20

25

20

27

20

29

20

31

20

33

20

35

20

37

20

39

20

41

20

43

20

45

20

47

20

49

PRIVATE SECTOR OUTPUT Data and Model Predictions


IES = 1.0




Data

Ln(ytpr)


-4.00

-3.50

-3.00

-2.50

-2.00

-1.50

20

03

20

05

20

07

20

09

20

11

20

13

20

15

20

17

20

19

20

21

20

23

20

25

20

27

20

29

20

31

20

33

20

35



IES = 1.0

Ln(xt)





Data


-0.47

-0.45

-0.43

-0.41

-0.39

-0.37

-0.35

-0.33

-0.31

-0.29

-0.27

-0.25

-0.23

-0.21

-0.19

20

03

20

05

20

07

20

09

20

11

20

13

20

15

20

17

20

19

20

21

20

23

20

25

20

27

20

29

20

31

20

33

20

35

20

37

20

39

20

41

20

43

20

45

20

47

20

49



IES = 1.0

Ln(ct)

Data

Data: C + NX





• Model with IES =1 and targeted revenues criterion aboveproduces unreasonable results.

• Alternative approach:

• Search over capital income tax rate that approximatesdynamics of investment in the data.

• Check predictions for labor input.



• Alternative approach:

• Search over capital income tax rate that approximatesdynamics of investment in the data.




• Alternative approach:• Search over capital income tax rate that approximatesdynamics of investment in the data.




• Alternative approach:• Search over capital income tax rate that approximatesdynamics of investment in the data.



-2.20

-2.10

-2.00

-1.90

-1.80

-1.70

-1.60

-1.50

20

03

20

05

20

07

20

09

20

11

20

13

20

15

20

17

20

19

20

21

20

23

20

25

20

27

20

29

20

31

20

33

20

35

IES = 1.0 PRIVATE GROSS DOMESTIC INVESTMENT


(detrended levels)

τk2013-2022 = 0.54

Ln(xt)

Data

steady-state level for low capital income tax rate regime (τk = 0.4, τh = 0.23 )





-1.52

-1.50

-1.48

-1.46

-1.44

-1.42

-1.40

-1.38

20

03

20

05

20

07

20

09

20

11

20

13

20

15

20

17

20

19

20

21

20

23

20

25

20

27

20

29

20

31

20

33

20

35

IES = 1.0 LABOR INPUT

(fraction of time spent working) Data and Model Predictions

Fully Anticipated Switch to Higher Capital Income Tax Regime τk

2013-2022 = 0.54

Ln(htpr)

Data


steady-state level for low capital income tax rate regime (τk = 0.4, τh = 0.23 ) 2nd order perturbation

solution


Documents

Fiscal Sentiment and the Weak Recovery from the Great ... · the Great Recession: A Quantitative Exploration Finn E. Kydland University of California, Santa Barbara Carlos Zarazaga