L ECTURE 6 The Effects of Fiscal Changes: Fiscal Consolidations - - PowerPoint PPT Presentation

LECTURE 6

The Effects of Fiscal Changes: Fiscal Consolidations September 28, 2016

Economics 210c/236a Christina Romer Fall 2016 David Romer

• I. OVERVIEW OF THE IMPACT OF FISCAL

CONSOLIDATIONS

How could fiscal contractions be expansionary?

• Wealth effect: With Ricardian consumers, a

decrease in G makes people expect lower future

• taxes. As a result, wealth rises, consumption rises,

and labor supply falls. Effects on Y could be positive if prices are sticky.

• Confidence effect: If budget problems are severe,

dealing with them may prevent having to take more extreme measures later on. Thus, consolidation can have positive confidence effects on C and I.

How could fiscal contractions be expansionary? (continued)

• Interest rate effect: Fiscal consolidations may lower

risk premium and so lower long rates. This may raise both I and C.

• Coincidence: Budget problems are a symptom of

dysfunctional government. Fiscal consolidation is a sign that the government is functioning, and so may be correlated with other measures that are good for growth (i.e. relationship could be present but not causal).

• II. GIAVAZZI AND PAGANO: “CAN SEVERE FISCAL

CONTRACTIONS BE EXPANSIONARY? TALES OF TWO SMALL EUROPEAN COUNTRIES”

Giavazzi and Pagano’s Regression

Possible Omitted Variable Bias

• Cyclical adjustment may not fully deal with cyclicality
• f revenues.
• Countercyclical discretionary policy.
• Other policies (like labor market and trade) may be

correlated with fiscal reforms.

From: Giavazzi and Pagano, “Can Severe Fiscal Contractions be Expansionary?”

From: Giavazzi and Pagano, “Can Severe Fiscal Contractions be Expansionary?”

Research Strategy

• Look at Denmark in early 1980s and Ireland in late

1980s.

• Is this a sensible research strategy?
• What might be a more sensible strategy?
• What are they trying to learn from these case

studies?

From: Giavazzi and Pagano, “Can Severe Fiscal Contractions be Expansionary?”

From: Giavazzi and Pagano, “Can Severe Fiscal Contractions be Expansionary?”

From: Giavazzi and Pagano, “Can Severe Fiscal Contractions be Expansionary?”

Alesina and Ardagna’s Measure of Fiscal Consolidations

• A year when the cyclically adjusted primary balance

(CAPB) improves by at least 1.5% of GDP.

• Primary balance is the budget position net of interest

payments.

• Cyclically-adjust the budget data using simple

regression against the unemployment rate. (CBO and OECD uses more detailed methods.)

From: Alesina and Ardagna, “Large Changes in Fiscal Policy: Taxes Versus Spending”

• III. GUAJARDO, LEIGH, AND PESCATORI: “EXPANSIONARY

AUSTERITY: INTERNATIONAL EVIDENCE”

Possible Problems with the Conventional CAPB Indicator of Fiscal Consolidations

• Something like a stock market boom may raise CAPB

and be correlated with other factors raising output.

• Discretionary changes in CAPB may be taken in

response to the state of the economy.

• One-time accounting changes may lead to

attenuation bias.

Narrative Approach

• Based on real-time OECD, IMF, and country budget

reports and documents.

• 17 countries for 1978–2009.
• Try to determine when there were deliberate fiscal

consolidations and whether they were taken in response to the economy.

• Use ex ante estimates of size of actions.

Evaluation of the Narrative Measure

From: Guajardo, Leigh, and Pescatori, “Expansionary Austerity”

Some Examples of Large Differences between the Two Measures

• Germany (1995 and 1996)
• Ireland (2009)
• Denmark (1986)
• Italy (1993)

From: Guajardo, Leigh, and Pescatori, “Expansionary Austerity”

• Is this test sensible?

Single-Equation Specification

• Where Yi,t is an outcome in country i in year t and F is

the change in CAPB ratio.

• k is set equal to 2 (using annual data).
• Run using both OLS and instrumenting with their

narrative measure of fiscal consolidations.

• Would it be more sensible to enter narrative

measure directly?

From: Guajardo, Leigh, and Pescatori, “Expansionary Austerity”

VAR Specification

• 4 variables (in this order):
• Narrative measure of consolidation shocks.
• Change in CAPB ratio
• Change in log consumption
• Change in log GDP
• 2 lags
• Consider an impulse to equation 1 (narrative

measure) such that it results in a contemporaneous rise in CAPB ratio of 1% of GDP.

For Comparison with CAPB Measure

• Same VAR (in same order):
• Change in CAPB ratio is second equation
• Consider an impulse to equation 2 (change in CAPB

ratio) of 1% of GDP.

• Is this sensible? What observations are they using to

estimate the IRF to a change in CAPB ratio?

From: Guajardo, Leigh, and Pescatori, “Expansionary Austerity”

From: Guajardo, Leigh, and Pescatori, “Expansionary Austerity”

Calculating a Fiscal Multiplier from a VAR

• Over some horizon, take the cumulative percentage

change in GDP and the cumulative change in the fiscal variable (as a percent of GDP) in response to impulse (either in itself or in the instrument).

• Calculate the ratio. In this case −1.57/1.68 = 0.93.

Extension: Splitting Narrative Measure into Spending-Based and Tax-Based Consolidations

• How do they do it?
• Is this sensible?
• Run a 5-variable VAR
• GLP don’t tell us timing/ordering assumption; might

it matter?

From: IMF, “Will it Hurt? Macroeconomic Effects of Fiscal Consolidation”

From: Guajardo, Leigh, and Pescatori, “Expansionary Austerity”

Figure 9 VARs for the Two Types of Exogenous Tax Changes and Real GDP

From: Romer and Romer, “ A New Measure of Fiscal Shocks”

• 5.0
• 3.0
• 1.0

1.0 3.0 5.0 7.0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Percent Quarter

• d. Response of GDP to Tax

Using Deficit-Driven Tax Changes

Extension: Role of Monetary Policy in Explaining Different Results with Spending-Based and Tax-Based Consolidations

• Add the change in the policy interest rate to the VAR.
• Order it last.

From: Guajardo, Leigh, and Pescatori, “Expansionary Austerity”

From: IMF, “Will it Hurt? Macroeconomic Effects of Fiscal Consolidation”

Extension: Splitting Narrative Measure into High and Low-Sovereign Default Episodes

• How do they do it?
• Is this sensible?
• Run a 5-variable VAR
• What would be an obvious alternative approach?

From: Guajardo, Leigh, and Pescatori, “Expansionary Austerity”

• IV. ALESINA, FAVERO, AND GIAVAZZI: “THE OUTPUT

EFFECT OF FISCAL CONSOLIDATION PLANS”

A, F, and G’s Focus

• Interested in fiscal plans – multiyear programs of

fiscal consolidation.

• Follow WEO’s narrative identification of fiscal

consolidations.

• Interested in effects of tax increases versus spending

cuts.

Notation

• τ means tax increases, g means government spending

cuts.

• e = τ + g.
• Superscripts: u means unanticipated (announced in t,

implemented in t). a means anticipated (announced at least a year before being implemented).

• Subscripts: i countries, t years, j years in advance.
• Example: ea

i,t,j (j > 0) is consolidation in country i

announced in (by?) year t to be implemented in year t + j.

A Couple of Their Equations

Notation (continued)

• TB means tax-based, EB means expenditure-based:
• Why 0-1 rather than continuous?

Specification

Restrictions That Are Imposed

• Only the ϕ’s vary by country.
• The ϕ’s are the same for tax-based and expenditure-

based consolidations.

• The impact of anticipated changes does not depend on

how long they have been anticipated for.

• The impact of anticipated changes when they are

implemented interacts with the current EB-TB dummies.

Results

Implications for Selected Countries

Some Issues

• Why are the estimates so similar across countries?
• Why are the standard errors so small?

Why Are the Standard Errors So Small?

From: Guajardo et al. (1.65-s.e. confidence intervals); Alesina et al. (1-s.e.)

Why Are the Standard Errors So Small? (continued)

Additional Issues

• Does monetary policy explain the differences in the

effects of tax-based and expenditure-based plans?

• Final thoughts.

• V. BLANCHARD AND LEIGH: “GROWTH FORECAST

ERRORS AND FISCAL MULTIPLIERS”

Blanchard and Leigh’s Basic Equation

Where:

• ΔYi,t:t+1 is real GDP growth in country i from 2010 to

2011.

• ΔYi,t:t+1|t is the forecast of that GDP growth made in

April 2010.

• ΔFi,t:t+1|t is the forecast of the change in the cyclically-

adjusted budget surplus in 2010-11 as a percent of potential GDP made in April 2010.

Blanchard and Leigh’s Basic Results

Robustness

Considering Different Time Periods