Economics 210c/236a Christina Romer Fall 2016 David Romer L ECTURE 6 The Effects of Fiscal Changes: Fiscal Consolidations September 28, 2016
I. O VERVIEW OF THE I MPACT OF F ISCAL C ONSOLIDATIONS
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. G IAVAZZI AND P AGANO : “C AN S EVERE F ISCAL C ONTRACTIONS B E E XPANSIONARY ? T ALES OF T WO S MALL E UROPEAN C OUNTRIES ”
Giavazzi and Pagano’s Regression
Possible Omitted Variable Bias • Cyclical adjustment may not fully deal with cyclicality of 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. G UAJARDO , L EIGH , AND P ESCATORI : “E XPANSIONARY A USTERITY : I NTERNATIONAL E VIDENCE ”
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)
• Is this test sensible? From: Guajardo, Leigh, and Pescatori, “Expansionary Austerity”
Single-Equation Specification • Where Y i,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 d. Response of GDP to Tax 7.0 5.0 3.0 Percent Using Deficit-Driven Tax Changes 1.0 -1.0 -3.0 -5.0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Quarter From: Romer and Romer, “ A New Measure of Fiscal Shocks”
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. A LESINA , F AVERO , AND G IAVAZZI : “T HE O UTPUT E FFECT OF F ISCAL C ONSOLIDATION P LANS ”
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: e a 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. B LANCHARD AND L EIGH : “G ROWTH F ORECAST E RRORS AND F ISCAL M ULTIPLIERS ”
Blanchard and Leigh’s Basic Equation Where: • Δ Y i,t:t+1 is real GDP growth in country i from 2010 to 2011. • Δ Y i,t:t+1|t is the forecast of that GDP growth made in April 2010. • Δ F i,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
Recommend
More recommend
