FISCAL STIMULUS IN ECONOMIC UNIONS: WHAT ROLE FOR STATES. Gerald Carlino Federal Reserve Bank of Philadelphia. Robert Inman University of Pennsylvania

Transcription

1 FISCAL STIMULUS IN ECONOMIC UNIONS: WHAT ROLE FOR STATES by Gerald Carlino Federal Reserve Bank of Philadelphia Robert Inman University of Pennsylvania

2 MOTIVATION FISCAL POLICY IN ECONOMIC UNIONS The Great Recession and US Fiscal Policy (ARRA) $797 Billion $381 B in Tax Relief + $98 B in Federal Purchases + $318 B in State Transfers (48%) (12%) (40%) LONG RECOGNIZED THAT STATES MIGHT BE IMPORTANT: BUT HOW?

3 On Their Own? When confronted by severe depression state-local governments as a whole have not had and cannot be expected to develop the fiscal ability to expand countercyclical spending. The few that are strong enough tend to hold back because expansion or contraction by them can affect the level of income and employment within their own area to only a limited degree. As Agents for Coordinated Action by the Federal Government? Financial assistance from the federal government by grants and loans has been suggested as a technique for getting concerted action by all levels of government. James Maxwell, Federal Grants and the Business Cycle, p. 100, NBER Fiscal Studies, No. 4, 1952 This Is The Received Wisdom as Delivered by: Richard Musgrave (Theory of Public Finance) Wallace Oates (Fiscal Federalism) THE AGENDA HERE: IS MUSGRAVE-OATES CONJECTURE CORRECT?

4 OUR CONCLUSIONS 1. From an analysis of state budgets and state economies for the years , we find that state deficits can improve job growth in the deficit state, but that there are also significant spillovers to economically neighboring states. 2. The deficit cost per job created in the deficit state is (on average) about $71,800/job. Including jobs created in economically neighboring states lowers the cost per job to (on average) $50,000 per job. Spillovers are significant, statistically and economically. Coordinated fiscal policies are more cost-effective. 3. Intergovernmental fiscal assistance (A) is the most plausible means for a coordinated policy. Current federal assistance is paid to states either as unconstrained ( fungible ) project aid (AP) or as targeted, matching ( price incentive ) aid (AW): A = AP + AW. 4. From an analysis of the impact of A, AP and AW on national income, we find that aggregate aid does positively impact income growth, and that AW assistance is more effective than AP assistance. For two reasons. AW aid is a price incentive so money is spent and AW aid is targeted for income and services provided to credit constrained lower income households. 5. The ARRA policy mix might have been 30 percent more effective (or 30% less expensive) in stimulating the aggregate economy had only AW assistance been used to stimulate the economy.

5 STEP 1: CAN STATES IMPACT THEIR OWN ECONOMIES? Estimation of: ( N H ) = f(ownd(-1), AP(-1), Spillovers; Controls) + υ st, where: OwnD(-1) = State own cash-flow deficit lagged one year. AP(-1) = Federal unconstrained aid lagged one year. Spillovers = Job (or Population Growth) of Economic Neighbors lagged one year. Controls = Changes in world energy prices; changes in state productivity; lagged weather disasters υ st = v t + v s + v st Estimated for the years,

6 OVERVIEW OF ESTIMATES FOR A TYPICAL STATE: FOR A 1 SD INCREASE IN OwnD(-1) = $390/person will: L Increase job growth by 1.2% one year later or by 34,000 jobs; L Increase in-migration implies 7,000 of those jobs filled by in-migrants; L Increase the aggregate state deficit by $2.44 billion; L Imply the average cost per job created within the state will be $71,800/job. BUT IMPACTS ARE SHORT-LIVED Figure 2

7 Figure 2: Responses of State Job Growth to an Increase in State Own Deficits* % Change in Job Growth Years Source: Carlino and Inman (2013) * The solid line represents the time path for changes in the rate of state job growth in response to a 1 percent change in the state's own deficit introduced at the start of year 0. The dashed lines represent the 95 percent confidence band for each year's projected change in job growth.

8 SPILLOVER EFFECTS ARE IMPORTANT, SO FREE-RIDING A RISK Table 2 REGION LARGEST JOBS STATE (ΔDeficit/Own Rev) LARGE STATE S OWN JOBS (Deficit Cost/Job) JOB SPILLOVERS TO OTHER STATES (Deficit Cost/Job) REGION S TOTAL JOBS (Deficit Cost/Job) OTHER STATES OWN JOBS (Deficit Cost/Job) JOB SPILLOVERS TO LARGE STATE (Deficit Cost/Job) REGION S TOTAL JOBS (Deficit Cost/Job) REGIONAL POLICY JOBS (Deficit Cost/Job) FAR WEST California (.10) 158,483 ($90,956) 108,561 ($0) 267,043 ($53,980) ($84,234) 43,836 ($0) 152,158 ($49,990) 419,201 ($52,532) Source: Carlino and Inman (2013)

9 THEREFORE: COORDINATE FISCAL POLICIES THROUGH STATE GOVERNMENTS USING INTERGOVERNMENTAL AID AP = Project Aid (Education, Highways, GRS, Model Cities, TANF, Tobacco Settlement) AW = Welfare Aid (AFDC, Medicaid) AID = AP + AW HISTORICAL ALLOCATIONS TO AID Figure 3 ARRA ALLOCATIONS: AID = $318 B = $217 B + $101 B = AP + AW

10 Federal Aid, Federal Purchases, and Federal Net Revenue: * (Per Capita, 2005 Dollars) $/person 2,500 Figure 3 Total Aid, Welfare Aid, and Project Aid 2,000 Total Aid (A) Welfare Aid (AW) Project Aid (AP) 1,500 1, * Recession years shown as shaded bands Figure 4 Federal Purchases and Federal Net Revenue $/person 6,000 5,000 Federal Purchases (G) Fed Net Revenues (R) 4,000 3,000 2,000 1, * Recession years shown as shaded bands

11 ESTIMATING THE EFFECTS OF A, AP, and AW on GDP Impact Estimates: Wilson (2012) Feyrer-Sacerdote (2011) Chodorow-Reich, et. al. (2012) Dynamic Estimates: Our Approach using Structural VAR Tables 3 and 4

12 TABLE 3: SVAR Estimates: GDP Responses to Fiscal Shocks for Blanchard-Perotti and Aid Specifications Specification Sample Period Blanchard- Perotti 1947:1 to 2010:3 (1) Blanchard- Perotti 1947:1 to 2010:3 (2) Blanchard- Perotti 1960:1 to 2010:3 (3) Blanchard- Perotti 1960:1 to 2010:3 (4) Aid Specification 1960:1 to 2010:3 (5) Aid Specification 1960:1 to 2010:3 (6) Aid Specification 1960:1 to 2010:3 (7) Fiscal Policy (R - A) G (R - A) G R G A IMPACT * (-1.72, -1.47).944* (.94,.947) * (-1.98, -1.55).959* (.95,.96) * (-2.84, -2.77).564* (.56,.57).528* (.52,.53) 4 Qtrs * ( -2.24, -1.65).772* (.46, 1.08) * ( -2.45, -1.72) 1.000* (.59, 1.40) * ( -3.90, -2.67).447 (-.24, 1.13).713* (.56,.86) 8 Qtrs * (-1.57, -.87).294 (-.07,.66) * (-1.70, -.75).619 (.20, 1.03) * (-2.96, -1.41).404 (-.30, 1.11).499* (.33,.67) 12 Qtrs (-1.06, -.34).023 (-.36,.41) (-1.05, -.11).340 (-.11,.75) * (-2.26, -.74).309 (-.44, 1.06).360* (.18,.54) 20 Qtrs (-.47,.06) (-.38,.24) (-.34,.16).058 (-.38,.50) (-1.61, -.23).165 (-.64,.98).234 (.04,.43) Peak * (Q2) (-2.35, -1.86).963* (Q2) (.77, 1.17) * (Q2) (-2.55, -1.99) 1.078* (Q2) (.67, 1.41) * (Q2) (-4.22, -3.29).564* (Q1) (.56,.57).802* (Q2) (.68,.93)

13 TABLE 4: SVAR Estimates: GDP Response to Fiscal Shocks With Disaggregated Aid Specification Sample Period 1960:1 to 2010:3 (1) 1960:1 to 2010:3 (2) 1960:1 to 2010:3 (3) 1960:1 to 2010:3 (4) Identification Strategy SVAR SVAR SVAR SVAR Fiscal Policy R G AP AW IMPACT * (-3.06, -2.91).807* (.80,.82) -.108* (-.11, -.11) 1.637* (1.61, 1.66) 4 Qtrs * (-3.71, -2.67).884 (.27, 1.51).919 (.02, 1.78) 2.108* (1.80, 2.42) 8 Qtrs * (-2.73, -1.40).677 (.07, 1.28).908 (.05, 1.77) 1.453* (1.07, 1.83) 12 Qtrs * (-1.97, -.067).498 (-.15, 1.15).886 (-.03, 1.80).988* (.58, 1.40) 20 Qtrs (-1.15, -.14).301 (-.40, 1.01).843 (-.16, 1.84).548 (.15,.95) Peak * (Q2) (-3.98, -3.23).884 (Q4) (.27, 1.50) (Q2) (.19, 1.82) 2.315* (Q2) (2.08, 2.55)

14 INSIDE THE MACRO ECONOMIC BLACK BOX: RATIONALIZING MACRO MULTIPLIERS WITH STATE FISCAL BEHAVIOR STEP 1: HOW DOES FEDERAL AID IMPACT STATE BUDGETS? BUDGET IDENTITY: AP + (rs - b) - (gs + k) / SURPLUS = Δc - Δd + Δf ($504) + ($ $276) - ($ $312) / (-$24) = ($81) - ($55) + (-$50), b = BC(1 - m) STATE FISCAL BEHAVIOR (rs, b, gs, k, Δc, Δd, Δf) = f(ap,1 - m; I, é; c -1 ; X) + (v t + v s + v st ), (Results in Table 6)

15 STEP 2: IMPLIED 4-QUARTER MULTIPLIERS FOR FEDERAL AID For a $1 Increase in Federal Project Aid: ΔAP = 1* Δrs/ΔAP.00 ΔB/ΔAP.02 ΔG/ΔAP.50 ΔW/ΔAP.48 Simulated Four Quarter Project Aid (AP) Multiplier dgdp/dap = (dgdp/drs)c[δrs/δap] + (dgdp/db)c[δb/δap] + (dgdp/dg)c[δg/δap] + (dgdp/dw)c[δw/δap],.52 = (-3.19)C[.000] + (1.59)C[.02] + (.88)C[.50] + (.10)C[.48] For a $1 Increase in Federal Welfare Aid: ΔAW = 1 from a small increase in Δm Δrs/ΔAW = (-52.58)/(135.50) = -.39 ΔB/ΔAW = (95)/(135.50) =.70 ΔG/ΔAW = (-53.30)/(135.50) = -.39 ΔW/ΔAW = (41.30)/(135.50) =.30 Simulated Four Quarter Welfare Aid (AW) Multiplier dgdp/daw = (dgdp/drs)c[δrs/δaw] + (dgdp/db)c[δb/δaw] + (dgdp/dg)c[δg/δaw] + [dgdp/dw)c[δw/δaw], 2.04 = (-3.19)C[-.39] + (1.59)C[.70] + (.88)C[-.39] + (.10)C[.30]

16 LOOKING FOR THE COST-EFFECTIVE POLICY MIX ARRA Policy Mix: ΔR = $45.2 B in 2009:2 ΔG = $11.8 B in 2010:1 ΔAP = $8.7B in 2009:2 and $18.75 B in 2010:1 ΔAW = $37.03 B in 2009:2 GDP Growth Over Benchmark With ARRA Policy Mix: 1.95% Targeted Policy Mix: ΔR = $57 B in 2009:2 ΔAW = $64.5 B in 2009:2 GDP Growth Over Benchmark With Targeted Policy Mix: 2.60%

17 A LINGERING (Last) QUESTION: WHY, THEN, USE AP AID? L TO BE QUICK, LEVERAGE EXISTING UNION POLICIES L FOR PASSAGE, GIVE STATE POLITICIANS WHAT THEY WANT Grunwald, The New, New Deal Boone, Dube, Kaplan, The Political Economy of ARRA Brookings Papers

18 LESSON NO. 1: Musgrave and Oates Were Right, BECAUSE OF SPILLOVERS, COORDINATED FISCAL POLICIES IN ECONOMIC UNIONS ARE MOST COST-EFFECTIVE AN EU FISCAL UNION? LESSON NO. 2: In Designing Stimulus Fiscal Policy, IF USING STATES, PRICE INCENTIVES AIMED AT STATE LOWER INCOME ASSISTANCE ARE MOST COST-EFFECTIVE EU-WIDE SOCIAL INSURANCE? LESSON NO. 3: But Remember, in a Union Democracy UNION POLITICS WILL SET UNION POLICIES EU PARLIAMENT or COUNCIL OF MINISTERS?