Why has the response to fiscal stimulus been so hard to detect? Perhaps because, on net, there wasn't much stimulus. According to this research by Joshua Aizenman and Gurnain Kaur Pasricha, the federal stimulus filled holes created in state and local budgets, and that was helpful -- I don't think the article does enough to point out what would have happened if the federal government had not offset these cuts. But how much did the stimulus do over and above simply simply offsetting the negative effects of cuts at the state and local government levels? Apparently, not much:
The net fiscal expenditure stimulus in the US 2008-2009: Less than what you might think, by Joshua Aizenman and Gurnain Kaur Pasricha, Vox EU: Bailout packages have dominated political debate in the US and elsewhere. The global financial crisis led to a massive bailout of the US financial system and significant fiscal stimulus efforts by the US federal government to offset the resulting severe economic downturn. The sheer size of the federal commitments, at a time when the unemployment reached two digit figures, has led observers to question the efficacy of fiscal policy. Moreover, questions were raised with respect to the size of the fiscal multiplier in the US, as well as about possible adverse effects of higher future debt overhang (see de Resende et al. 2010, Barro and Redlick 2009, Spilimbergo et al. 2009 and the references therein).
Given that the counterfactual of the performance of the US economy in the absence of the fiscal stimulus is hard to ascertain, one may thus question its effectiveness, and hence the logic of continuing it. Before taking a position on these vexing issues, it is vital to ascertain the net size of the fiscal expenditure stimulus of the real sector. This issue is of key importance in a federal system like the US, where the fifty states are restrained from borrowing in recessions, and frequently refrain from raising taxes at times of collapsing tax bases. While stabilising the financial system is useful in preventing bank runs, deepening credit constraints facing key sectors like local government expenditures imply that financial bailouts would not prevent, in the short-run, a sizable contraction of aggregate demand.
In order to address these issues, in recent research (Aizenman and Pasricha 2010) we analyse the patterns of fiscal expenditure of the federal government, the state and the local governments, and the consolidated fiscal expenditure. We distinguish between the “pure fiscal expenditure” and the published total expenditure. The “pure fiscal expenditure” or simply, fiscal expenditure of the textbook variety is defined as the sum of government consumption and government gross investment whereas the published total expenditure equals this pure fiscal expenditure plus transfers. Excluding transfer payments (i.e. transfers to financial sector and automatic stabilisers like higher unemployment benefits that were a consequence of the higher unemployment levels) allows us to consider the impact of the discretionary fiscal stimulus.1 That is, “pure fiscal expenditure” is the government expenditure relevant for computing the Keynesian fiscal multiplier. While a large literature defines countercyclical fiscal policy as one with positive correlation between fiscal surplus and output, we focus on actual government spending. We agree with Kaminsky, Reinhart and Vegh (2004) that one needs to focus on government instruments to smooth business cycles, not on outcomes like fiscal deficit, that are endogenous. For example, the government may be raising tax rates in the recession and cutting expenditure, yet running a fiscal deficit because the tax base is smaller.
Figure 1a reports the seasonally adjusted patterns of the pure fiscal expenditure in the US between 1995 and 2009, at three levels: (a) consolidated, (b) federal, and (c) state and local (all data are from Bureau of Economic Analysis (BEA) website). Figure 1b plots the real expenditures relative to their respective values in the third quarter of 2008. The Figures reveal that during the crisis, state and local fiscal expenditures dropped from $1547 billion in real terms in 2008Q3 to $1545.5 billion in 2009Q3 while the federal fiscal expenditures rose from $991.6 billion to $1043.3 billion over the same period. The consolidated fiscal expenditures therefore increased by a mere $48.9 billion in real terms between 2008Q3 and 2009Q3. Moreover, all the three series fell before rising – as the economy was in a tailspin in the first quarter of 2009, both federal and state fiscal expenditures were falling with it.
Figure 1a. “Pure” fiscal expenditures (consumption + gross investment) (Billions of 2005 US dollars, seasonally adjusted)
Figure 1b. Pure fiscal expenditures (2008Q3 = 100)
Figure 2 plots the fiscal expenditures plus transfers for each level of government. During the crisis, from 2008Q3 to 2009Q4, state and local fiscal expenditures plus transfers rose by $20.43 billion. The federal fiscal expenditures plus transfers increased by $415.39 billion, resulting in a net increase of the consolidated expenditure plus transfers.
Figure 2a. Total expenditures (Pure fiscal expenditures + transfers) (Billions of 2005 US dollars, seasonally adjusted)
Figure 2b. Total expenditures (pure fiscal expenditures + transfers) (2008Q3 = 100)
In order to better understand the actual magnitude of the fiscal stimulus, we proceed by focusing on the patterns of the three (consolidated, federal, and state and local) fiscal expenditure time series relative to GDP. Considering policy lags, fiscal expenditure today may reflect decisions undertaken a period ago, based on GDP at that time. Figure 3 reports the three fiscal expenditure time series, normalised by four-quarter lagged GDP (Each chart reports the actual time series, the predicted series of one-step-ahead forecasts using the appropriate ARIMA structure, and the 95% confidence interval around the predicted value). It reveals that the consolidated fiscal expenditure was well within the confidence interval of prediction for the entire time period between 2006Q1 and 2009Q3. This was largely because even as federal expenditures were rising, state and local government expenditures went into freefall, stabilising only after 2009Q1. In fact, for all of 2008, state and local fiscal expenditures were significantly below what could be predicted using historical data. The figures clearly reveal that the fiscal expenditure stimulus did not expand the consolidated fiscal expenditure above the predicted level.
Figure 3. Pure fiscal expenditures/lagged GDP; 1995-2009
In Figure 4, we examine the actual and predicted values of fiscal expenditures plus transfers, along with the confidence intervals around one-step-ahead predictions. While the consolidated fiscal expenditure plus transfers was outside the confidence interval in two of the last five quarters, it exceeded the upper bound by at most 0.6 percentage points, and fell back within the confidence interval in 2009Q3. Thus, even taking into account the transfers to the financial sector and the automatic stabilisers built into the system, the federal stimulus did not expand the consolidated fiscal expenditure significantly above the predicted level. Even this limited impact may now be over.
Figure 4. (Pure fiscal expenditures+ transfers)/lagged GDP; 2006-2009.
Obstacles for net stimulus
The case for net stimulus in the US is debatable. On the one hand, if the US is already in a robust recovery, as is presumed by more optimistic observers, then net fiscal stimulus may be redundant, and could lead to inflationary pressures down the road. On the other hand, double digit unemployment, and uncertainty regarding the strength of the recovery, may suggest the need for a second US federal fiscal stimulus package. Independently of this debate, understanding the reasons for the lack of greater net fiscal expenditure stimulus in the aftermath of the deepest recession of the last fifty years is essential. One explanation is provided by the moral hazard concerns associated with common pool challenges of a fiscal union. Another explanation for the lack of a larger stimulus is that the present trajectory of the US public debt/GDP, in the absence of concrete plans for fiscal consolidation, is a cause for concern.
The limits on state borrowings in a federal union may be rationalised by the concern that the absence of such limits may induce competitive borrowing by states, with the expectation that the federal government will bail them out if necessary.2 This is an important concern in a highly centralised federal union, where most of the tax revenue is “owned” by the federal government (von Hagen and Eichengreen 1996). There is also concern that transferring federal resources to states with deeper tax revenue shortfalls would “reward” states that were less prudent, and penalise states that have developed a more stable tax base and accumulated precautionary reserves.
While valid, these considerations need not prevent a deeper federal stimulus in the unique circumstances of an unanticipated, deep crisis propagated by financial factors beyond the control of each state. However, moral hazard concerns call for designing the federal fiscal stimulus in ways that would not reward states for their past fiscal mistakes. We suggest two schemes that would mitigate moral hazard. One scheme would involve channelling funds to states on a per-capita basis, so that each state gets an allocation proportionate to its population, as long as the state government is committed to spending these funds and not using them to repay past debts. One may expect that the equitable per capita treatment of this scheme would facilitate greater support for the net fiscal expenditure stimulus by the Congress. Under this scheme, the federal government borrows for the states, in a way that equalises the borrowing per-capita, independently of the quality of the domestic public finance mechanism of each state. Another scheme would involve making the allocation to each state proportional to the federal tax revenue from that state, on average during the few years prior to the crisis.
Another concern restraining public support for greater federal stimulus may be the long run implications of a net stimulus, i.e. the resulting increase in future debt overhang. Observers noted that even before the crisis, the public debt trajectory was unsustainable, and fiscal reform was needed (Auerbach and Gale 2009). Recognising the gravity of the recession induced by the financial crisis may call for coupling any federal fiscal stimulus with outlining a credible medium-term plan for fiscal consolidation. In fact, independent of the fiscal stimulus triggered by the great recession, concerns about the future path of the public debt/GDP remain a serious policy challenge for the US.
To summarise, so far, the federal fiscal expenditure stimulus has mostly compensated for the negative state and local stimulus associated with the collapsing tax revenue and the limited borrowing capacity of the states. While this is a significant accomplishment, the net effect is that the consolidated fiscal expenditure stimulus is small relative to the sharp fall in private aggregate demand. Consumption plus gross investment expenditures of all levels of the US government were only $47.6 billion higher in 2009 than in 2008 whereas total expenditures (which include transfers) were $330 billion higher. Thus, the fiscal expenditure stimulus did not manage to provide a viable cushion for the negative stimulus associated with private sector’s declining demand. This observation is pertinent in explaining the anaemic reaction of the overall US economy to the alleged “big federal fiscal stimulus”.
Aizenman, Joshua and Gurnain Kaur Pasricha (2010), “On the ease of overstating the fiscal stimulus in the US, 2008-9”, NBER Working paper 15784.
Auerbach, Alan and William Gale (2009), “An update on the economic and fiscal crises: 2009 and beyond”, Brookings Papers, October.
Barro, Robert and Charles Redlick (2009), “Design and effectiveness of fiscal stimulus programmes”, VoxEU.org, 30 October.
De Resende, Carlos, René Lalonde, and Stephen Snudden (2010), “The power of many: assessing the economic impact of the global fiscal stimulus”, Bank of Canada Discussion Paper No. 2010-1.
Kaminsky, Garciela, Carmen Reinhart and Carlos Vegh (2004), “When it rains, it pours: procyclical capital flows and macroeconomic policies”, NBER Macroeconomics Annual, Mark Gertler and Kenneth Rogoff (eds.), Cambridge, MA: MIT Press.
von Hagen, Jürgen and Barry Eichengreen (1996), “Federalism, fiscal restraints, and European Monetary Union”, The American Economic Review, 86(2):134-138.
Spilimbergo, Antonio, Steve Symansky, and Martin Schindler (2009), “Fiscal multipliers”, IMF Staff Position Note, SPN/09/11.
The Economist (2010), “Greece and the euro”, 18 February.
1 This observation does not negate the possible benefit of stabilising the financial system by means of federal bailouts. Yet, liquidity transfers to financial institutions do not amount to a net fiscal stimulus that increases the fiscal expenditures in ways that compensate for the impact of borrowing constraints on state and local governments.
2 Examples for gaming a Federal Union by means of “competitive borrowing” include Argentina, where there have been frequent federal bailouts of provincial governments, leading ultimately to higher inflation, and to the ultimate financial meltdown of the early 2000s. See also the recent concerns regarding the euro zone’s rescue plan for Greece (The Economist, Feb 18th 2010, Greece and the euro).