FRASER RESEARCH BULLETIN June 2020 Is Fiscal Stimulus an Effective Policy Response to a Recession? Reviewing the Existing Research by Jake Fuss, Alex Whalen, and Tegan Hill Summary This report examines existing academic Research also shows that the Canadian studies that analyze evidence regarding the government’s stimulus in response to the 2009 effectiveness of fiscal stimulus—additional recession contributed little, if anything, to the government spending and/or tax relief—as a economic turnaround. mechanism to mitigate the impact of a reces­ Harvard University professor Alberto Ale­ sion and speed up economic recovery. The re­ sina finds that increases in government spend­ search raises significant doubts about whether ing are associated with lower growth, while fiscal stimulus can achieve this objective. stimulus based on decreases in personal and Evidence from University of California— business taxes are associated with higher rates San Diego professor Valerie Ramey and Har­ of economic growth. In addition, Alesina’s work vard University professor Robert Barro dem­ suggests that reductions to government spend­ onstrates that the fiscal multiplier created by ing are an important part of reducing fiscal increased government spending is below 1.0, deficits following a recession. indicating that stimulus measures actually Before implementing any fiscal stimulus crowd out economic activity that would other­ packages, Canadian policymakers must con­ wise have occurred. sider the potential implications on both the Empirical research from Stanford Univer­ economy and government balance sheets. Past sity professor John Taylor shows that the US history suggests that stimulus will not improve stimulus package during the 2008-09 recession the Canadian economy and may even be a det­ failed to increase consumption and had little to riment to it. no effect on economic growth. fraserinstitute.org FRASER RESEARCH BULLETIN 1 Is Fiscal Stimulus an Effective Policy Response to a Recession? Introduction fessors Robert Barro and Alberto Alesina, and Is fiscal stimulus—additional government Stanford University professor John Taylor. spending and/or tax relief—an effective mecha­ Their findings have important implications for nism by which governments can mitigate the Canada as policymakers in this country con­ impact of recessions and speed up economic template their potential response to the recent recovery? This bulletin provides background recession. on the critical question of stimulus, which is al­ ways of interest but particularly so when econ­ Can government spending help the omies are in recession. economy? During the 2008-09 recession, Canadian gov­ University of California—San Diego professor ernments enacted discretionary fiscal stimu­ Valerie Ramey and Harvard University profes­ lus packages in response to the downturn. The sor Robert Barro have conducted decades of stimulus included a mix of tax and spending empirical research assessing the economic im­ measures that included tax cuts, public infra­ plications of fiscal stimulus, specifically, in­ structure spending, and industry subsidies. creases in government spending. The primary purpose of these measures was to Economists generally use a concept called the jump-start the economy. “fiscal multiplier” to help illustrate their find­ There are already calls in Canada and the US ings. The fiscal multiplier shows the impact that for fiscal stimulus in reaction to the recession each additional dollar of government spending induced by the global outbreak of COVID-19 can have on the economy. The logic is that in and the related government policies that have times of economic downturn or recession there slowed or shut down large sectors of the econ­ are underused resources in the economy and omy. Given the importance of this issue, it is that government spending can work to mobi­ appropriate to review the existing research on lize these resources, create jobs, and increase the impact and effectiveness of fiscal stimulus incomes. The theory assumes that once the ef­ programs at achieving their objective to im­ fects of government spending are felt, people prove the growth of the economy. will spend more of their additional income and create a positive ripple effect throughout the This report examines existing academic stud­ economy. Person A receives new money from ies that analyze evidence of the effectiveness the government and spends a significant por­ of fiscal stimulus. The spectrum of research re­ tion of it on goods and services provided by viewed is not meant to be exhaustive and fo­ persons B and C, creating income for persons B cuses narrowly on empirical assessments of and C, who likewise spend a significant portion the effect of fiscal stimulus based on historical of the income on goods and services, and so data and experience, rather than model-based on, and so on. If the multiplier is greater than work. Although the latter literature has impor­ 1.0, then a $1 increase in government spending tant contributions to make, it is less helpful in will increase overall economic output by a value confirming real-world effects from stimulus. greater than $1. For example, a multiplier of 1.5 Notable authors using empirical assessments means that a $1 increase in government spend­ include University of California San Diego pro­ ing raises overall economic activity (i.e., GDP) fessor Valerie Ramey, Harvard University pro­ by $1.50. fraserinstitute.org FRASER RESEARCH BULLETIN 2 Is Fiscal Stimulus an Effective Policy Response to a Recession? It’s important, however, to understand that in­ that the implied multiplier on total GDP is about creased government spending often is financed 0.3, and that tax changes do not have a signifi­ by reductions in spending in other sectors of cant effect on the spending multiplier. Her em­ the economy such as private consumption and pirical data indicated that government spend­ investment. Simply put, government spending ing would not stimulate the private sector, and is not a free lunch. The resources for govern­ in fact, likely crowds out private activity, such as ment spending must come from somewhere. As consumption, investment, or net exports. such, if the multiplier is less than 1, then gov­ ernment spending crowds out economic activ­ ity that would have otherwise taken place. This ... government spending tends is contrary to the objective of fiscal stimulus. to reduce private spending. A study by Ramey (2011) surveyed more than 30 studies to assess the existing literature on Put differently, $1 of additional deficit-financed temporary spending—typical­ ly used in stimulus—and found that the multi­ government spending led to less plier for the United States is likely between 0.8 than $1 of private activity. to 1.5. The study notes that large increases in government spending are typically followed by increases in economically harmful taxes (Burn­ side et al., 2003). For this reason, Ramey con­ The study also explored the impact of govern­ cluded that any estimate should account for the ment spending on labour markets. It found that effect of taxes on the economy. Ramey found while there is an increase in employment from that the existing empirical research indicates government spending, it is almost entirely in the tax multipliers between -0.5 to -5.0, meaning government employment, as opposed to private that even if the spending multiplier is positive, employment, again indicating limited effects on there will be countering effects for future tax the private sector from government spending. increases which may reduce or cancel out any Ramey and Zubairy (2018) conducted further stimulus effects. Overall, the study concluded research to test whether government spending that there is not clear empirical evidence that multipliers differed based on the level of un­ the government spending multiplier is great­ used resources in the economy (ie., “slack”) or er than one, and thus government spending based on interest rates reaching the zero lower does not necessarily stimulate private activity, bound. “Slack states” are defined by an unem­ so overall economic activity might not be im­ ployment rate above 6.5 (higher thresholds are proved from additional government spending. used in robustness checks). The study also re­ In another study, Ramey (2012) used several dif­ viewed multipliers at the point of military news ferent samples, models, and specifications to shocks—using Business Week and other news­ calculate a spending multiplier. She found that papers to identify when individuals expect a across these estimates government spending change in government spending based on news tends to reduce private spending. Put different­ of military spending. The study also reviewed ly, $1 of additional government spending led to multipliers using the Blanchard-Perotti meth­ less than $1 of private activity. Ramey concluded od, which uses information on tax and transfers fraserinstitute.org FRASER RESEARCH BULLETIN 3 Is Fiscal Stimulus an Effective Policy Response to a Recession? to infer the point of actual government spend­ the direction of causation between spending ing.1 These methods of identification work to and the state of the economy. Put differently, capture both delayed increases in government these studies assume that government defense spending and immediate rises in government spending is less likely to respond to economic spending. events, which makes it easier to determine if such spending has caused a change in econom­ Simply put, the study found that government ic activity. spending multipliers were below 1.0 regardless of the level of slack in the economy. Generally, estimates for the multiplier range between 0.3 and 0.8. These findings are close to the esti­ ... the results indicate that mates from Barro and Redlick (2011). For inter­ government spending had est rates near the zero lower bound, there is also limited evidence that multipliers are great­ a dampening effect on the er than one. The results of this study imply that government spending multipliers are not neces­ economy as opposed to an sarily larger during periods of recession.