Alberto Alesina on Fiscal Policy and Austerity
What this covers
Russ Roberts interviews Harvard economist Alberto Alesina on the relative costs of different routes to fiscal consolidation. Alesina presents empirical evidence from OECD countries showing that jargon:austerity achieved through spending cuts carries far lower economic cost than austerity through tax increases — in many cases near zero, sometimes even boosting growth. The central claim rests on a methodological foundation: the Narrative Approach, which identifies actual policy decisions made for reasons independent of the business cycle, thereby avoiding Reverse Causality and Endogeneity problems that plague simpler ratio-based studies. Roberts and Alesina discuss the mechanisms behind this asymmetry, concrete episodes including Greece, Italy, and Britain, and the structural factors that make some countries more resilient to high debt than others.
The conversation moves through several distinct layers. Alesina explains why spending cuts outperform tax increases: the former signal lower future taxes to investors and consumers, boosting Forward-Looking Expectations and investment, while tax increases carry distortionary supply-side costs and depress confidence broadly. He addresses the role of monetary policy, structural factors, and the specific problem that governments facing sudden crises tend to raise taxes quickly rather than implement carefully designed spending cuts, leading them to choose the costlier path. The discussion touches on expansionary austerity — growth-compatible spending cuts observed in Denmark, Ireland, Canada, and others — and examines how welfare-state efficiency, not spending level alone, determines distributional outcomes. Alesina also addresses US fiscal dynamics, distinguishing the manageable current debt-to-GDP ratio from the unsustainable trajectory of Medicare and Social Security, and reflects on Paul Krugman's public criticism of his work, noting that later IMF research validated the core finding.
Alesina argues that austerity achieved through spending cuts is far less economically costly than austerity achieved through tax increases, often having near-zero output cost and sometimes being expansionary, because spending cuts shift expectations about future taxes and boost investor confidence in ways tax hikes do not.
- Empirical narrative-approach evidence across OECD countries shows tax-based austerity caused multi-year recessions while spending-based austerity had near-zero average output cost
- Spending cuts signal lower future taxes, raising investment and confidence, while tax increases distort the supply side and depress both consumer and business confidence
- Even after controlling for monetary policy and structural reforms, the asymmetry between spending cuts and tax increases remains large
The fundamental fiscal danger is unsustainable entitlement program growth, not current debt levels, since high debt-to-GDP ratios are sustainable when credibility exists.
- The danger for US fiscal sustainability is not the current debt level but its dynamics: Medicare and Social Security, especially Medicare, are exploding per CBO projections, so without policy change either debt-over-GDP becomes unsustainable or Medicare crowds out all other spending — though no bankruptcy is imminent given US credibility.
“if you read the projection of the Congressional Budget Office, it's really, really scary. So, if policies do not change, either the debt-over-GDP ratio may grow to a level that becomes unsustainable, or we don't spend on anything else other than Medicare, because Medicare is growing at an incredibly fast rate.”
- There is no single debt-over-GDP threshold that signals crisis; sustainability depends on the credibility and investor confidence a country has accumulated, which is why Latin American countries hit crises at much lower ratios than the US or Europe, and why the US can sustain very large debt without crisis.
“the level of the debt-over-GDP ratio, whether it is sustainable or not depends a lot on the credibility and the confidence that investors have on a particular country. And the United States has a lot of confidence on the part of investors, so the United States can sustain a very large debt without a problem of a crisis.”
- Today's very low interest rates make borrowing cheap and mask the burden of debt, but since the relevant horizon is medium-to-long run, rates should be expected to return to normal, at which point the burden of accumulated debt becomes much higher.
“the very low interest rate that we have today in some sense is another [?] confounding factor, because now debt is cheap... in the medium and long run we should expect interest rates to go back to a normal situation, and when interest rates go up then the burden of the debt becomes higher, too.”
- In the medium and long run, fiscal adjustment based on spending cuts has been more successful than tax-based adjustment at keeping the debt-over-GDP ratio constant, because entitlement programs like Social Security and pensions grow automatically, forcing ever-rising taxes unless their growth is halted.
“fiscal adjustment based on spending cuts has been more successful at keeping debt-over-GDP ratio constant rather than spending adjustment based on tax increases. And the reason is pretty obvious: that... there are various entitlement programs--Social Security, pensions--that are increasing at the rate that if you don't stop them, you are forced to raise taxes forever and ever”
- Paul Krugman holds rather extreme views and presents them as if obviously true and anyone who disagrees as obviously wrong, exaggerating to push a position that, while respectable, is not actually proven by the facts or supported by research.
“I think Paul Krugman has rather extreme views. But more importantly, he talks about his views as if they were obviously true, and anybody who would disagree with him was obviously wrong. And he exaggerates.”
Spending cuts produce minimal output costs while tax increases cause severe recessions, a robust empirical asymmetry across austerity episodes.
- When the Cameron government implemented sizeable spending cuts in 2011, the IMF predicted a major recession for Britain, but the UK economy did reasonably well in 2012-2014 without entering recession, prompting the IMF to publicly apologize for having under-predicted UK growth.
“the IMF (International Monetary Fund) predicted a major recession for Britain... in the end we discovered them not to be right; and in fact the IMF apologized publicly for having under-predicted the United Kingdom economy, who did quite well in 2012, 2013, 2014”
- Across episodes of austerity, spending-based fiscal consolidations have close to zero average output cost and have sometimes been expansionary, whereas tax-based consolidations have produced recessions of 2-3% of GDP lasting 2-3 years — a large and robust asymmetry that any unbiased observer would acknowledge.
“We are talking about the effect of spending-based austerity having close to zero cost on average in terms of output; and tax-based austerity having recession of 2, 3% of GDP for 2 or 3 years. So very large, very large differences.”
- Although the IMF in 2010 wrote a pointed criticism of Alesina's work, after all the research was done the IMF's own published paper acknowledged that spending cuts are much less costly than tax increases — the fundamental result on which Alesina's work is based — making Krugman's claim that the work was discredited wrong.
“even the IMF acknowledged--had to conclude that this was indeed the case: that spending cuts were much less costly than tax increases... if you read the published paper of this IMF research, you will conclude that they also agree with this conclusion”
- Expansionary austerity — where spending cuts are accompanied by increasing rather than declining growth because the small Keynesian contraction is more than compensated by other effects — has occurred in real episodes including Denmark and Ireland in the 1980s, Canada and Belgium and Finland in the 1990s, and the UK during the financial crisis.
“in the 1980s there has been an example in Denmark and Ireland; in the 1990s there has been an example in Canada. And in Belgium the case of--and in Finland if I remember correctly; in the case of England even in the middle[?] of the financial crisis is one example.”
Forward-looking expectations and confidence channels explain why spending cuts can be expansionary while tax rises depress both investment and consumption.
- The asymmetry arises because the simple Keynesian model is static: if people perceive spending cuts as permanent, they expect lower future taxes, so consumers consume more today and — more powerfully — entrepreneurs invest more for the future, because investment takes time to materialize and forward expectations matter greatly.
“if you cut government spending today, and people think that this cut is permanent, they will think, 'Okay, in the future taxes will be lower.'... if you had an investor or entrepreneur and you see government spending going down and you expect your taxes to go down in the future, and then you invest in and plan for the future, considering that investment takes a while to materialize, the future matters a lot.”
- Empirical evidence shows that during spending cuts investor confidence rises while consumer confidence falls only slightly, whereas during tax increases both consumer and business confidence drop substantially — so confidence is one of several real channels driving the asymmetry, not a mythical 'confidence fairy.'
“when there are spending cuts, the confidence of investors actually goes up, and the confidence of consumers goes down very little; while with tax increases, confidence of both consumers and business investors goes down quite a bit”
- Although monetary policy may have been slightly more relaxed during spending cuts than during tax increases, holding it constant shows it is not enough to explain the output difference; moreover monetary policy is itself endogenous, since a central bank can relax rates when it sees credible permanent spending cuts but tightens when it fears a tax-driven crisis.
“if monetary policy has been indeed slightly more relaxed during spending cuts than tax increases; but not enough, it is not enough to explain the difference. And, in addition... monetary policy is itself endogenous”
- In the basic Keynesian model, spending cuts should actually be more harmful than tax increases, because cutting spending removes the entire amount plus its multiplier effect, whereas raising taxes reduces aggregate demand less since people absorb some of it by cutting savings.
“in the basic Keynesian model should actually be even more harmful than tax increases, because spending cuts, entire spending is cut; plus the multiplier effect. While if you raise taxes, people reduce some of their savings, so aggregate demand goes down less with tax increases.”
Fiscal crises force governments to choose tax increases for immediate budget effects, even though spending cuts would be economically preferable.
- When a country faces a sudden fiscal crisis with no choice but to act quickly, it tends to raise taxes rather than cut spending, because tax increases produce an immediate budget effect while identifying and implementing spending cuts that minimize social cost takes time the government does not have — leading to the wrong type of austerity.
“when you have no choice and you have to do something quickly, in most cases it is actually much easier to raise taxes, because you have an immediate effect on the budget. Just ask people: Go to the cash machine of the taxpayer, and you get money out of them. To cut spending: Figure out what to cut, how to cut, how to do it.”
- Italy's 2011-2012 technical government, facing near-default with interest rates spiking from Greek contagion, was forced to raise taxes by almost 2% of GDP, which triggered another major recession that left Italy down roughly 9% of GDP per capita while other countries recovered in 2013-2014.
“in the end of 2011, and 2012, they raised taxes by almost 2% of GDP, in a variety of different taxes. And Italy entered another recession; and Italy has been in a recession for, basically, almost until now, and from the beginning of the financial crisis up to now it has lost 9% of GDP per capita, roughly”
- The mere announcement of fiscal policies may itself have an effect on the economy, independent of how much is actually implemented — so a government that announces large spending cuts but delivers less may still influence economic outcomes through the announcement.
“The announcement of policies may actually have an effect on the economy as well.”
- Reinhart and Rogoff were badly and undeservedly mistreated by the press over an error in one particular paper's calculations, but the main message of their work on debt-to-GDP and crises remained solid.
“I think there was an error in their calculations in one particular paper, but the main message of their fantastic book and fantastic work remained solid.”
- The cleanest solution for Greece would have been to default, but this was avoided either because German and French banks held much Greek debt (cynical view) or because a Greek default would trigger contagion in high-debt countries like Italy, Spain, Portugal and Ireland, potentially causing a second financial crisis or even Euro collapse.
“the cynical view is that the German and the French did not want their banks to suffer from a Greek default... The less cynical view would be that the default in Greece would have triggered contagion in other high-debt countries, like Italy”
- Greece's strong growth in the first decade of the Euro was illusory, driven by borrowing and consumption, with a deficit hidden at a supposed 3% of GDP that was actually 13%, alongside major tax evasion and declining productivity, leaving the country in complete economic disarray.
“they were doing it just because they were borrowing and consuming what they were borrowing. They were even hiding their deficit--at the beginning of the financial crisis was supposed to be 3% of GDP... instead it was 13% of GDP.”
Tax increases impose distortionary supply-side effects that reduce work and investment incentives, particularly in unionized economies.
- Tax increases have distortionary supply-side effects left out of the basic Keynesian model: when taxes rise, people may choose to work less and invest less, and in highly unionized European economies cutting public-sector wage growth can lower private-sector wages, raising firm profitability and investment.
“Taxes, on the other hand, have a distortionary effect on the economy, so there is a supply side of the economy that is left out of this model; but if you raise taxes, people may want to work less, may want to invest less”
- The narrative approach — pioneered by Romer and Romer for US taxes and extended by the IMF and Alesina to OECD spending and taxes — addresses identification by documenting the actual policy actions governments decided to take for reasons unrelated to the business cycle, rather than reading endogenous spending-over-GDP and tax-over-GDP ratios.
“you don't look at how government spending over GDP has gone, but you look at what policies have actually been implemented. 'We have cut $30 billion dollars in government salaries,' for example. So you look at the actual policy action.”
- A central measurement problem in austerity research is reverse causality: a falling deficit-over-GDP ratio may reflect rising GDP (the denominator) rather than fiscal tightening (the numerator), so one cannot infer that deficit reduction caused growth without identifying actual policy decisions.
“if you see the deficit-over-GDP going down, and GDP going up, you could not conclude that... the deficit is going down therefore GDP is going up. The deficit-over-GDP may be going down simply because GDP is going up for whatever reason... So it's the denominator, not the numerator”
Spending level alone reveals little about redistributive effectiveness; efficient welfare systems matter more than fiscal size for poverty reduction.
- The level of government spending and taxation reveals little about redistributive outcomes: Southern European countries (Greece, Italy, Portugal, Spain) spend nearly as much as Northern Europe yet have inefficient welfare systems that redistribute poorly, while more fiscally responsible Northern European countries redistribute much more efficiently — so modest spending cuts of 3-5% of GDP need not throw people into poverty.
“countries like in southern Europe--Greece, Italy, Portugal, Spain... are also the countries that have the worst welfare systems in Europe... despite spending so much they are not doing a lot of good in terms of reducing inequality. And the northern European countries, who are actually more fiscally responsible, they have a large government but they are much more efficient at redistributing wealth.”
- Austerity is fundamentally an extraordinary correction that only becomes necessary when a government has previously violated optimal fiscal policy rules — running deficits during recessions or emergencies but failing to run surpluses during booms — so under proper fiscal management austerity would never be required.
“In principle, there should never be any need for austerity if government followed the rules of optimal fiscal policy--namely, having deficit during recession or in period of exceptionally high government spending needed for a war or for a natural calamity or whatever; and then having surpluses during booms”
- From a policy-advice standpoint, rather than merely controlling for structural reforms one can embrace them: the evidence suggests the least costly austerity plan combines spending cuts with structural reforms such as labor and goods market liberalization, which is the recommended approach especially for European governments.
“By looking at the data, we should cut spending and engage in this structural reform, because it would appear from the evidence that the combination of spending cuts and structural reform is the one that works best.”