Austerity Versus Stimulus – the Reinhart/Rogoff Controversy

May 16th, 2013 by under Faculty Commentary. No Comments.

Longbrake3by Bill Longbrake, Executive-in-Residence, Center for Financial Policy

This is an excerpt of the May 2013 Longbrake Letter. To read the letter in its entirety, click here.

Much ink has been spilled in recent weeks about computational errors, which have been acknowledged, and methodological errors, which have been disputed, embodied in a paper published, “Growth in a Time of Debt” by Carmen Reinhart and Kenneth Rogoff.¹ The latest round of controversy was ignited by a paper published by Thomas Herndon, Michael Ash and Robert Pollin.² The controversy has been heated because it is more than an esoteric intellectual debate among economists.

Reinhart and Rogoff’s research finds that economic growth declines as the government-debt-to-GDP ratio increases. Specifically, when the ratio eclipses 90%, economic growth drops sharply.  The computational error had to do with the importance of the 90% cliff. Reinhart and Rogoff have acknowledged this error and agree that there is not a cliff effect, but argue vehemently that economic growth declines, nonetheless, as the government-debt-to-GDP ratio rises.

Reinhart and Rogoff’s research, and especially the so-called 90% cliff effect, has been cited by some as justification for curtailing government deficits. The name for the fiscal policy that focuses on driving down government deficits is referred to as “austerity.” Austerity can involve spending cuts or tax increases or both. In the U.S. austerity policy has been championed by Republicans and is expressly embedded in Paul Ryan’s House of Representatives version of the 10-year federal budget. Ryan’s austerity plan encompasses only spending cuts.

European Union (EU) fiscal policy is decidedly focused on austerity but encompasses spending cuts, tax increases and other kinds of economic reforms. The EU has set a 3% annual budget deficit target and member countries are expected to adopt policies to attain that level within a relatively short period of time, generally two years.

Critics of austerity argue that withdrawing fiscal stimulus through spending cuts and tax increases will depress GDP growth and could decrease growth to such an extent that the debt-to-GDP ratio would rise, rather than fall, because the denominator falls faster than the numerator.

To put this debate into perspective requires a discussion of macroeconomic theory.

1.    Keynesian Macroeconomic Policy

There has been an ongoing policy debate among economists and policy makers about the conduct of fiscal policy in times of economic duress. The Keynesian response is that when private aggregate demand declines government must intervene and replace the lost spending power. Then, as private sector demand improves, government stimulus can gradually be withdrawn. Keynesians view fiscal policy as a macroeconomic stabilization instrument. Government stimulus should be added when the economy is at less than full employment but be withdrawn when the economy is at full employment. Government fiscal policy should be neutral over the entire cycle.

Keynesians further argue that without government stimulus there is no assurance that an economy operating at less than full employment will be able to return to full employment on its own. The economy can get caught in a liquidity trap or worse in which a negative reinforcing downward spiral takes hold. The famous economist, Irving Fisher, described this phenomenon in a seminal paper published during the Great Depression which described the debt-deflation process.

Politicians, being the people who they are, especially in democracies, have difficulty withdrawing government stimulus when the economy is strong with the result that a permanent structural budget deficit tends to become embedded. Then, when the economy falters and the budget deficit explodes, the government-debt-to-GDP ratio spirals out of control. And, as Reinhart and Rogoff argue, economic growth may be permanently depressed.

2.    Austerity Macroeconomic Policy

Those who espouse a policy of austerity generally are skeptical of government interference in the marketplace. They argue that government intervention is inefficient and messes things up. Worse, those who promote government intervention seek to reallocate resources in ways that diminish aggregate social welfare. In other words, proponents of intervention sponsor income transfer programs that may increase opportunity for those who are less well situated but that reduce opportunity in the aggregate for everyone – that is, total potential GDP contracts.

Thus, those who favor austerity tend to be advocates of the view that government is better when it is small than large and, thus, a policy of “starve the beast” should be pursued. This interconnects with a bias to minimize deficits even in a time of economic duress. In this context evidence that growth is slower as deficits pile up is taken as direct proof that deficit reduction needs to be pursued vigorously.

3.    Good Deficits and Bad Deficits – Fiscal Multipliers

There is plenty of evidence that government replacement of lost spending power in an economic downturn is essential to interdict the vicious circle inherent in a liquidity trap and to prime a virtuous circle. In other words, government deficits are an essential ingredient in arresting an economic decline and initiating recovery.

But not all forms of government fiscal policy intervention have the same impact dollar for dollar. The important aspect of stimulus is that it leads to spending that creates jobs. If the stimulus goes into the bank without being spent its impact on economic activity will be limited. The relationship between a type of government stimulus and its longer-term impact on economic activity is measured by the fiscal multiplier. If a dollar of stimulus stimulates a dollar of economic activity, the multiplier is equal to one. Multipliers greater than one are highly desirable while multipliers less than one are not.

When the EU reinforced discipline to reduce budget deficits a couple of years ago it believed fiscal multipliers were less than one. Were this actually true, then debt-to-GDP ratios would have contracted even as GDP declined. In addition, per Reinhart and Rogoff, future GDP growth potential would rise as deficits fell.

Unfortunately, we know from the experience of the last two years that the belief that multipliers were less than one was false. European countries forced to adopt stringent austerity measures have seen their economies collapse with limited if any improvement in their debt-to-GDP ratios. The International Monetary Fund (IMF) owned up to this outcome in research published last year which concluded that fiscal multipliers in times of enormous economic slack are much greater than one.

What all of this means is that when there is a large output gap fiscal policy should focus first on boosting aggregate demand but the mix of fiscal policy programs should emphasize high-multiplier initiatives. Then, only as economic recovery gains traction should deficit reduction rise to the fore.

4.    U.S. and European Experiences

Europe has pursued austerity aggressively and has limited its use of monetary policy. Many European countries are still mired in recession and prospects for imminent turnaround are doubtful. Little progress has been achieved in addressing government deficits and high levels of debt. In the meantime social unrest is building and political stability is slowly unraveling.

In contrast, the U.S. initially pursued a traditional Keynesian stimulative fiscal policy. However, many argue that the policy was deficient both in scope and also in composition as the program did not allocate significant resources to high multiplier programs such as investment in infrastructure, education and research.

The U.S. emerged from recession nearly four years ago, but growth has been disappointingly slow since then. Could the recovery have progressed more rapidly in the U.S.? Keynesians argue in the affirmative, but that would have required greater amounts of stimulus.

Now U.S. policy has switched from stimulus toward austerity while the output gap remains extremely large. In the short run this will slow growth. What is important is whether the slowdown is temporary or protracted. My sense is that the negative consequences will linger and the expected strong bounce back in GDP growth that most expect may not materialize. I continue to be concerned about a sustained decrease in productivity growth because of insufficient investment. When deficit reduction is the policy of choice government investment suffers along with other forms of government spending. Reduced investment spending will depress productivity growth and that will decrease potential GDP growth.

5.    Does a High Debt-to-GDP Ratio Cause Slow Growth or Does Slow Growth Cause a High Debt-to-GDP Ratio?

Although Reinhart and Rogoff’s 90% cliff finding that economic growth drops dramatically after the government-debt-to-GDP ratio reachs that level has been discredited, GDP growth is still negatively correlated with that ratio. Critics have correctly pointed out that the statistical analysis merely reveals that a correlation exists but does not prove that high debt ratios cause slower growth. The reverse could be true – lower growth leads to higher debt ratios. This debate will continue and I cannot shed any light on the issue of causality.

However, from my own statistical analysis I can corroborate Reinhart and Rogoff’s finding that there is a strong negative correlation. Reinhart and Rogoff analyzed three data sets. The data set that Reinhart and Rogoff (RR) and their critics, Herndon, Ash and Pollin (HAP) focused on covers 20 advanced economies over the period 1945 to 2009. Data were grouped by ranges of debt-to-GDP ratio and the arithmetic mean and median were provided for each range. Table 4 shows the results of both RR’s and HAP’s calculation based on this data set.

Because RR and HAP used the same data set the differences in medians has to do with HAP’s and RR’s disagreement about calculation methodology. To my way of thinking the differences are not particularly material. The negative correlation is apparent and that correlation is monotonic, which means that growth steadily declines as the debt ratio rises. Clearly, there is no 90% cliff effect and RR acknowledged that fact when they corrected their data.

Table 4

GDP Growth for Developed Countries Covering the Period from 1949 to 2009 Classified by the Size of the Government-Debt-to-GDP Ratio

 May 2013 Longbrake Letter Table 4

*Originally reported by RR as -0.1%; this was the errant calculation that led to the assertion that there is a 90% cliff effect

#Bill’s analysis is a statistical correlation for the U.S. only covering the time period from 1971 to 2013; debt/GDP ratio is public debt only rather than total debt

Also shown in the last column of Table 4 are the results of statistical analysis I conducted on U.S. GDP and the U.S. public-debt-to-GDP ratio for the period 1971 to 2013. For a lot of reasons my analysis is not strictly comparable to that of RR and HAP. However, the direction of correlation and the general order of magnitude among all of these results are similar.

6.    Conclusion

Whether high debt-to-GDP ratios lead to slower GDP growth will continue to be debated. However, what is clear from the analyses is that neither low growth nor high debt ratios is a desirable place to be. However, driving down debt ratios without understanding the transitional impacts on GDP growth can have disastrous consequences. Austerity entails high risks. Outcomes for countries that have pursued austerity are discouraging. Keynesian stimulus appears to achieve better outcomes but a Keynesian fiscal policy, too, can be designed poorly and produce troublesome results.

Read more…

1 Carmen Reinhart and Kenneth Rogoff. “Growth in a Time of Debt,” American Economic Review” Papers and Proceedings, May 2010.

2 Thomas Herndon, Michael Ash and Robert Pollin. “Does High Public Debt Consistently Stifle Economic Growth?” PERI Working Paper 322, April 2013.


Tags:, , , , , .


Leave a Comment