Paul Krugman expressed his view against austerity measures at this time. Even though I also think that we should not take austerity measures right now, I have to disagree with the argument he gives in his posting entitled “Self-defeating Austerity”. The point of the article is that austerity measures have a toll on output which will decrease the government revenues (taxes). Therefore a tight fiscal policy, at this time, will be detrimental for the economy as well as the budget deficit.
Is this right? The devil lies in the details and indeed this is where it again lies in the argument Krugman gives. The key assumption to his argument is the spending multiplier, which is assumed to be 1.4. This means that every dollar spent by the government generates 1.4 dollars in the economy. It is not difficult to understand that if such a multiplier is really at works, then cutting spending really hurts the economy, while pilling up debt “pays for itself”.
Spending Multiplier Debate
So the question is whether the spending multiplier is as high as 1.4. The answer is that, at best, the academic community has been inconclusive about the effectiveness of spending on stimulating output. This means that there is a lot of doubt, in the academic community, that spending works, i.e. that spending has a multiplier value greater than one. There are also people who argue that the multiplier is negative, in which case, spending by the government reduces output. For a more complete discussion of this debate, click here.
In addition, a rough estimate I come up with by reversing Krugman’s argument I find that the spending multiplier is 0.6, much less than the 1.4 used in his article, rendering spending an ineffective policy.
Back to Krugman’s Argument
The whole argument is that since 1% of GDP spending could produce 1.4% GDP growth, cutting it, shrinks the economy by 1.4%, while the savings amount to only 0.65% of GDP, since a 0.25 marginal tax effect implies that the government losses 0.35% of GDP in taxes.
Now, using an arguably more accurate multiplier estimate of 0.6, we have that a 1% of GDP cut in spending, shrinks the economy by 0.6% and results to savings of 0.85% of GDP. These numbers describe a much more appealing picture in favor of fiscal tightening.
Evaluating the Benefits of Spending Cuts
In addition, Krugman calculates the current benefits of not getting additional debt. Assuming that an increase in spending is coming from borrowing at a real interest rate of 3%, the government by not borrowing saves the expense of 3%*1% of GDP or 0.03% of GDP. This amount, using the same marginal tax rate as above, implies that the government does not have to generate an additional 0.03%/0.25 of GDP or 0.12% of GDP. Given that the current interest rates are way lower than 3%; this number is actually an overstatement of the current debt benefits from cutting spending.
Even though the current interest rates are not high, so perhaps this exercise may not be relevant now, I would like to calculate the general benefits of spending cuts, i.e. the maximum benefits, by considering a case in which the interest rate constraint is binding, i.e, when debt reaches its ceiling, so that the reader gets a feeling of the benefits of these policies in extreme cases, when they are really needed. In those days interest rates increase rapidly to very high values, like 10%-20%. With a 10% real rate, the same computation done above implies that, the savings from servicing the new debt amount to 0.4% of GDP.
However, measuring the benefits of spending cuts just by looking at how much less debt we need to service, in a strictly output-measure sense, is not complete. There are other benefits which are difficult to quantify,[1] like the increased confidence of the bond community on the sovereign entity’s bonds and the benefits from escaping default. Alleviating fears and escaping from the worst, which are worthy way more than 0.4% of GDP.
Parameters as Functions of the State of the Economy
My next comment has to do with the partial approach that we usually employ in economics.[2] The costs calculated above are marginal costs when all the other parameters are fixed to the assumed values. However, these parameters are not constant in the economy. They too change with the state of the economy. For example, the marginal tax rate changes; it increases (or decreases) if taxes are increased and it declines if output declines. The same is true about the spending multiplier; it can increase if the government becomes more efficient, if the economy improves, or if the quality of projects increases. Also, it decreases as it crowds out private sector’s investments and it perhaps increases as the private sector holds on on investing, anticipating high taxes in the future. The same, at last, is true about the interest rate; it is considered to be procyclical, it increases as the credit quality of the country deteriorates, and it increases as fears or expectations about inflation increase.
Acknowledging this fact but also ignoring to some extent, for now, and taking his argument to the extreme can reveal some of its weaknesses, which I have already mentioned. Assuming that the marginal tax effect and the multiplier are as above, the new revenue is always 0.35% of the previous period’s GDP. What is the interest rate that would make taking additional debt impossible to service? This rate would be such that r%*1% = 0.35%, hence r=35%! This means that this scheme could be played as long as real long interest rates are less than 35%. Of course, we all know that a scheme like this would have stopped way earlier, most probably at rates around 15% (this is itself arbitrary but historically justifiable).
A Spending Multiplier Estimate
Reversing the above exercise, we can estimate maximum multiplier that can sustain the above pilling up debt scheme. As the country’s balance sheet deteriorates, both the marginal tax effect and the multiplier change. Assuming that the tax effect is less volatile than the spending multiplier, we can estimate the maximum value the multiplier can get in order to facilitate servicing debt at 15%. Multiplier = 15%/0.25 = 0.15/0.25 = 0.15*4 = 0.6. So, an estimate for the spending multiplier we get by reversing Krugman’s argument is 0.6. This means that for each dollar spent in the economy by the government, the economy increases its output by 0.6 dollars, revealing that government spending is not a very efficient use of the public money.
A More Robust Answer Why We Do not Need Austerity Measures
Instead of going on and giving examples that are based on questionable estimates, a much more direct and unquestionable approach is to look at what the markets say. Interest rates are bottom low. For more, read on my answer to this question.
[1] Putting a number on these benefits depends on the level of interest rates, the amount of debt, the growth rate of debt, the state of the economy and other factors.
[2] This is a general comment, not necessarily against Krugman’s argument, of course. After all, he warns against generalizations of his argument and taking it to the extreme.
