On the heels of yesterday’s post on risk, outliers, and uncertainty, I saw an interesting article today in the New York Times. It discusses a recent paper highlighting potential errors in the work of Carmen Reinhart and Kenneth Rogoff, authors of the influential 2010 economic study “Growth in a Time of Debt” and the book This Time Is Different.
Covering financial crises in many countries over the past several centuries, Reinhart and Rogoff’s (RR) massive study draws the conclusion—controversial in certain circles—that, when a country’s overall debt exceeds a certain level with respect to the size of the country’s economy, expressed as GDP, future growth declines. It is used as an argument against excessive government spending and debt, particularly here in the U.S.
Let’s set aside the mathematical details for a moment and engage with the basic idea behind the RR analysis: the more you borrow, the less you grow in the future. First, consider where growth comes from—population growth and growth in productivity, which stems from technological advancement and capital investment. We can ignore population growth in this context, so let’s look at what supports technological advances and capital investment. In short, it’s money.
More debt means higher debt payments. Given a stable income stream, the more money that’s spent on debt payment, the less is available for research and development or buying new equipment. This isn’t politics, it’s math. If future growth comes from R&D and capital investment, then more debt service means less cash available for investment and lower future growth.
While the fundamental argument is irrefutable, the give points are whether debt will, in fact, enable future growth and where the debt threshold is. If, for example, the money raised by debt was invested in public goods with a higher rate of return than the debt itself, that might well be true; the U.S. interstate highway system is a great example. The argument for borrowing to invest in infrastructure is a good one and would enable a higher threshold. If, on the other hand, the debt is used for immediate consumption, it will subtract from future growth as described above, enabling a lower threshold.
Nonetheless, unless investment opportunities are unlimited, the ability to borrow to invest profitably will be exhausted at some point, generally speaking—and at that point we will hit the threshold and additional debt will slow growth. We can argue about where the threshold is, but that it exists seems undeniable.
The paper in question, which I have not yet read, deals with some supposed mathematical errors in the RR study. If its assertions are true, it’s embarrassing. The paper doesn’t appear to address the study’s fundamental argument, however. At bottom, it raises the issue of precision versus accuracy, as well as the issue of narrative economics versus mathematical economies, both of which I have written about before.
The fundamental problem with economics is that it’s not a hard science but a people science. Although we may not have the necessary data or rules to be precise, we can be accurate in general terms. We can say west or east, to some extent, but not “32 degrees north of east.” Economics is best evaluated in terms of fundamental reasonableness, as are investment strategies. At Commonwealth, for example, we try to be generally right (as opposed to precisely wrong).
According to the NYT article, the paper attacking the RR methodology notes that the study may have caused great suffering by justifying unnecessary austerity. I don’t agree. What caused the collapse of Greece, for example, or the slowing of the eurozone economies in general, was accumulating and unsupported debt. With or without the RR study, a reckoning was coming.
The U.S. approach, as messy as it has been, has reflected a healthy debate that precludes reliance on the false sense of precision that models generate, and it bodes well for a solution that is based on reality rather than models. We are in the process of marking our decisions to reality, which is good, because reality is much less susceptible to manipulation.