Free Lunch: Should the government be bigger?

Δημοσιεύθηκε: 9 Απριλίου 2015 - 11:29

I'm from the government and I'm here to help (really)

It had to return: the debate on the appropriate size of the state. Not courtesy of a revived revolutionary left, but from within mainstream macroeconomics (admittedly one of its more thoughtful corners). Bradford DeLong has, in a wonderfully open-minded spirit, posted a draft for comment of his talk on fiscal policy at next week's IMF conference on rethinking macroeconomics. Free Lunch is happy to oblige.

DeLong makes three claims. First, that the government's share of the economy should be bigger in the 21st century than in the last one. Second, that government debt as a share of GDP should also be higher. Third, that the risks of a sudden flight from government debt are not big enough to undermine the previous two points. We'll focus here on the first two claims.

Should the state be bigger? DeLong makes a very good argument: the government should be involved in the spheres of the economy where private markets typically perform poorly. The most obvious of these are health, education, information-intensive goods and services, and public goods in the technical sense - those it costs very little to multiply and/or where it's hard to stop people from using them without paying. The latter will include a lot of what can be called the internet infrastructure and digital products more generally. Now with advanced economies ageing and becoming more knowledge-based and digitised, these types of activity will take up a larger share of the overall economy and so, therefore, should the state. This is particularly interesting from a UK perspective where more than a third of voters want higher taxes and spending, yet no big parties have this on offer.

In a comment that largely endorses the argument, Paul Krugman rewords this as saying "we're going to need even more of this kind of paternalism". That's a sharp, if slightly provocative, way of putting it. But it is also a revealing one. For it is not clear that "more paternalism" is synonymous with a bigger government share in the economy. In an international perspective, many governments - try France and Denmark - have such a large share in their national economies that it's hard to believe they should be bigger still. DeLong and Krugman obviously have the US at the front of their minds.

But even for the US, the question of how big the state actually is is a thorny one. Eduardo Porter recently had an enlightening column in the NY Times on the massive use of tax breaks in the US tax system. This makes the state look a lot smaller than it actually is. The effect is quantified in research by the Peterson Institute (cited by Porter) showing that properly accounting for "hidden" spending, the US is much less different from western Europe and even Scandinavia than we tend to think.

So while DeLong's premise is spot on, the correct conclusion may well be that the composition of government state spending (properly accounted for) should be massively reprioritised. That is to say, more spending on the goods and services listed above, and less on other things (mortgage subsidies for the upper middle class, for one) - not necessarily a bigger aggregate amount.

The second argument is more technical - and perhaps its most important function is to parry an argument against the first, which would state (wrongly) that a bigger state must lead to more government debt. DeLong points out the astonishing fact that since 1890, the interest burden on US government debt (the ratio of interest costs to total debt outstanding) has always been lower than the economy's growth rate when measured at 10-year moving averages. In the jargon popularised by Thomas Piketty (though he thinks the opposite relationship holds now), DeLong finds "r" has been less than "g" at all times for the past 125 years. He claims something similar is true for other north Atlantic economies.

This matters because it means the debt was always more than sustainable: the debt burden (as a share of GDP) would fall by itself even without any principal ever being paid back. DeLong argues that the substitutability of government bonds for private investments means something similar must have been true in the private sector. But that means the economy is "dynamically inefficient" - with returns to capital so low, current generations could increase their consumption more today than future ones would have to cut theirs in the future.

That is a clever way of undermining arguments that debt is too high. But it's not entirely convincing. In a second comment, Krugman points out that the r<g finding doesn't hold across all advanced countries. He still finds the argument powerful for the Anglo-American economies and for Germany. But he does so for what seems like an erroneous reason - that it "lets the government do useful things, like invest in infrastructure". But the dynamic inefficiency story DeLong is telling does not imply that, because it's a story about capital accumulation having gone too far. Using the r<g finding to say we should borrow more is supposed to tell us that we have invested too much and should instead shift consumption from the future to the present.

This perspective also shows a problem with the argument. Is it really believable that the US and western European countries have invested too much? True, some academic research suggests as much by showing that capital returns are consistently lower than capital investment. But this jars with the widely shared impression that the opposite is the case. Perhaps again it's a problem of composition not of totals - we haven't invested too much overall but too much in the wrong things.