Business cycles: In search of a theory of deep downturns
Dec 19th 2011, 20:37 by R.A. | WASHINGTON
LAST week, I commented on Joseph Stiglitz's unorthodox view of the Depression and the more recent Great Recession. The former, he writes, was largely attributable to the knock-on effects of a broad structural transformation in the economy, from agriculture to manufacturing. And the latter, he argues, is similarly due to the shift from manufacturing to service employment. I continue to see this as a mistaken view of economic history, but following on an exchange between Brad Delong and Scott Sumner (here, then here, then here) I'd like to add a few more thoughts.
Mr Stiglitz is after a theory of deep downturns. His logic is tempting in its way; the modern economy has suffered two deep downturns and two broad sectoral shifts, and they conveniently line up with each other. The trouble is that this tempting narrative is factually questionable. The modern American economy has suffered many deep downturns, most of them just happen to have occurred before the second world war. If we look internationally, we many more examples of serious recessions outside of the big two that register prominently in our memory. Some overlap with sectoral changes, but many do not. Indeed, the whole notion that there have been two large, distinct sectoral changes in America's history is odd. "Agriculture" and "manufacturing" aren't really economic concepts; they're categorisations of business that only loosely correspond to the activities in which people are actually engaged.
Sectoral change is a constant presence in the modern economy. Within the major shift from rural to urban employment that characterised the period from about 1850 to 1950, there were several iterations of industrialisation, innovation, and displacement. And it is worth pointing out that the postwar period from 1945 to 1970, which is widely considered a golden age for rich-world growth, was an era of major structural transformation. New transportation and communication technologies erased whole industries and spurred a major shift in America's economic geography. Industry vanished from urban centres and many cities experienced a wrenching depopulation. No deep downturn followed. On the contrary, this is considered one of the most prosperous and stable economic periods in America's history.
Structural changes are economically meaningful. Major shifts can change the return to different kinds of labour, with potentially momentous impacts on the economy. I agree that recent changes have had just this effect, increasing the return to some kinds of skilled labour while eroding the return to moderate skill levels, and this has helped fuel dramatic changes in the income distribution over the past three decades. It's possible to argue that this could impact the macroeconomy. If income systematically shifts toward a class with a greater propensity to save and invest than to spend, then demand might well fall. One response to this might then be to encourage low-income households to take advantage of low borrowing costs to take on debt in order to supplement their incomes. In this way, demand might be maintained, but only at the cost of household debt burdens that grow in an unsustainable fashion. When the unsustainable is no longer sustained a crash occurs and the economy slumps.
That's a decent (though incomplete) description of the mess America got itself into in recent years. Yet it is incorrect, in my view, to attribute the mess to the sectoral shift. It would be more appropriate to blame policymakers. Congress amplified underlying income trends with changes in tax policy, on the one hand. Perhaps more importantly, the Fed and federal regulators opted to hold wage inflation at a low level and encourage borrowing when they might have instead pushed for a larger level of wage inflation and used regulatory tools to limit borrowing. In particular, the decision to tolerate a credit-fueled boom in asset prices as a salve for dormant wage inflation looks very foolish.
It's important to recognise, though, that wages for low income workers have been stagnant for some time. The biggest rise in income inequality within the bulk of the income spectrum (within the 99%, that is, rather than between the 99% and the 1%) occurred in the 1980s. And household indebtedness rose to a relatively high level in the late 1980s, which was also a decade of gutting deindustrialisation. America avoided a deep downturn, however. The early 1990s brought on a serious recession, of course, but unemployment peaked at 7.8%. And if you step back and look at the period from 1929 until now, you see lots of episodes of negative real economic growth but only two episodes of serious contraction in nominal demand: the 1930s and the period from 2008-2009. That strongly suggests that a theory of deep downturns needs to focus on the demand side and in particular on the money side.
If I were going to try to generalise this a bit, I'd suggest that a theory of deep downturns needs to have two key ingredients. First, it needs a theory of big demand shocks. And second, it needs a theory of bungled policy reponses.
A good and reliable candidate for the former might be a major episode of financialisation in the economy that coincides with a surge in capital flows. If there is then a sudden change or reversal in the pattern of capital flows, as often occurs, the financialisation of the economy amplifies the impact and a big demand shock is the result. I wouldn't begin to suggest this is the only thing that ever generates big shocks. It is perhaps the most common and significant of the possible suspects, however.
A big shock will not, on its own generate a deep downturn, especially one of a sustained nature. To do that it needs an accomplice: bungled policy. Governments can and do get many things wrong in the wake of economic shocks, but few mistakes are powerful enough to reliably generate depressions. Monetary policy mistakes can. In the Depression, the specific mistake was the near-religious adherence to an institution—the gold standard—that frequently forced central banks to do the opposite of what the economic situation demanded. We can generalise the policy blunder simply by noting that central banks are very uncomfortable tolerating above-normal inflation and even more uncomfortable with the task of actively generating inflation. Faced with a major demand shock, however, that is precisely what an economy needs.
The downside to the above theory of deep downturns is that it seems to be bordering on the tautological. Big downturns, I'm arguing, are caused by big shocks to demand that lead to big declines in the amount of money changing hands, and policymakers contribute to the downturn by failing to ensure that the amount of money changing hands moves back to the normal, expected level. Look more closely, however, and there are meaningful premises here around which to build defences against deep downturns. Be wary of episodes of financialisation. Lean against large capital flows, perhaps through fiscal and macroprudential policy. React to big shocks with overwhelming monetary force. And don't take your foot off the monetary accelerator until demand (as measured by nominal spending, say) returns to the previous trend.
I can certainly see the ways in which broad structural trends can contribute to deep downturns, in the way that accidents may become more likely as a highway moves from flat plains to winding mountain passes. At the same time, drivers are typically capable of navigating mountain passes safely. It takes some other, distinct error to cause an accident.