The recent slowdown in the US economy raises the question of whether the growth of output may have dropped below the so-called “stall speed”. This is the growth rate at which a healthy expansion can no longer be maintained, after which the economic engine misfires and the US heads back towards recession. For as long as this risk remains, markets are likely to remain nervous, and will certainly be extremely focused on the minutiae of weekly and monthly activity data. Up to now, the data suggest that the engine of the economy has not stalled, though it is making some ominous spluttering noises.
What is the “stall speed” of an economy? Until recently, this concept had not been treated very seriously by economists. However, when the global economy appeared to be grinding to a halt in the summer of 2010, central banks started paying serious attention to the concept, and the Fed’s Jeremy Nalewaik has recently published a detailed paper on the topic. This concludes that the concept of a stall speed has some empirical backing in the US, and that it can help to predict the onset of recessions.
The stall speed is the critical growth rate below which an economic upswing turns first into a period of much slower growth, and then shortly afterwards into a recession. It has important predictive content if a dip below the critical growth rate signals, more often than not, that a cumulative process of economic retardation is setting in. If, on the other hand, the economy frequently drops below the critical rate, but then re-accelerates after a brief pause, the concept is either completely useless or downright misleading.
Why should such a “critical” growth rate exist? One possibility is the behaviour of the labour market. In the US, unemployment tends to rise when GDP growth falls below about 2.5-3 per cent (under Okun’s Law, explained in this blog). If the growth rate falls a lot further than this – say to below 2 per cent for a couple of quarters – then unemployment rises sufficiently for this to impact consumer confidence, and households respond by reducing expenditure. This can lead to a further rise in unemployment, so the downward process becomes self-feeding.
One stylised “fact” about the US economy is that a rise of about 0.5 percentage points in the unemployment rate represents a sort of tipping point. This is often sufficient to trigger a downward spiral, so even such a small rise in the jobless total may not reverse itself, but instead eventually causes a much bigger rise, of at least 2 percentage points in the unemployment rate.
The Fed study suggests that the stall speed for US output growth might be around 2 per cent, annualised on average over two quarters. (Incidentally, the variable used to measure output growth is real gross domestic income, rather than the more commonly used expenditure based estimate of GDP. See this paper by Nalewaik.) Here is the relevant graph:
The graph shows the growth rate of real GDI, along with the critical 2 per cent stall speed, on a two-quarter annualised basis since 1947. It is clear that whenever the economy dips below that growth rate, a period of recession (marked in blue) becomes fairly likely. According to the Fed research, nine out of the 11 post-war recessions have been signalled in this way. (The exceptions were 1948/49 and 1953/54.)
The rule is not perfect – few rules are. It is far too mechanistic to be taken literally. But it may capture a basic truth which should not be ignored. In the last couple of quarters, the economy has dipped to within an inch of breaching its stall speed, but reassuringly it has not actually done so. Up to now, the slowdown seems somewhat similar to the mid course corrections which occurred in the early stages of the last two recoveries, which were subsequently shrugged off.
The Fed research argues that it is possible to improve the accuracy of the rule by adding other variables into the mix, including the shape of the yield curve, housing starts and the change in unemployment. My colleague at Fulcrum, Vasileios Gkionakis, has done a similar exercise to produce probability estimates of of the economy falling into recession, either contemporaneously, or within six months into the future. Here are his results:
The blue line says that the chances of a recession within 6 months are basically zero. However, this estimate is dominated by the level of short term interest rates and the shape of the yield curve, both of which are obviously heavily distorted by the fact that short rates are currently constrained by the zero bound. (See Paul Krugman’s demolition of the yield curve indicator here.)
The red line might be more informative, since it contains only economic and market data which have in the past been indicators which suggest that the economy is already in recession. This indicator, which does not include the yield curve, reckons that there is about a 30 per cent chance that forthcoming data will reveal that a recession has already started. Although we have seen false spikes in this indicator in the past, especially early in long recovery periods, this is clearly something which needs watching.
In conclusion, it seems clear that the US economy has slowed down enough this year to be flirting with its stall speed, without actually breaching it so far. Since the reasons for the slowdown include temporary factors like the Japanese earthquake and severe weather conditions, the most likely out-turn is that the economic growth will remain above the stall speed in the remainder of 2011, as expected by the Fed and the forecasting consensus. But there is not that much room for error.
