This has been the monotonous story for more than two years. Just to mention two very recent posts: Christian Menegatti (Aug 29) asks “where is this recession already”, while David Wyss (Aug 29) sees “recession returning in the third quarter”. All this reminds me of the popular Looking for Wally game from seven or eight years ago: fun for a while but fast reaching decreasing returns.
I find the alternative question: why, in spite of the powerful cocktail of shocks – oil, commodities, housing and credit – recession is not yet in the data much more interesting. If the so called great moderation of the last twenty five years was for many the result of luck in the form of blander shocks since the early 80´s, the shocks experienced in the last eight years that include, in addition to those mentioned, the bursting of the internet bubble in 2000, the terrorist attacks of September 11, the unveiling of the corporate shenanigans and a host of geopolitical manifestations, certainly cannot be classified as “weak”.
Figures 1 and 2 show the marked differences in the macroeconomic outcome – inflation and growth – observed in the 1970´s (“Great Inflation”) and more recently (“Great Moderation”) following the energy shocks in the two periods.
Professor John Taylor has for the last ten years argued in favor of the improved quality of monetary policy as bearing most of the responsibility for the observed outcome. On the other hand, for many years several analysts confused much lower (more than 50%) output volatility in the last twenty five years with subpar growth, despite the fact that mean growth since 1984 is not statistically different from that observed in the previous thirty years!
Granted, the US economy is much less dependent on oil today then thirty years ago; nevertheless since oil started its persistent climb in 2002, shortly followed by commodity prices and more recently by the housing and credit market problems, recession and/or stagflation predictions have become prevalent. Maybe that´s the problem: prediction. Everyone is after the title “Was the first to predict…”, forgetting to understand the underlying process; that´s harder and of little interest to the media.
Take, for example, house prices. Since the housing problems erupted, everyone has latched on to the S&P Case-Shiller index, mostly because it is more conformable to the “bubble and burst” view of house prices. Figure 3 attests to the large differences between the indices, both in levels and rates of change. Different methodologies in the construction of the two indices help explain most of the differences between them. The C-S index is value weighted; more expensive houses have greater weight. The OFHEO index on the other hand is unit weighted. While C-S gives greater weight to census regions that have higher priced homes (California and the Northeast, for example), the OFHEO index distributes weights according to the number of residential units.
The methodological differences mentioned above allow us to understand, for example, the different behavior of the two indices during the first half of the 1990´s. While the C-S index was flat (even falling somewhat), the OFHEO index was growing.
With the S&L crisis, the West (California) and Northeast were the regions most affected. Those are also the regions with higher priced homes (in no small measure due to zoning laws that restrict supply). When the housing market picked up again, supply restrictions (while housing starts blossomed in the South, for example, they were stagnant in the Northeast) pushed the rate of change in prices up strongly in those regions (W and NE) that have greater weight in the C-S index. Just as house price increases were magnified in the C-S index so we now observe the reverse effect due to price falls.
The upshot is that the magnitude of the wealth effect based on house price falls as calculated by the C-S house price index may be grossly overestimating the true wealth effect being experienced by homeowners.
This brings us naturally to consider the consumer – king of the US economy “owning”, through his expenditures, a 70% share of the economic pie. In most analysis one speaks of consumer expenditures bundling together durables with non durables and services. But that can be quite misleading since the behavior and determinants of the two kinds of expenditures are quite different. Their respective shares are also very different: about 60% of GDP for non durables and services and about 10% for durables.
Figure 4 indicates that the “great moderation” had a very positive effect on consumer durables expenditure. The decrease in economic uncertainty extended planning horizons increasing demand for durables and both the demand and supply of credit.
The start of the “Great Moderation” also coincided (not coincidentally) with the surge in (productive) financial innovations, beginning with Michael Milken´s early “junk bonds” followed by the development of MBS´s and graduating to CDO´s and such. These are all good developments and were certainly factors behind the “surprising” economic outcomes experienced in what came to be known, in the words of Joseph Stiglitz the “Roaring Nineties” or, less euphemistically, according to Blinder and Yellen, the “Fabulous Decade”.
Competitive pressures made each new product more complex. That, together with increasing leverage and tight coupling made the system more risky (much like the launching of spaceships and workings of a nuclear plant). The ensuing problems are being addressed and will follow their course, but they should not be viewed as the end of the world or the harbinger of a great disaster. In some sense the US is lucky to have a specialist in credit crisis as Fed Chairman!
Figure 5 plots growth in GDP, consumer durables expenditure and nondurables and services expenditures. A few things are notable:
1. 1. After 1984 all growth volatilities are significantly reduced.
2. 2. Durables expenditures are relatively volatile.
3. 3. Except for the 2001 recession, consumer expenditure growth (both durables and nondurables and services) drops significantly during recessions (shaded areas). All the fallout in 2001 was due to the marked decrease in nonresidential investment as a share of GDP.
4. 4. Except for the 2001 recession GDP always turns negative (year on year) during recessions.
5. 5. Less evident is the fact that potential growth appears to have decreased after the 2001 recession from what it was in the second half of the 90´s.
For the first time since 1992 in the second quarter of this year growth of durables expenditures in real terms has turned negative (-1.1%). The credit crunch has reached the consumer. But what about the 60% share of nondurables and services; what is it telling us about consumption (excluding durables) and GDP going forward? There I go giving a stab at prediction!
Figure 6 is a visual impression of the cointegration property of consumption (ex-durables) and GDP. Cointegration means that two series evolve in a parallel fashion, or that they possess a common trend. Individually they are non stationary, but their ratio is stationary, i.e. it doesn´t show a trend, fluctuating around a constant mean value. Formal statistical tests confirm the cointegration properties of consumption and GDP.
For this to happen there must be an adjustment mechanism that keeps them “together” (much like track beds and nuts and bolts keep train tracks parallel). Since we are not dealing with physical objects like train tracks, the important question is: which series bears the brunt of adjustment – consumption or GDP?). Put differently, when the consumption (C)/GDP (Y) ratio departs from its (stationary) mean, which variable moves to bring the ratio back towards its mean?
A visual interpretation can be gleaned from figure 7, which shows that whenever the C/GDP (Y) ratio departs from its mean (60.3%) GDP growth adjusts (i.e. decreases when C/GDP (Y) is below the mean and increases when it is above). From this we may say that consumption is useful to predict income (GDP).
Since 2003 the C/GDP (Y) ratio has remained pretty close to its mean. This is an indication that GDP has evolved very close to potential (small fluctuations or more stable economy despite the intensity and severity of shocks), and since C/GDP (Y) is still slightly above its long term mean, this augurs well for GDP in the third quarter.
It´s funny to observe that recently potential GDP has been revised down from something close to 3.5% that prevailed for most of the 90´s to something closer to 2.5%. Even so, there is a widespread belief that growth has been anemic! Even funnier, in the 90´s when most believed that potential was close to 2.5%, it was generally thought that the economy was growing too fast and that the Fed was behind the curve (see, for example Krugman and his “Speed Trap” series in Slate and HBR in 1996/97). It seems that the economy, one way or another, is always a source of frustration and that this frustration will be validated in the “next quarter”!
 Arthur Laffer (of Laffer Curve fame) has a nice argument, explaining how international trade (in addition to financial innovations) was able to “transform” an excess supply of housing in Japan and several European countries after their housing bust in the early 90´s into an increased supply of housing in the US in order to satisfy the local excess demand for housing. Now it could be that the US will “export” its excess supply of houses and durables and so “improve” net exports.
 The theory behind this result goes by the name of Permanent Income Rational Expectations.
 Notice that early in 2007 when rose, growth in the second and third quarters was “surprisingly” high.