Uncertainty and the credit crisis

Michelle Alexopoulos, Jon Cohen

23 December 2008

a

A

It’s official. As everyone now knows, the US economy is in recession and has been since December 2007. If the contraction continues for another four months, which at this point seems inevitable, this downturn will match the two longest peak to trough slides in the Post-WWII period, the first from November 1973 to March 1975, the second from July 1981 to November 1982. Whether the current recession achieves the dubious distinction of matching unemployment rates of the earlier ones (9% in May 1975 and 10.2% in November 1982) remains to be seen, but the dramatic rise in unemployment announced on 5 December 2008 is worrisome. As for the stock market, the decline of the Dow Jones Industrial Average from its peak in July 2007 to its low point on 20 November 2008 actually exceeds by a few percentage points the 40% drop between October 1972 and October 1974. And, of course, the catastrophic fall in house prices continues unabated. In short, the economy is in serious trouble and it is likely to get worse before it gets better.

Nick Bloom, in a recent Vox column, uses links he has identified in his academic research between uncertainty shocks as measured by changes in expected volatility of the S&P 100 – the so-called investor fear index – and GDP growth to predict the length and depth of the current downturn. He predicts, on the basis of a dramatic jump in expected volatility caused by the credit crunch, a GDP decline of 3 percentage points in 2009 with recovery starting at the very end of year, assuming favourable government policies and a drop in volatility. Fear and uncertainty with all its associated collateral damage – postponed investment, limited structural change, and delayed consumption – indeed would seem to stalk the land.

From Wall Street to Main Street

Bloom’s argument depends heavily on the reliability of the expected volatility index as an indicator of uncertainty. Although his research results are compelling, it is still reasonable to wonder if his results are sensitive to his uncertainty measure. Or, to put it another way, are the forces that shape expectations among the Wall Street crowd the same as those that affect the folks on Main Street? In short, would the use of a more broad-based indicator of uncertainty alter the observed link between uncertainty shocks, output, and productivity? This is the question we are addressing in our current research. In brief, here’s what our preliminary results tell us.

Uncertain times, uncertain measures

We base our index of economic uncertainty on the number of articles that appear in the New York Times which use the terms uncertain and/or uncertainty and economic and/or economy. The beauty of the measure is that it is consistent over a very long time span (the New York Times searchable data base extends back into the nineteenth century), it is transparent and unfiltered (as they say, all the news that’s fit to print), and it approximates what the average Main Street resident knew about current events. In Figure 1, we present our monthly uncertainty index (adjusted for the days of each month) with NBER business cycle reference dates in the background, to show it adheres closely to business cycle dates. Moreover, as Figure 2 illustrates, the timing of the uncertainty shocks identified by Bloom’s uncertainty index (based on S&P volatility) and ours are quite similar.

Our statistical results suggest that uncertainty shocks act on economic activity with remarkable swiftness (the shock has an almost immediate negative impact on growth and productivity), strength (they explain over 25% of the variance of output and productivity within two years), and durability (the effects linger for a number of quarters). Moreover, the current uncertainty shock - that effectively dates from the Bear Stearns bailout - is the largest of the twentieth century, greater even than that associated with the October 1929 stock market crash.

The dreaded D-word

Of course, the obvious question is what does all of this mean for Main Street and its inhabitants? Many are now prepared to put the current crisis in the same league as the dreaded Great Depression. They point out that in both periods there were significant bank failures, sharp declines in equity and housing prices, and a severe credit crunch. However, the current policy responses have been vastly different, in large part because Federal Reserve Chairman Bernanke, an expert on the Great Depression, has used his deep knowledge of that event to avoid the errors of the past. Unlike in the 1930s, the monetary authorities have moved swiftly to increase liquidity, push down interest rates, and bolster the stability of the financial system. As it happens, our regression results suggest that these policy responses make a big difference. That is, when we introduce interest rates (the policy variable) into our regressions, we find that the economic contraction is likely to be closer to the 1% predicted by the OECD than to the 3% predicted by Bloom.1

Could the worst be over soon?

In spite of the very real threats to the US (and world) economy, a little perspective is in order. First, we have survived sharp jumps in uncertainty in the past – July 1971, January 1991, September 2001 (see Figures 1) – and will do so again. In this respect, it is also worth noting that the pattern displayed by our uncertainty index for the current crisis resembles more the sharp, short-lived ups and downs of the 1970s and early 2000s than it does the long drawn out rise and sluggish fall of the Depression years. This would seem to suggest that the current crisis, despite its gravity, does not mark the end of the world as we know it. Second, in keeping with the old adage that it is often darkest just before the dawn, the numbers in Table 1 indicate that the light of a new day may just be visible on the horizon. Our uncertainty index, in this case based on data from six major US newspapers, shows a sharp run-up in uncertainty through October 2008 and a modest decline since. Two months do not make a trend but the drop is definitely encouraging. Although the negative economic consequences of the severe shock are likely to dog the economy for some time, we would guess that the worst is, indeed, behind us. Or, to employ Bloom’s horror film metaphor, the credit crisis has us (with good reason) perched on the edge of our seats, white-knuckled and wide-eyed. But, it is well to remember that the heroine while a little worse for wear, usually lives to welcome the dawn of a new day.

Newspaper Average daily number of articles with keywords
("uncertainty” or “uncertain” & “economic” or “economy”)
Average week-day circulation
  2007 2008a Selected months 30 Sept. 2008
      October November Decembera  
New York Times 1.09 2.33 4.13 3.87 3.11 1,000,665
LA Times 0.65 1.07 2.00 1.27 1.44 739,147
USA Today 0.23 0.47 0.68 0.60 0.75 2,293,310
Wall Street Journal 1.98 3.38 5.29 3.43 2.56 2,011,999
Washington Post 0.76 1.58 3.48 1.90 2.22 622,714
Chicago Tribune 0.56 1.07 1.32 1.60 1.67 516,032
Circulation-weighted average 0.95 1.75 2.88 2.10 1.85 na

a. Values reported for 2008 are through 9 December 2008.

References

Alexopoulos, M. and Cohen, J. 2008. Uncertain Times, Uncertain Measures. Manuscript. University of Toronto, 2008.
Bloom, N. 2007. The impact of uncertainty shocks. National Bureau of Economic Research, Working Paper W13385. Issued in September 2007.
Bloom, N. 2008. The credit crunch may cause another great depression. VoxEU, 8 October 2008.
Bloom, N. 2009 will be the Nightmare on Main Street. VoxEU, 18 November 2008.
OECD. 2008. Economic Projections for the US, Japan & Euro area. Press Conference 11 November 2008.


1. Our predictions are made from standard Vector Autogression (VAR) forecasts. While our bivariate analysis suggests a decline of approximately 3%, the addition of interest rates into the system cut the forecasted decrease to 1%.

 

a

A

Topics:  Global economy

Tags:  credit crisis, uncertainty, US economy

Associate Professor of Economics, University of Toronto

Emeritus Professor of Economics, University of Toronto