Eichengreen and O’Rourke have produced three columns (6 April, 4 June, and 1 September 2009) comparing today’s crisis to the Great Depression via the synoptic analysis of world industrial production, trade, and stock market. Their main conclusion in the latest update was that the most recent crisis remained dramatic on 1 September 2009 by the standards of the Great Depression.
Bloom wrote two columns on 4 June 2008 and 12 January 2009 concentrating on uncertainty, as measured by stock market volatility, as a cyclical economic indicator. The evidence at the time of the first column led the author to label the situation as “a perfect storm” resulting from “…a huge surge in uncertainty that is not only generating a rapid slowdown in activity but also limiting the effectiveness of standard monetary and fiscal policy to prevent this”.
However, at time of the second column, the reduction in uncertainty led the same author to choose the title “The recession will be over sooner than you think”.
It seems interesting to add the synoptic analysis of the evolution of uncertainty over the two crisis periods to the indicators selected by Eichengreen and O’Rourke.
In this short column, I propose a measure of uncertainty that closely covaries with the stock market volatility and it is easily comparable in the relevant historic periods. This indicator suggests some interesting developments in the “tale of two depressions”.
Measuring uncertainty: The BAA-AAA spread versus VIX
My proposed measure of uncertainty is the spread between Moody's Seasoned Baa Corporate Bond Yield and Moody's Seasoned Aaa Corporate Bond Yield, available via the FRED database and the Federal Reserve Bank of St-Louis. The spread between bonds with the same maturity issued by firms with different credit ratings measures directly the premium that uncertainty commands in the bond market. This measure has the advantage of being homogeneous over the long relevant sample (whereas realised volatility has to be combined with implied volatility in options (available only for the most recent crisis) to have stock market volatility measures for the same periods), while keeping the same properties of the stock market volatility in terms of co-movement with output growth.
Figure 1 reports the BAA-AAA spread along with the VIX volatility taken from Bloom, Floetotto, and Jaimovich (2009) over the period 1960-2009. The two series strongly commove and have the same cyclical properties during NBER-dated recessions. The main exception to the co-movement of the two series is the spike in stock market volatility in October 1987, which it is not matched by any similar movement in the BAA-AAA spread. Interestingly, that strong spike in stock market volatility was not mirrored by any major fluctuation in GDP growth. Figure 2 demonstrates the relevance of the BAA-AAA spread as a cyclical indicator, especially during recessions, by showing the co-movement between the spread and annual GDP growth.
Figure 1. Two measures of uncertainty
Figure 2. Uncertainty and economic fluctuations
The BAA-AAA spread from 1928onwards and 2006 onwards
In the light of that evidence, it might be interesting to apply the Eichengreen and O’Rourke synoptic analysis to the BAA-AAA spread. I do so in Figure 3 by plotting the monthly BAA-AAA spread from June 1928 forward and from April 2006 forward. The results suggest a striking difference between the situation now and in the early 1930s. The evolution of uncertainty suggests that the current situation is much less dramatic by the standards of the Great Depression.
Figure 3. The BAA-AAA spread: Today vs. the Great Depression
Acknowledgment: This column was generated by discussions with Paolo Mottura, to whom I am very grateful.
Bloom, Nicholas, Max Floetotto, and Nir Jaimovich (2009), “Really Uncertain Business Cycles”, Stanford University mimeograph.