My brother is smart, so smart.
I love my brother; I send him money every week.
-- David Brenner
Cause and EffectLet's pose a Clintonesque retort to the key question on which market to believe: "Depends on what you mean by smart." After all, the stock and bond markets have different objective functions. Bonds need to assess the probabilities of default and inflation, and unless they're convertible, they have a fixed upside. Bonds tend to be purchased on the motivations of income and capital preservation. Stocks are exposed to the same forces of business failure and inflation, but they have a substantial upside. Their purchase is motivated by growth.
|The Moody Blues |
|Source: Federal Reserve|
If we restate the data above to display the spread as a percentage of the underlying Treasury yield, (corporate - Treasury) / Treasury, the picture changes. We are still within an upward channel of rising credit spreads that began, incredibly, with the breakout to the upside of the stock market in January 1995. Equally surprising in this chart is how the dreadful stock market of the 1970s coincided with a period of decreasing corporate spreads as a percentage of the Treasury yield.
|Spreads Not as Low as They Look |
|Source: Federal Reserve|
The ResultsThe monthly returns on the two corporate bond series and on the S&P 500 were analyzed from the beginning of 1928. A simple visual inspection of the series on a coincident basis didn't hold out a lot of promise for a strong statistical relationship. That a stronger relationship between lower-quality bonds and stocks existed compared with the one involving higher-quality issues should surprise no one.
|Coincident Monthly Returns, 1928-2003 |
|Source: CRB Infotech|
When the data were examined for Granger causation, the results were similarly inconclusive. None of the cross-correlations were as strong as the autoregressive (the series itself) correlations. We can't say that corporate bond returns cause stock returns or vice versa. All we can say is that the stock market does a better job of forecasting the corporate bond market -- R-squares of 1.2% and 4.7% for the AAA and Baa, respectively -- than either bond market does for stocks, 0.4% and 2.7% for the AAA and Baa, respectively. The final tidbit is that the S&P 500's autoregressive R-squared of 7.9% is higher than that of either bond market.
|Forecasting Advantage |
Stocks get the edge?
|Source: Federal Reserve, CRB Infotech, TSC research|
On the basis of these data, we have to grant stocks an advantage over corporate bonds as the smarter market. Not a win, but an advantage, one derived from a single test over a very long sample and at a monthly frequency. But there's no crying in baseball, and no ties, either. Stocks get the win by default in this, the first game of the World Series. I'll be back shortly with other tests.