Monday, November 17, 2008

The Extremely Anxious Index

The Philadelphia Federal Reserve Bank released its Survey of Professional Forecasters for the fourth quarter of 2008. As part of this survey the Philly Fed creates the Anxious Index. The Anxious Index is the forecasters probability of a recession occurring in the next quarter. The advantage this survey has over models based on the yield curve is the yield curve model assumes banks are borrowing short and lending long, thus stimulating the economy. This has not been the case recently and is a huge shortcoming of any yield curve forecasting model.

The Anxious Index for Q4 2008 came in at the highest reading in its history. Professional forecasters see a 74.78% chance that the economy will be in recession the next quarter. The last time the reading was this high was in the fourth quarter of 1974 when the index reached a high of 74.06%.

The Philly Fed also asked a few extra questions this quarter, in particular, they asked when the forecasters thought the recession had started and how long will it last. The forecasters believe that the recession started in April of 2008 and will last for 14 months. 31 of the 51 forecasters surveyed stated that the recession estimates included the potential for a new fiscal stimulus plan.

The following chart shows the Anxious Index since its inception in 1968 with recessions in red. Historically a reading above 40% probability has been a good indicator that a recession has begun. The Anxious Index reached 42.92% in the first quarter of 2008 which is consistent with the forecasters belief that the recession started in April of 2008.



Using the forecast of a 14 month recession would put the end of the recession coming in June 2009. The Anxious Index has been accurate in predicting recessions one quarter ahead, however, several Federal Reserve studies suggest that the accuracy significantly deteriorates after one quarter.

While this is useful information there are two shortcomings of this survey. First, the professional forecasters use the yield curve in their forecasts, therefore some of the inputs to the forecast are based on flawed assumptions. Second, this survey only goes back to 1968 and does not take into account forecasts during the Great Depression which are probably more applicable to the current economy.

The best way to use this survey may be to think of it as a map without road names. It may tell you where you are and what direction to go, but it will not give a directions to a specific address.

To that end, looking at the Dow Jones Industrial Average in 1974 one finds remarkable similarities. Coming into the fourth quarter of 1974 the DJIA was down over 30% for the year and the Anxious Index was above 70%. On December 13, 1974 the DJIA hit a low of 570.00 which turned out to be the absolute low of the bear market and the beginning of the bull market that lasted until 2000. Interestingly, the National Bureau of Economic Research (NBER) stated that the recession ended in March of 1975, one quarter after the Anxious Index hit an all time high.



Overlaying the DJIA in 1974 and 2008 illustrates the similarities between these years. A simple linear regression reveals the two markets had a correlation coefficient of 0.81 and an R Squared of .66. Perhaps not enough of a match to develop a tradable model, but the information can be used as a guide. Using the past as a guide (always a dangerous and thrilling assumption) we might expect the DJIA to retest the lows and then begin a bull move. If the retest occurs one quarter before the economy turns in June 2009 (according to the Philly Fed Survey) one might expect the DJIA to bottom around March 2009.

Looking further back in history, the Great Depression ended in March 1933 according to the NBER. The DJIA bottomed in July of 1932, two and half quarters before the depression ended. Comparing the twelve month period before the July 1932 bottom and 2008 reveals similar results to the 1974 comparison.


The correlation coefficient of .73 suggests that this period of time could be useful as a guide to the 2008-2009 market. Using 1932 as a guide (here we go assuming again) and a two quarter lead for the bottom of the DJIA, one might expect that the DJIA is currently in the bottoming process.

The point of this exercise is not to attempt to pinpoint a bottom, but to illustrate that the market turns before the economy and well before economic forecasts. In fact, the extreme readings of the perception of economic weakness could be used as a contrary indicator. This is a classic example of perception diverging from reality.

Disclosures: I am short SPY.

No comments:

World Clocks