The Next Six Months Have Promise
11/14/2011 8:15 am EST
Using a newly developed indicator, it looks to Elliott Gue of Personal Finance like there will be some solid buying opportunities in the coming months.
The stock market is often a leading indicator for the economy. Since 1929, the US economy has experienced 14 recessions, and all but one of those downturns—a short contraction in 1945—was accompanied by a pullback of 15% or more in the S&P 500.
On average, stocks topped out about four months before the economy slipped into recession and bottomed five months before the economy hit its trough.
The Great Recession of 2007-09 is a good example of the stock market’s predictive powers. The S&P 500 peaked in October 2007, two months before the economy officially entered recession. The index then fell to multiyear lows in March 2009, about four months before the economy exited the downturn.
There is no perfect economic indicator. Economist Paul Samuelson once famously quipped that the stock market has forecast nine of the last five recessions. Since 1980 alone, the S&P 500 has experienced severe corrections in 1984, 1987, 1998 and 2002…and none of these preceded economic contraction.
Nevertheless, a statistical technique known as logistical regression can transform stock-market performance into a fairly reliable economic indicator. Armed with a half-century of economic data, I’ve constructed an equation—known as a logit model—that estimates the probability of recession over the next six months.
Three recessions have struck the US economy since 1990, and my model showed a greater than 50% probability of recession in the early stages of the 1990 and 2007-09 recessions. It also predicted the 2001 downturn months ahead of time.
This indicator hasn’t eliminated false alarms…the model incorrectly called for a recession in 1998 and 2002. But when the predictions were correct, the signals came early enough to spare investors the bulk of the market’s downturn.
The stock market isn’t the only leading indicator of economic performance. I also closely watch initial jobless claims, which spike just before or in the early stages of an economic downturn.
I constructed a similar logit model based on jobless claims data stretching back to the 1960s. The logit model based on initial claims spiked above 50% in the early stages of all three recessions since 1990.
Although this model called the 1990 and 2007-09 contractions later than the stock market did, it actually predicted the 2001 downturn months before the stock-market model.
Even better, initial claims never spiked enough to indicate recession in 1998 or 2002—an investor following this model might have correctly seen both as stellar buying opportunities.
The stock market and jobless claims are flawed indicators. But together, they provide a more accurate picture of the actual risk of recession.
These are just two of the data sets I used to create the PF Recession Radar. This proprietary metric also tracks eight additional economic and financial indicators that cover credit and commodity markets, monetary conditions, manufacturing activity, and confidence.
Of course, no model is perfect. And investors should not rely solely on any single indicator when making investment decisions. But the PF Recession Radar is a quick and dispassionate gauge of the risk of an economic contraction over the ensuing six months—another tool that will help us take the necessary steps to reduce risk.
At this point, the Radar is posting a score of 25%, which puts us in a medium-risk environment over the next six months. When the model hits the upper 40s, it’s time to start worrying. Until then, oversold stocks in promising sectors can be good bargains here.
- 2012 Could Be a Good Year for Stocks
- New Dangers of Credit-Default Swaps
- Barrel Into This Large-Cap Fund