“The four most dangerous words in investing are ‘This time it’s different.’”-Sir John Templeton, yet a nuanced approach tells us that times are often different. The current talk about a recession may be one of those different times. Many factors are the same with identifying a recession but the causes seem to change. The credit crunch has had an impact on the real economy, but for what seems like the longest time the sectors outside the housing industry have been doing reasonable well. Consequently, some of the models that try and provide a probability of a recession have not sent clear signals of a potential recession.
The quantitative approaches are in some cases being more cautious then opinions of market analysts. The number of stories concerning a US recession has increased substantially if you look at the internet and the number of economists who are not seeing a recession in 2008 is getting fewer and fewer, yet when you look at objective measures on the probability of a recession based on the behavior of past economic data the story is mixed. The betting market in the Intrade betting exchange shows that the odds of recession are very high. This may be a better measure of public opinion than looking at the number of stories of recession.
The yield curve probability model shows much different results than what economists are thinking. Up until this summer the yield curve was flat to slightly inverted and had been calling for a greater than 20% probability of recession for almost a year. It was almost getting to the point that the yield curve model was giving a false signal. Only now are we getting the slowdown that has been expected, but the steepening of the curve now makes the probabilities start to change and move to a low chance of a recession. By this measure the Fed has reacted and has taken action which will change the course of the real economy by the end of the year.
More complex recession probability models are showing mixed results. The Jeremy Piger model of recession probability has inched up but is nowhere near levels that would suggest that we are in a recession. For more details, look at his homepage http://www.uoregon.edu/~jpiger/. Piger is a former economist at the St Louis Fed. Coincidently, the other model used to forecast recession probabilities is also by a former economist from the St Louis Fed, Michael Dueker, who is now at Russell Investments. He developed a qualitative VAR model for recession forecasts and finds a much higher probability that is above his threshold for a recession. He is calling for a two quarter recession for the first half of 2008. His work can be read in more detail on econbrowser.com;
http://www.econbrowser.com/archives/2008/02/predicting_rece.html. His approach employs a business cycle index with a vector autoregressive model to find recession probability forecasts. He uses GDP growth, core CPI inflation, the slope of the yield curve and the Fed funds rate. The fed funds rate and the slope of the yield curve are common to all of these models so the action is in the fundamental data. He notes that there is a difference with a recession recognition model which have not shown high probabilities and recession prediction models like his which are giving clear signals.
What you do find is that recession probability models can move from low or no probability to an extremely high number in a very short time. The probability of recession is also based on the data being used because there just are not that many recessions to look at. Each one may have unique features for its cause, but if you leave off some of the data from the 1970’s you will get different probabilities.
The nice thing about these types of models is that they truly give a quantitative measure of a recession chance and can be the basis for meaningful discussion on the particulars of what makes this time different. Unfortunately, the numbers are truly leaning to a recession regardless of how you cut the data.
No comments:
Post a Comment