Princeton Economist: Why Don’t My Equations Work?
By Vedran Vuk
Yesterday, I mentioned the futility of utilizing mathematical equations to model economics and the social sciences. Here’s a great example of what I’m talking about in a Bloomberg article by Alan Krueger, professor of Economics at Princeton:
From late 2008 to the end of 2009, the U.S. unemployment rate jumped by 3.1 percentage points. What was so shocking about this was that the relationship between unemployment and the drop in the gross domestic product during the recession would have predicted only a 1.2 percentage point increase. This was the largest forecasting error in decades.
Since then, the baffling moves have been in the opposite direction, with unemployment falling from 9.8 percent in November to 9.4 percent in December. Since then, the unemployment rate has dropped even further, to 8.9 percent in February. We’ll get the March figures tomorrow.
Regardless of whether the rate ticked up or down, the 1.2 percentage point drop in the unemployment rate from its peak in 2009 is a puzzle because economic growth hasn’t been strong enough to generate such a large decline.
Let me translate. Krueger put variables into his math equations. The number that came out on the other end had nothing to do with reality. How shocking!
Actually, it’s not shocking at all to anyone who realizes that the social sciences are not physics. One can’t plug numbers into an equation and actually expect results to reflect reality. Economic variables and their relationships are always changing. Furthermore, the economy is the complete opposite of a controlled experiment. It’s a chaotic storm of different variables and factors determining an uncertain future.
Now, of course, economic modeling has some value. At Casey Research, we clearly use mathematics for all sorts of things. However, an equation alone will never determine our investment recommendations. Unlike Princeton’s Alan Krueger, we are not shocked to find that math models fail to meet reality.
Despite these obstacles, many investment companies rely more heavily on mathematical models, and they can do so at their own risk. The real problem occurs when national decisions are based on the same approach and decision-makers have the utmost trust in a fallible system. Do you remember the promise that the bailouts and stimulus were supposed to keep unemployment below 8%? How do you think they came up with that number?
If an investor messes up an equation, they’ll get hurt. If the Federal Reserve and government mess up an equation, everyone gets hurt. Despite much advancement in econometric techniques, economics isn’t an exact science and never will be.
Simply put, there are no equations that can consistently predict future economic events. If this wasn’t true, Bernanke would not have been recruited for the Fed position from a professor post. Instead, the Fed would have begged him to give up his life of luxury as a billionaire.