Lies, Damned Lies, and Statistics

If you've been spending time tracking economic indicators such as unemployment, capacity utilization, and monetary aggregates to divine what will happen with inflation, maybe you should just get a Ouija™ board instead.


If you’ve been spending time tracking economic indicators such as unemployment, capacity utilization, and monetary aggregates to divine what will happen with inflation, maybe you should just get a Ouija™ board instead. According to a recent study, such indicators, in isolation, are no more effective at predicting inflation than simply using past inflation figures to predict future trends.

The study, authored by Stephen Cecchetti, a professor at Ohio State University and a former director of research at the Federal Reserve Bank of New York; Rita Chu, a graduate student at New York University; and Charles Steindel, a senior vice president of the Research and Market Analysis Group for the New York Fed, confirmed what they already suspected, according to Steindel. “In my view, this all just means that the economy changes all the time and you can’t rely on things you relied on in the past,” he says.

In an ironic commentary on the uselessness of such indicators, the study showed that one of the most effective single inflation indicator variables was the commodity price for gold and other precious metals–but it works in reverse. When the price of gold went up, inflation went down. Another relatively effective predictor was the Journal of Commerce price index for industrial material, which outperformed autoregression, or extrapolations based simply on past inflation patterns, 9 times out of 13.

So, what if you’ve been waiting for the next Fed decision on interest rates before building that new headquarters building? The bad news is that the Fed has traditionally tracked those indicators closely to help determine its monetary policy. This study suggests that the Fed may be tilting at windmills by watching for signs of inflation in those numbers.

“It’s hard to generalize about what the Federal Open Market Committee relies on to make its decisions,” says Steindel. “There are 12 principals sitting around a table, and they all see the world in different ways. They all have different views on what the major forces are.”


Another group, the Mark Twain Institute, a think tank in Washington, D.C., has taken a swing at the traditional, manufacturing-based indices used to predict growth in the GDP. And the Institute’s founder, Harry Freeman, claims that forecasters rely too heavily on non- service-sector variables, even though the GDP is now made up largely of service components.

“Services often aren’t measured as part of productivity, because it’s too difficult to measure,” says Freeman, a former tax attorney who was an executive vice president at American Express Co. “It permeates all of our policy-making. The Fed is working hard to get good statistics, but they would have a better handle on it if they had better service statistics.”

The Institute, named to honor the quote attributed to Mark Twain, “There are three kinds of lies: lies, damned lies, and statistics,” has created a new index, called the Mark Twain Economic Indicator (MTEI), which predicts GDP growth nine months out based on a set of eight variables, including service-sector indicators such as the total number of telephone minutes used, and the Manpower Inc. hiring survey, which is largely service-based. The other component is GDP trends in the previous nine months.

According to Freeman, the index has been back- tested over the past 10 years, and has proved “remarkably accurate.” But those who are responsible for the other indicators question whether the MTEI passes the “so what” test.

“I don’t think it’s that big a deal,” says Ken Goldstein, an economist for The Conference Board, which generates the Leading Economic Indicators, which seeks to predict short-term economic development. Goldstein admits that a big part of the reason The Conference Board does not aggressively track service indicators is because they’re difficult to quantify. “Plus,” he adds, “this is another instance where, yes, there’s a new economy, but there’s not much new about it.”

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>