Lies, Damned Lies, and Statistics

In his most recent column in the New York Times David Leonhardt presents the history and inadequacies of some of the official economic statistics. It’s worth reading just for the history but I want to underscore his peroration:

The whole point of statistics is to describe reality. When a statistic no longer does so, it’s time to find a new one — not to come up with a convoluted rationale that tries to twist reality to fit the statistic.

The notion that our most prominent economic indicators are problematic has been around for a long time. Kuznets himself, the economist who invented G.D.P. as we know it, cautioned people not to confuse it with “economic welfare.” Most famously, Robert F. Kennedy liked to say during his 1968 presidential campaign that G.D.P. measured everything “except that which makes life worthwhile.”

That so many people especially people who quote statistics have so little understanding of basic facts about statistics (like the difference between the average and the median or what a standard deviation is) doesn’t help. And there is no way to depoliticize economic statistics. They are inherently political.

The Dow-Jones Industrial Average is a lousy proxy for national economic welfare and has become worse over time. Gross domestic product is also a lousy proxy for national economic welfare for the simple reason that it is an aggregate. Even the famous question, “Are you better off now than you were four years ago?”, has its problems.

All of this is not to say that statistics are unimportant. I’m just suggesting that they must be understood in perspective.

3 comments… add one
  • Guarneri

    As I’ve pointed out, GDP was originally just a measure of output, not general welfare. I’m not sure when it morphed into the be all and end all. It’s not.

    That said, I’m not sure a better one exists. Probably better to look at a number of stats and weight them by judgment and personal values. But that takes work, and it’s icky and subject to politics.

  • TastyBits

    … the average and the median …

    When somebody says “pass the salt”, I am fairly certain they do not want baking soda, but technically, it would be correct.

    I bothers me that most people do not know the difference between a second cousin and a first cousin once removed. I correct my wife, but I know what she means.

  • The reason that the different between mean and median is important, obvious to you and me, is that it is quite possible for the average income to increase while the median income remains the same or even decreases and those would not be benign eventualities.

    The average increasing while the median stays the same or decreases probably means that the tails on both sides of the distribution are getting longer, i.e. rich people getting richer and poor people getting poorer.

    I don’t think that should be an objective of policy but, sadly, it appears to have been. NAFTA is the clearest example. In fairness I guess there’s a more charitable explanation: ideologically-blinded stupidity.

Leave a Comment