Friday, October 26, 2007

Three rules for reporting on statistics 

One thing you try to teach students is how to put numbers in context. Based on the idea that "figures don't lie, but liars sure can figure," part of educating citizens in a democracy, and of critical thinking, is to ask questions about numbers presented by political candidates and media.

Some simple rules I teach:
  1. The Carl Sagan rule. For older readers, what's the one phrase you remember from him on the Tonight Show with Johnny Carson? "Millions and millions" of planets/stars/galaxies... The idea was to imply "a lot, a real lot, bigger than you can imagine." The use of large numbers is usually meant to impress or scare you. So for example, remember yesterday when I mentioned that the subprime mortgage crisis would cost $14 million in local property tax revenue in Minnesota? You should ask "out of how much total is that?" The answer is that total property tax revenue collected in the state is almost $7 billion, so that the hit is around one fourth of one percent. (That is .0025% (thanks for the correction, Dale and Peter!) for those of you scoring at home.) Doing things in terms of shares (like spending or taxes) is a help in putting things in the right context.
  2. The "keeping it real" rule. You might think this is a simple matter of deflating everything by CPI, but that is a bit difficult. What's the real cost of a personal computer now versus one five years ago? That question is almost nonsense, because the computer on my desk five years ago does nowhere near what mine does now. One of the things I really like to do with this is to show students some of the annual reports of the Dallas Fed, written by Michael Cox and Richard Alm, which attempt to put the cost of things in something that really is constant, like human effort. Or by understanding that productivity at work isn't just measured by output per unit of input, but includes the quality of time at work and at home. The point is that to really keep it real, you need a real numeraire that people can hang on to, not just some "absence of money" definition. Keynes wrote the General Theory using the wage rate as his numeraire. Austrian economists have a point in referring to the use of a numeraire as "misleading". I don't eschew one, but it should be used with care.
  3. The adjective rule. Pay very close attention to the adjective used in defining the term to be measured. Take yesterday's editorial in Investor's Business Daily about President Bush's spending. Spot the adjective:
    Bush is "arguably an even bigger spender than LBJ," says a story from McClatchy Newspapers on the president's fiscal record. Pretty tough words, given that LBJ conducted both a war in Vietnam and a War on Poverty simultaneously, racking up huge gains in spending over his term and a half in office.

    The McClatchy piece says discretionary spending under Bush has risen an inflation-adjusted 5.3% in his first six years, outstripping the 4.6% under Johnson � and way above President Reagan's meager 1.9%. By "almost any yardstick," the article continues, Bush "generally exceeds the spending of his predecessors."

    "Any yardstick," that is, except the most important of all � spending as a share of GDP. On this, Bush is actually lower than most of his predecessors.
    What do we mean by "discretionary spending"? The definition used by the federal government (via C-SPAN) is the spending which is annually appropriated by Congress and signed into law. Entitlement programs are out as are any other mandatory spending; combined that's about half the federal budget. But discretionary spending includes all defense spending (including supplemental spending for the war on terror.) It also includes the money spent on cleanup for Katrina, which puts a different spin on the word "discretionary". So which do you want to look at? Nondefense discretionary? Nondefense, non-disaster relief discretionary? And if that's the yardstick, why wouldn't policymakers simply choose to move more and more spending into mandatory programs so that they escape such scrutiny?
Now there's no way around Bush as someone who has spent a great deal of money and been at best a 'C' student as an economic policymaker. But looking at this chart from Heritage indicates to me that the rise in nondefense discretionary spending isn't a recent phenomenon. One can be rightly disappointed that nothing's been done to stop the rise. But when people want to talk about spending, be sure you keep it real, and keep the context in mind.