How to lie with statistics
“How to Lie With Statistics” by Darrell Huff is a venerable book on the many ways we as end-users of statistics can be misled by “the numbers.” As participants in the capital markets we are bombarded with various statistics on a daily basis. From various governmental sources we receive data on the state of the economy. From companies we receive earnings report that purport to tell us how the firm is doing.
This post is not about the deliberate manipulation of statistics. One can surely think of any number of examples of misinformation and fraud that have occurred in government, corporate America and in the fund world that have come about through a willful manipulation of the numbers. Rather this post is about when there are genuine disagreements on how to accurately measure economic reality.
For example, Holman Jenkins at WSJ.com notes how “mark to market” accounting is at the heart of the many writedowns occurring at major financial companies. Financial accounting is, at best, an agglomeration of various financial estimates. A large company’s income statement and balance are the sum total of any number of accounting assumptions. Reasonable accountants can, and do, differ on the application of the many accounting rules.
That being said, although Wall Street spends its time splitting hairs over pennies in each quarterly earnings announcement, it misses the bigger picture. As Jenkins writes:
Overstating the importance of accounting rules is, indeed, the essential error that leads to excessive twiddling with accounting rules. Whether a company values its assets at historic cost or market value or a value derived by some other formula, investors still have to make their own forecasts and judgments. A thermometer is equally useful whether it says water freezes at 0 degrees or 32 degrees — though it still doesn’t tell what the temperature will be next week.
Another realm in which statistics play a large role is in measuring the state of the economy. Global financial markets can hinge on the fractional changes in measures like GDP, inflation, and payrolls. David Leonhardt at NYTimes.com has an article on how changes in the nation’s economy have affected jobless statistics. The way Americans work makes the incumbent statistical framework less useful. So much so that:
Yet there is no doubt that the unemployment rate is a less telling measure than it once was. It’s simply no longer the best barometer of the country’s economic health. A truer picture can be found elsewhere, by looking at compensation growth, for instance, or to changes in the percentage of the employed.
Another measure at the forefront of policy makers and investors alike is inflation. One would that looking at a market-derived measure of inflation expectations would be the gold standard estimate. However, even there reasonable analysts can differ.
The spread in yield between nominal and real Treasury securities is indicating higher future inflation. Greg Ip at Real Time Economics notes how this spread has been rising since the beginning of 2008. While the Federal Reserve has good reason to play down the implications of this rise, there are some good reasons why even this number has some play in it:
Mr. Mishkin says it’s because the breakeven doesn’t represent just expected inflation, but compensation for other factors including uncertainty about inflation. By this reckoning, if investors have the same forecast for inflation but less confidence in that forecast, they will pay extra (i.e. accept a lower TIPS yield) for insurance against inflation. This is true even if there’s an equal chance of it coming in lower as higher than expected
Another example more market-related springs to mind. A cadre of bloggers, way smarter than us, have been exploring the inner workings of the VIX, i.e. the CBOE Volatility Index. The VIX is a widely quoted market “fear indicator” that for all intents and purposes is a black box to the vast majority of investors. It may be enough for most of us to know that when it goes up, options volatility is higher, and vice versa. Its inner workings are understood only by those well-versed in options theory. The VIX is only a sampling of implied options volatility and is at best a mere proxy for investor fear.
With April 15th quickly approaching, we can guarantee that there will be newspaper and magazine articles that send a family’s tax information to a handful of tax preparers. Inevitably they will get back a range of estimates for taxes the family should pay. Absent fraud, and/or error, the differences will come down to the differing application of various tax rules and regulations. In short, the reality of personal taxation is ambiguous.
Whether it is corporate earnings, jobless claims, the VIX or a family’s taxes there is an inherent ambiguity in application of statistics. Unfortunately there is no simply way to wish this problem away. Indeed, one could argue that this problem will only get worse with time.
In today’s increasingly global and dynamic economy it is going to become ever more difficult to precisely measure the changing state of the economy and corporate performance. Much ink, and pixels, will be spilled parsing future data releases. Smart and well-intentioned analysts can, and will, disagree about them. This is all well and good, and healthy to boot. However for investors the future is what matters. Focusing on the big picture and those measures that are unambiguous and well-understood will help in sifting through an increasingly murky statistical world.
Filed under: Economics | 2 Comments