When Statistical Measures Fail to Capture Reality
in Data Analysis | Economy | Employment | Inflation | Psychology/Sentiment
The Sunday New York Times had an interesting story by Dan Gross about one of our favorite themes: What happens When Statistical measures fail to comport with experienced reality?
This phenomena is the result of how the economy got to where it is today: Post crash, the massive government stimulus created an artificial recovery. The details of the stimuli -- ultra low rates leading to a real estate boom, tax cuts that primarily benefited those in the highest tax brackets -- are why the aggregates present a misleading picture.
The typical measure was never designed to capture the details of such a bifurcated economy. Perhaps these models are creatures of an era when wealth distribution was far less concentrated. They seem to be unable to keep up with the present shift, and the downsizing of the middle class.
"This strange and unlikely combination — strong and healthy aggregate macroeconomic indicators and a grumpy populace — has been a source of befuddlement to the administration and its allies. It's not unreasonable to assume that Mr. Snow is being replaced as Treasury secretary in part because he couldn't make Americans appreciate just how well the economy is performing. And it's possible to detect among Bush partisans an element of frustration at the public for what they see as its failure to do so. In Iowa last month, Rudolph W. Giuliani bluntly dismissed concerns about the economy and higher gas prices by saying, "I don't know what we're all so upset about."
Gas prices and the Iraq war have surely contributed to this disconnect. But a lesser-known factor is also at work: the misleading aggregates.
Aggregates — big-picture figures like the unemployment rate, productivity and growth in the gross domestic product — are highly useful to economists. But to most people, they're abstractions. You can't use a low unemployment rate to pay a mortgage.
As a result, large aggregates "are something that people may hear about in the news, but don't have a direct impact on how people feel," said Lynn Franco, director of the Consumer Research Survey at the Conference Board." (emphasis added)
How can this be? Low unemployment (NILF), Low Inflation (Ha!), strong GDP (carried over from Q4). It turns out there is a simple explanation -- the data is "simply misleading:"
"Aside from being abstract, many of the most popular aggregates are simply misleading. Dean Baker, a director of the Center for Economic and Policy Research in Washington, puts the Consumer Price Index — the main gauge of inflation — at the top of the list.
"It has no direct relationship to what people perceive as inflation," he said. Mr. Baker notes that the index doesn't take account of rapidly rising co-payments and higher insurance deductibles when it calculates health and medical costs. And to gauge inflation in housing, the index approximates a measure of rent instead of looking at home purchase prices.
"We've had a huge run-up in the price of housing, and that doesn't show up in the C.P.I.," he said. So while the index shows that inflation is elevated but still under control — up 3.5 percent from April 2005 to April 2006 — many Americans find themselves paying sharply higher prices for essential goods and services."
Remember the concept of substitution: If beef prices rise, but chicken doesn't, BLS allows a substitution in their basket of goods -- therefore showing no price gains. But the shopper in the Supermarket says, "Damn! These meat prices keep going higher!" -- hence the disconnect.
That's only part of the explanation as to why we have no inflation (ex inflation). Real income has been negative, and people feel that. Most people don't care about the BLS data, they are concenred with how much money they have at the end of the month after they pay their bills:
"In addition, aggregates generally are averages, which are of declining utility in an economy characterized by greater inequality of income and assets. In an interview with The Wall Street Journal in March, Mr. Snow took pains to point out that there had been substantial gains in per-capita income (8.2 percent, after inflation) and net worth (24 percent, before inflation) from the beginning of 2001 to the end of 2005.
The data he cited were averages, or means, and that can be misleading. "The average wage is a useful indicator if you want to know what's happening to the tax base, but it might not tell you what's going on for the individual worker," said Alan B. Krueger, an economics professor at Princeton and a former chief economist at the Labor Department."
As every student of statistics knows -- and the chart below reveals -- there is a big difference between median and average. Average is skewed by outliers (Bill Gates walks into a bar . . .) whereas Median is not:
>http://bigpicture.typepad.com/comments/2006/06/when_statistica.html |
