By almost any measure, the American economy entered 2009 in need of a rebuild. One area that sorely stands out among those in need of overhaul that hasn't gotten much attention: the nation's economic statistics. It's understandable that data collection and analysis becomes an afterthought in times of crisis. But policy makers and business leaders need good numbers to ensure good outcomes and to make sure down cycles don't become worst cycles. So eliminating the guessing and complaining about the jobs numbers, the CPI and the PPI, the construction numbers, and all of the other data would go a long way to getting to the right solutions for the problems the economy faces. You can't plan well with bum numbers. GIGO - Garbage In Garbage Out - still applies.
Right now government data gathering, analysis, and deployment is scattered among various cabinet departments and agencies, and is often hamstrung by constituent-group biases and political agendas.
Take the much maligned inflation numbers. "There has been an upward bias in our measures of inflation," says Timothy A. Canova, associate dean and professor of international economic law at Chapman University Law School. "The U.S. uses the so-called Laspeyres index, which shows what it would have cost to buy a certain market basket of goods and services in an earlier year. But consumers don't continue purchasing the earlier basket but rather will substitute things that decline in price or go up more slowly for those products that are rising faster in price," Canova points out.
And the country's unemployment statistics enjoy a downward pitch, many economists complain. "I've long argued that the official unemployment rate seriously underestimates the amount of actual unemployment for both labor and capital," says Canova. "For a time, there was a measure of unemployment, the U-7 rate, that included 'discouraged workers'-those who would like to work but have simply given up looking, and part-time workers - those who work even one hour per week but who would like to work full-time and are unable to find full-time jobs." U-7 numbers were politically unpalatable, and the Bureau of Labor Statistics discontinued U-7 calculation and reporting in January 1994.
The history of capacity utilization is also instructive as to how data can be politicized, says Canova. "At least as early as 1968, the Federal Reserve began revising its capacity utilization series to exclude supposedly antiquated capital and thereby increased the recorded capacity utilization rate." The change was made at a time when a "decline in capacity utilization indicated considerable slack in the economy and a dramatic slowdown in labor productivity, as early as 1966." Canova argues that the Fed's approach "may have its own statistical biases that have at times reinforced arguments for tighter monetary policy."
Equally problematic is the very structure of the data-mining engine. "We don't do a good enough job running and gathering data," states David Wyss, chief economist at Standard & Poor's. "We're still operating on a structure established 50 years ago. The government focuses far more on data for agriculture than on the service sector. There's a vested interest in those databases."
The paltry amount of federal money spent on the U.S. statistical picture is divvied up along those 50-year-old lines, too. "Fifty percent of the spending goes to agriculture, and most of the rest goes to measuring manufacturing," Wyss notes. "It's easier to count tons of steel than counting doctor's visits."
Even more troubling: "Input measures are not a priority, and timeliness is not a priority," according to Wyss. "There are real problems with quality and the number of sources. You do what you can do, given the budgetary issues."
How can the sorry state of U.S. statistics be straightened out? The solutions are far simpler than trying to unravel a single mortgage-backed CDO. Create an independent Government Data Office, a joint Congressional-Executive agency perhaps, and fork over all government data series to its many economists and database managers. Implement state-of-the-art IT, and make sure it always stays that way. Put people in the field, and involve industries and scholars proactively.
The expense would hardly show up on the budgetary radar next to even relatively small federal outlays, never mind TARP. In the shadow of trillions of dollars of rescues, backstops, and stimuli, robust data gathering, analysis and sharing is even more essential. Otherwise, how do policy makers really know whether the help is working?
