October 12, 2005
Statisticians like to say that if you torture the data long enough it'll confess, which is a cute way of saying that you can prove almost anything if you play with the numbers. But data massaging can be a two-way street. If the creative mind can manipulate data, there's also the possibility that the data is less than pure before the analyst starts to work on it.
The issue of misleading numbers comes up anew with the recent government revisions to gross domestic product data for the past three years through this year's first quarter. For example, growth of the GDP price index over the three-year period was raised from two percent to 2.3 percent, reports the St. Louis Fed in its October issue of Monetary Trends. Another adjustment includes the “core” personal consumption
expenditures (PCE) price index, which is among the Fed's preferred metrics for assessing inflation.
"Although the core PCE price index was revised upward by only about 0.25
percentage points to about 1.75 percent over the three-year period, what is perhaps worrisome is that the largest revisions to the core PCE inflation rate have occurred since mid-2003," writes Kevin Kliesen, an associate economist with the St. Louis Fed, in Monetary Trends. "This is the period when monetary policy was characterized as accommodative and economic growth was generally robust. From 2003:Q2 to 2005:Q1, the annualized growth rate of the core PCE price index was revised up 0.5 percentage points to 2 percent."
Does this mean that monetary policy was more accommodative than originally billed? And if so, does that mean that the monetary tightening now under way should be more aggressive, run longer, or both in order to correct for past data sins?
It's a question worth pondering once you realize that the real (inflation-adjusted) Fed funds rate has been lower than government data previously suggested. The disconnect between what was reported and what now appears to have been unfolding was at its worst in late 2003 through early 2004. The government data dispensed at the time showed a real Fed funds rate rising, moving from negative to positive. In other words, it appeared that the central bank was substantially tightening the monetary strings when in fact, based on the revised data, the exact opposite was in progress, i.e., the real Fed funds rate became more negative (or more "accommodative") for a time.
Kliesen observes that "current estimates show that a negative real Federal funds rate persisted until the December 2004 meeting, probably a more stimulative stance than the FOMC intended. While the policy implications of this particular data revision
may not be large, it nevertheless reinforces the difficulties that FOMC policymakers confront when implementing policy in real-time." And, we might add, it's no picnic for bond investors either.
Calculating what a debt security worth is tough enough in the best of circumstances, thanks to the ongoing mystery otherwise known as the future. As it turns out, the past isn't necessarily crystal clear either. Knowing that, who'll step right up and take a guess at whether the current yield on the 10-year Treasury (roughly 4.44% at the close of today's session) is overvalued, undervalued, or fairly value?
Posted by jp at October 12, 2005 4:52 PM
Right you are...noted and fixed. Thanks.
Posted by: JP at October 13, 2005 1:43 PM
correction on your 10-yr usd bond assertion - the closing yield on oct12/05 was more in the neighbourhood of 4.44
Posted by: franko at October 13, 2005 1:08 PM