The Retail Revolution and Food-Price Mismeasurement

In December 1995, the U.S. Senate Finance Committee’s Commission to Review the Consumer Price Index issued its final report. This report stated that the U.S. Consumer Price Index (CPI) was an upwardly biased measure of the cost of living that most likely exaggerated inflation by 1.1 percentage points a year. Exaggerating inflation means that we underestimate the purchasing power of our money and thus reduce gains in output when we measure them in inflation-adjusted dollars. The report attempted primarily to estimate the current and future bias of the CPI; it did not discuss the historical bias nor whether the bias had increased. This article presents evidence that the upward bias in measures of U.S. inflation worsened in the late 1970s. A rising bias would support the argument that the slowdown in growth of U.S. inflation-adjusted output and labor productivity reported since the mid-1970s is an artifact of mismeasurement.¹

Nobel laureate economist Robert Solow has said of this slowdown, “We see computers everywhere but in the statistics.” This so-called Solow paradox draws its power from the observation that not only is there no apparent pickup in productivity growth from the widespread adoption of personal computers that began in the late 1970s, but, instead, our official statistics report a drastic slowdown. I argue here that one of the impacts of the new electronic technology has been a retail revolution that has made price measurement more difficult. Our failure to see faster growth from computerization is thus a side effect of the confrontation between an outmoded statistical system and a rapidly changing economy.

This article appeared in the May/June 1998 edition of Business Review.