Does the consumer price index overstate inflation?

A US commission said it did, by more than a percentage point a year, and the question has not been fully settled since. In 1995 the US Senate appointed the Boskin Commission to study possible bias in the Consumer Price Index. Its 1996 report concluded that the CPI overstated inflation by about 1.1 percentage points per year in 1996 and about 1.3 percentage points before that.

Why would a basket of goods lean high? The commission pointed to four sources. Substitution bias comes from a fixed basket: when relative prices change, shoppers swap toward goods that have become relatively cheaper, but a longstanding basket keeps pricing the expensive version. If product A is bought by most consumers and a similar product B goes on sale, people will naturally buy what is cheaper. Outlet substitution bias is the same idea for shops: shifts to lower-price outlets can be missed. Quality change bias arises when improvements, such as better energy efficiency or less need for repair, are measured inaccurately or not at all. New product bias arises when new products reach the basket late.

The stakes were practical. CPI inflation sets the annual increases in Social Security and other programs, so, by the report's reasoning, the federal budget had risen by more than it should have.

Yet the fixes are themselves contested. Some commentators, including Austrian economists, feared that quality adjustments could be used to mask the true inflation rate. Others found that unadjusted methods can misstate inflation too.