Vendor switching and the flattening Phillips Curve

Posted by

A guest post on Econbrowser based on this paper, reports the latest results from price micro-data.

It finds that CPI indices overstate inflation at the purchaser level, as households switch to cheaper vendors of the same (or similar) goods in down-turns. This makes sense, as the demand for paid labour declines, the opportunity cost of shopping time declines – so we invest more in searching for the best price.


If one re-weights price indices to reflect this shopping around, one finds that actual prices paid decline to a greater extent in downturns. The above chart shows that the difference between the average price paid and the effective price increases as the unemployment rate rises.

A related finding is that sales are pro-cyclical (more of them when the economy is good) — possibly because as shoppers get more price sensitive, the returns to sales by expensive stores (who have most of the sales) are lower as they attract fewer customers.

Due to the decline in sales when there are large demand shocks the phillips curve will appear to flatten — with observed with inflation falling by less than you would expect given the increase in the unemployment rate.


  1. Good to see you looking at interesting blog sites. This is a good one.

    You are too young but in the early 90s Bernie Fraser determined the RBA should have an inflation target.

    This coincidentally was the period when inflation targeting by Central Banks was in vogue.
    The conclusion here is a variation on a theme brought up then that the CPI overstates inflation.

    as I recall this was an exhilarating period intellectually. Two of the main people on deciding what sort of target the RBA would target were Glenn Stevens and David Gruen.

    I wonder whatever happened to them.

    ( David Gruen has a PhD in both Physics and Economics.)

Comments are closed.