It should be clear to you from the text the important role the assumption of sticky wages plays in helping to explain both unemployment and the ability of fiscal stimulus in a period of unemployment to increase output and reduce that unemployment. This might lead you to wonder about the evidence of stickiness in the real labor market.

Until recently it has been hard to find the right data to answer the question of wage stickiness. After all, we cannot just look at average wages since labor is differentiated. We really want to ask for a set of workers how much wage change each of them experiences over time. A recent paper using a longitudinal survey done by the Census Bureau has done just that.1 The Census Bureau in its Survey of Income and Program Participation (SIPP) followed 39,095 people for periods of 24 to 48 months, interviewing them quarterly about their work and wages. All the workers were hourly workers, making the wage data easier to interpret than that of salaried workers. The data allowed Barattieri et al. to distinguish between wage changes achieved by changing jobs and wage changes within jobs.

The period of the survey was 1996–2000. Looking back on the data in Chapter 5 you will see this was a period of low unemployment and high growth, precisely the time we would expect to see wages going up. The researchers found, however, considerable evidence of wage stickiness, especially among those who did not change jobs. They found that the probability of a within-job wage change is between 16 and 21 percent per quarter. This translates into wages that remain unchanged for about a year, which tells us that average wages change much less than average prices.

Questions

Why do you think wages are sticky even during high growth periods?