The Wage Curve.

Ewing, Bradley T.

Few books have received as much attention as The Wage Curve. But then, few books have as much to offer as this one. What does this book offer? Blanchflower and Oswald might say that it merely documents the existence of an empirical "law" of economics. The opening quote to chapter 5 by Ronald Coase more appropriately sums up the books addition to economics: "Inspiration is most likely to come through the stimulus provided by the patterns, puzzles and anomalies revealed by the systematic gathering of data, particularly when the prime need is to break our existing habits of thought" [p. 209]. The Wage Curve lays the groundwork for new ideas.

The link between macroeconomics and labor economics is fundamental to the study of economics. However, traditional treatment of the subject has viewed the labor market as competitive. Wages and unemployment are positively related, stemming from a Harris-Todaro theory of compensating differentials. The macroeconomics literature is full of stylized facts, relationships that are found to hold so much of the time that economists accept them as fact. Stylized facts are not found as readily in the labor literature. Through the use of microeconomic data Blanchflower and Oswald document the existence of a negative relationship between local unemployment and wages. Labor markets do not work in the classic textbook manner. Wages are high in places with low unemployment and low in regions with high unemployment. The pattern appears to hold by industry sector, as well. A new stylized fact has emerged.

The first three chapters of the book provide the reader with an idea of what Blanchflower and Oswald set out to do and why. There is a thorough discussion of previous research on wages and unemployment across space. Given the results to date, one can not help but conclude that the evidence "is not reconciled with the textbook model of the labor market" [p. 36]. A chapter is devoted to theoretical issues and may be considered lengthy to most labor and regional economists. Basically, Blanchflower and Oswald describe and develop models that predict wages to be negatively related to local rates of unemployment. Contract theory, efficiency wage models, and bargaining models are all discussed in a regional setting. The authors are careful to point out that they do not totally disregard the theory of compensating differentials.

Evidence of a wage curve for the United States is presented in chapter 4. At this point the book becomes really exciting. A particularly interesting finding is that both regional unemployment and industry unemployment affect wages less in union jobs in the private sector. Further, it is found that in the public sector, both industry and regional unemployment have insignificant effects on the wage regardless of union status. Such an important finding is not given any real explanation, and the reader is left wondering why. Most of the results are found using Current Population Survey data. It is somewhat puzzling that Blanchflower and Oswald begin their study of the wage curve using a rather obscure data set known as the U.S. General Social Survey. Several problems are readily apparent and well noted by the authors. First, the GSS only contains information on nine regions. Certainly not enough to convince the hard line orthodoxy of a new empirical law. Second, the income variable is "grouped" and open-ended, a problem often dealt with in the literature by using a maximum likelihood method of estimation. In any event, the GSS results provide support for a wage curve, but their significance is only marginally convincing.

Further tests of a U.S. wage curve are presented in chapter 5 where three important issues are considered. The first issue concerns the interpretation of the pattern found in the data. The authors satisfactorily provide evidence that the wage curve is not a labor supply curve. The log of the unemployment rate does not appear to be a mis-measured labor force participation variable. Further, including profitability measures does not invalidate the wage curve finding. Perhaps most importantly, instrument variable estimations support the notion that simultaneity between wages and unemployment does not affect the wage curve estimates.

A professor I had in graduate school once told me that a macroeconomic relationship that is true in the United States ought to hold in other countries. Blanchflower and Oswald must be from the same school of thought. Chapters 6-8 describe their findings from multi-country analysis of the wage curve. The focus of these chapters is to provide evidence of a "deep economic phenomenon." The patterns uncovered are remarkably similar to that found in the United States. The empirical law of a wage curve stands up to the considerable differences in economic activity, cultures and institutions, and labor markets found in various countries.

There is an absence of diagnostic tests of the estimating equations common in the labor economics literature. Little, if any, discussion on the estimator properties and error processes is presented in the book. Though the brute force method is convincing enough in this case, statistical tests that provide support for the estimation procedures are lacking. This is particularly surprising given Blanchflower and Oswald's use of such modern estimation techniques as random effects models.

Labor economists can now point to a collection of empirical evidence that finds local wages are inversely related to the local unemployment level. A result that many who argue non-competitive models of the labor market more aptly characterize the real world find easy to believe, but others are often quick to discard. The data are consistent with both bargaining models and efficiency wage theory. The book does not go far enough in distinguishing which theory is driving the result. The empirical regularity of a wage curve may lead to a new macroeconomic theory where non-competitive models of the labor market play a dominant role. Thank you Professors Blanchflower and Oswald for providing this useful resource and for providing the inspiration for further research into what is behind the stylized fact that wages and local and industry unemployment are negatively related.

Bradley T. Ewing Georgia Southern University