Industrial and Labor Relations Review |
January 2001, Vol. 54, No. 2 Review of Why Wages Don't Fall
during a Recession. By Truman F. Bewley. Wage rigidity is something to be documented and something to be explained. By using the unusual method of listening to business people, Truman Bewley seeks to do both. The result, Why Wages Don't Fall during a Recession, spans most areas of labor economics and the fluctuations side of macroeconomics. The setting is the recession of the early 1990s. Bewley interviewed more than 300 business people in and around Connecticut. Rather than administer a formal survey instrument, Bewley simply introduced business people to his areas of interest and let them talk. He listened. The book contains an extensive set of quotations from business managers, union leaders, and placement people describing what they do and why they do it. The title of the book reflects a question of enduring interest: why don't firms cut wages to avoid layoffs? Chapters 1-3 set the context in terms of economic issues, interview and sampling methods, and the economic environment in and around Connecticut in the aftermath of the 1990-91 recession. The book's core contains one chapter for each of sixteen issues (for example, morale, internal pay structure, resistance to pay reductions, layoffs).Three closing chapters bring the interview evidence to bear on existing theories (Chapter 20), sketch a new model of workplace morale (Chapter 21), and offer suggestions for further interview-style work (Chapter 22). Each core chapter is organized into a series of sections and an appendix containing extensive links to related literature. For instance, the ten sections in the chapter on layoffs (Chapter 13) include "Why Layoffs?", "How Do Companies Choose Whom to Lay Off?", and "Can Workers Choose between Less Pay and Layoff?" Opening each section is a summary of interview findings (including a table of counts), followed by a series of quotations. Consider, for example, the section "Reasons for Laying off Workers Rather Than Cutting Pay." Here Bewley reports that asking "directly about the choice between layoffs and pay cuts risked alienating managers," and that a common reaction to this question was "puzzlement" (p. 181). Some responses focused on lack of work. A human resource manager reasoned, What do pay cuts have to do with layoffs? A layoff is used when you don't have sufficient work for certain skills. What would you do with the extra help? It would be unfair to them to keep them around....How could you ask everyone to take a pay cut in order to keep some idle people around? (p. 185) Other responses stressed that "layoffs do less damage to morale and productivity than do pay cuts" (p. 183). The owner of a small manufacturing company explained, "A wage cut would give rise to morale problems. The employees would have a chip on their shoulders and would lose the fire in their bellies" (p. 175). Apparently, a key advantage of layoffs is that the demoralized are ushered from the workplace. Bewley's interview evidence is a valuable supplement to the wealth of evidence we have from econometric sources. Some will be skeptical of interview evidence. But many of us take every opportunity to ask taxi drivers, waiters, hair cutters, construction workers, corporate executives, and others about their jobs, and the responses are sometimes insightful. The exhaustive breadth of the issues covered by Bewley's interviews, as well as their sheer number, magnify the insights to reveal important themes. (See section 21.1 for a lucid summary.) Bewley's presentation of the interview responses in relatively raw form (that is, series of lengthy quotations) contributes to the book's success. But my enthusiasm for the book is qualified by reservations regarding the survey method and the relevance of the interview evidence for modeling the labor market. An obvious disadvantage of Bewley's method is the short distance between the observer and the observation. The only source we have for the insightful quotations is Bewley's interview notes, which he wrote up carefully after each interview, "filling in missing information from memory" (p. 26) .Even one's notes filter actual responses, and this leaves too much leverage to the researcher. Archived tape recordings, which could provide an important safeguard, do not exist in this case. Yet I am not alarmed by this potential pitfall in Bewley's book, because I see no evidence of ax-grinding. The area of wage rigidity is new to Bewley. Indeed, the 48-page bibliography includes not a single reference to any of Bewley's numerous publications. Another problem is the question of the credibility of interviewees' self-reports of "why they do what they do." Business people know what they do, and they can report what guides the decisions. But such explanations can be superficial. In business, it pays to make the right decision; it does not pay to know why that decision is the right one. Perhaps managers understand some behavioral relationships without knowing the deeper structural relationships. In many contexts (for instance, in the absence of innovation) , they would gain little through such knowledge. They can get by with correlations. (Recall the common business slogan, "Analysis breeds paralysis.") So why should economists, in seeking to identify structural relationships, rely on the reasons offered by business people? In Chapter 20, Bewley uses his interview evidence to sort through competing models of the labor market and of cyclical fluctuations. He finds fault with virtually every existing model because of the "lack of realism of their basic assumptions" (p. 423). Here my enthusiasm fades, because Bewley applies faulty scientific method to reject theories. He argues that modeling assumptions must be realistic, accurate, or even true. Like Bewley, I firmly believe that assumptions are important; indeed, establishing the validity of a scientific explanation requires checking its essential assumptions as well as testing its key implications. But the realism of inessential, auxiliary, or simplifying assumptions is irrelevant. Models are always abstractions, and in evaluating modeling assumptions the standard is always determined by the purpose of the model: what is to be explained determines what is essential and what is not. In sorting through the models, one must not drop the context of the model's purpose -- and that is precisely what Bewley does. Since the strength of equilibrium business cycle models does not lie in the realism of their assumptions, Bewley's rejection of this class of models is not surprising. But nowhere does he note the key patterns these models hope to explain. Furthermore, to determine whether nominal wage rigidity is instrumental in explaining fluctuations, one needs to know what patterns would be implied without rigidities. If all the key features of fluctuations could be explained without resorting to wage rigidity, blaming fluctuations on wage rigidity would be at best premature. If wage rigidity exists, perhaps it is simply a bell or whistle. Another example is Bewley's application of interview evidence to implicit contract models. The business people, he says, were confused by the idea of employers providing insurance (p. 411). Since they do not see insurance premiums and claims, Bewley rejects implicit contracts as being unrealistic. But the-interviewees do see risk sharing, which is the central theme of implicit contract models. Business people do not call it insurance, but this is no reason to reject a class of models in which risk sharing illuminates key elements of the wage structure and employment fluctuations. The interview evidence of wage rigidity is important, especially since the evidence for downward nominal wage rigidity is underwhelming in panel data. Bewley, however, rejects most models of wage rigidity on the grounds that they assume the wrong source for the rigidity (for example, private information rather than morale) .Although I find this analysis valuable, I remain troubled that Bewley continues to drop context: he ignores each model's purpose when evaluating its assumptions. After rejecting nearly every existing model of employment and wages, Bewley sketches his own model of mood and morale in Chapter 21. The opening section, "Morale Theory," does not contain a theory of morale, although it is a superb summary of the emerging themes from the interview evidence. Few will find anything to disagree with, and fewer will find reason to change the way they model employment. subsequent sections contain a formal model of morale, emphasizing an inner conflict between conscious and unconscious selves. The model is a sequential game, with the unconscious mind moving first in choosing mood, anticipating how the conscious mind's decision will react to mood. Given the weight Bewley places on realism of assumptions, I find it odd that he would allow the unconscious mind to be so forwardlooking and strategic. My reading of Bewley's interview evidence takes me in a different direction. There is an intertemporal bargaining or hold-up problem. An employer who can commit not to behave opportunistically is at an advantage in attracting and retaining workers. Commitment takes the form of salary scales, pay equity, work rules, layoff rules, and so on. If the employer breaks the rules, violating an implicit contract, workers feel cheated and get angry; morale takes a dive, and productivity suffers. Workers' emotional reactions then discipline the employer, encouraging the employer to follow the rules. We might wonder why workers respond with anger ( as they surely do) ; here again we can frame an explanation in terms of commitment and the survival value of emotions (see, for example, Robert Frank, Passions within Reason: The Strategic Role of the Emotions [New York: W.W. Norton, 1988]). I am also struck by managers' fear of the legal consequences of decisions regarding employment, layoffs, firing, severance pay, checking references, and so on. Fear of frivolous lawsuits seems to dominate the current employment relationship in the United States. The book succeeds in its core mission of illuminating business people's impressions about wage rigidity and related issues -- what they do and why they do it. Although I disagree with Bewley's view of the role of realism of a model's assumptions and the way he applies that view in sorting through models, I find much to admire in the book's core -- especially the many illuminating quotations from business managers, union leaders, and placement professionals. I remain skeptical of the power of interview evidence to identify motives, but Bewley has certainly projected a panoramic view of what business people do in managing the workplace and what is in their minds when they do it. Even if you tend to distrust the impressions of business people, I am sure you will find more than a few convincing patterns in this book that deserve attention in economic models. Kenneth J. McLaughlin |