Why Wages Don't Fall During a Recession

by Truman F. Bewley
(Harvard, 576 pp., $55)

What do workers want? Most people give the same answer: more money. The answer is not entirely false, but it certainly is not right. If employers and government officials act on the basis of that answer, they will make serious mistakes.

To understand this point, it is necessary to back up a bit and to say something about an increasingly intense debate. Economics is known as "the dismal science," in part because many economists work with a dismal picture of human beings. According to a standard account, people combine a high degree of selfishness with an impressive ability to calculate and to promote their interests. They are "rational, self-interested profit-maximizers." Many economists see people's choices as following from that foundation--not merely investment decisions, but decisions about whether to get married, to have children, to commit crime, to discriminate, to have safe sex, to raise or cut wages, to work, to break the law, to have an abortion.

To say the least, this is a simplified picture of human beings; but economists have used it to provide remarkable insights into social problems. We now know, for example, that rent-control legislation will decrease the stock of available housing (a special problem for poor people); that the environmental regulation of new cars can increase pollution, by leading people to keep old, dirty cars on the road; that laws forbidding discrimination on the basis of disability can actually hurt disabled people, by making employers wary of hiring disabled people for fear of a costly lawsuit.

Yet the simplified economic picture of human behavior sometimes goes wrong; and in the last twenty-five years many people have been trying to infuse economics with a more accurate understanding of human beings. Gary Becker has emphasized the role of altruism and social influences on individual behavior. Other economists have worked to create a new field, known as "behavioral economics," which seeks to imbue economics with a more vivid and complicated sense of how people think and what they want. Behavioral economists draw particular inspiration from two cognitive psychologists, Daniel Kahneman and the late Amos Tversky. Through a series of careful experiments, Kahneman and Tversky have tried to provide a kind of "map" of human judgment and decision-making.

Kahneman and Tversky show that we often rely on mental short-cuts, or "heuristics," to help with difficult tasks. Perhaps the most important of these ordinary calculations is the "availability heuristic," by which we judge the probability of an event by asking whether an incident of its occurrence readily comes to mind. Hence, for example, some people think that AIDS is not a problem, while others think that it is a near-epidemic. Kahneman and Tversky also show that people usually hate losses from their current situation--in fact, they dislike losses far more than they like equivalent gains. If, like most people, you do not prefer to sell stocks for a loss, or you get upset whenever gasoline prices rise, you will know what Kahneman and Tversky have in mind. Behavioral economists refer to this phenomenon--the special antipathy to losses--as "loss aversion." Robert Shiller's Irrational Exuberance, his surprise best-seller about the stock market, draws on behavioral economics to question the rationality of stockholders' decisions, and ultimately to predict a large decline in market values.

Behavioral economists also emphasize that people care about things other than money or material self-interest. They are especially interested in status and decency--in being fair and, even more, in being treated fairly. As the economist Robert Frank has emphasized, people often care most not about how much they have in the abstract, but whether they have less or more than others. Many people would prefer earning $70,000 per year in an organization where most people earn $60,000 to earning $80,000 in an organization where most people earn $90,000. And, as Matthew Rabin and Richard Thaler have shown, individuals are sometimes willing to sacrifice their economic self-interest in order to be fair and to punish individuals who are acting unfairly. Most people tip, even in out-of-town restaurants. And many people will give up some money in order to harm people who have been unfair to them--not just in an acrimonious divorce, but in daily life.

With an understanding of the fundamental importance of fairness, we can obtain a better sense of some otherwise surprising behavior--why, for example, many companies are reluctant to raise prices even when they can; why some parties to a lawsuit do not settle when settlement clearly seems in their mutual interest; why people often comply with laws that are not enforced; why strangers cooperate for the public good; and why the laws of many states ban "price-gouging," or raising prices in times of emergency. The economist George Akerlof has even speculated that the workplace, supposedly the prototypical realm of purely self-interested behavior, contains a "gift exchange." According to Akerlof, many employers give employees a little more money than the market requires, and employees reciprocate by working a little harder than they otherwise would. The example is supported by experimental evidence that "reciprocal altruism"--generous behavior from one person, producing similar behavior from another--is a pervasive feature of social and economic life.

Behavioral economics is having a large and growing impact on many fields--not only economics, but also political science, sociology, and law. Indeed, it appears to be creating a kind of revolution. But many of its findings come from simple and somewhat artificial experiments with small groups of subjects. It is reasonable to ask how and whether the central findings map onto the real world. We could know a lot more about whether and when people sacrifice money for the sake of fairness, or show a special antipathy to losses, or prefer status--and hence favorable comparisons with others--over absolute gains.

Truman Bewley is one of the leading mathematical economists of his generation. For a long time he has been interested in a central conundrum for economic theory: that workers' wages do not fall during recessions. This is a conundrum because, on the conventional view, any recession should lead employers to reduce the costs of labor, first by cutting wages, and second by laying people off. To be sure, employers do lay people off. But they usually do not cut wages. Writing in the earnest, dry, inquisitive, somewhat personal tone that runs throughout his book, Bewley explains: "I have long wondered about wage rigidity. I was puzzled that firms and workers do not agree to cut wages and salaries during recessions when sales decline.... I wasted years inventing theories describing impediments to pay cuts."

In this book Bewley mostly avoids math, and he does not try to "invent theories." Instead he reports the results of more than three hundred interviews with employers, labor leaders, business consultants, and unemployment counselors. The interviews focus on employers' behavior during the American recession of the early 1990s. He asked people, especially employers, what they did and why they did it. Much of his book consists of actual quotations purporting to explain employers' behavior.

Most economists do not like to base their claims on interviews, which they believe to be unreliable. People might not know what they did or why they did it. They might not tell the truth even when they know it. Bewley's explanation of his use of interviews is disarmingly simple: he "could think of no other way to answer" the most basic questions about employer behavior. His interviews produce a clear answer to the conundrum with which he started; and in the process they also provide a number of important lessons about economics and human motivation, even when money is involved.

Begin with the basic conundrum. Bewley studied 165 businesses taking action during the recession. What did they do? 115 reduced raises or bonuses; 60 froze pay; 69 laid people off; and only 31 cut the pay of some or all employees. And why were wage cuts so unusual? Because wage cuts hurt workers' morale. When workers' morale suffers, their performance suffers. When their performance suffers, business suffers.

About seventy percent of Bewley's respondents emphasized this point. Consider the account of a high-level official in an insurance company with 50,000 employees: "Morale suffers from a pay cut for psychic reasons. It says, `I am doing badly,' even if everyone gets a pay cut. It is part of American culture." An owner of a restaurant with 30 employees spoke in the same terms: "I have never cut anyone's pay. I don't believe in it in principle. . . . A pay cut would be interpreted as a punishment, even if it were done across the board." The owner of a small manufacturing company spoke similarly: "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." The judgment of the vice president of human resources of a manufacturing company with 5,500 employees: "Pay cuts are extremely demoralizing. Everybody gets used to a standard of living."

Bewley concludes that pay cuts "would probably be viewed as disloyal, unfair, and as a capricious denial of earned rewards." Hence it "is anticipation of negative employee reactions that makes employers oppose pay cutting." While economists normally assume that people are self-interested, and responsive to financial rewards and punishments, "other motivations are useful as well," often having to do with the importance of "generosity." Hence employers widely believe that "punishment should seldom be used to obtain cooperation." High "morale motivates workers to perform well even when no supervisor can check on them." In these circumstances, any effort to cut pay carries with it "an insult effect and a standard of living effect." The insult effect comes from the fact that "workers associate pay with self-worth and recognition of their value to the company." The standard of living effect means that pay cuts produce "much more damage to well-being than increases of an equal size improve it." By contrast, the "level of pay itself has little impact on morale, unless pay is so low as to be perceived as grossly unfair." Apparently what matters, for many workers, is the direction of any change in pay, far more than the actual level of pay.

All this creates an additional puzzle: don't layoffs hurt morale, too? If so, then Bewley has not explained the anomaly after all. The answer is that any adverse morale effects from layoffs are far less serious than those produced by wage cuts. In the words of the owner of a small manufacturing firm: "It is easier to lay people off than to cut pay. It's like soldiers. If the guy next to me is killed, I say, `Wasn't I lucky it wasn't me.'" A high-level official in a large company reports that "with layoffs, people take an `It'll never be me' attitude. Pay cuts affect everybody. Layoffs affect a small percentage of the employees." In the same vein, a vice president of human resources at a manufacturing company remarks that "a pay cut would have a worse effect. The effect of a layoff on the morale of survivors is temporary, provided that they are taken care of."

These, then, are Bewley's principal findings, his solution to the puzzle. But his interviews uncover a great deal more. People who are interested in markets and the proper place of regulation have often been puzzled by the role of severance benefits. In theory, such benefits might well be offered by self-interested companies seeking to attract good workers. They might also be sought by self-interested workers, knowing of the possibility of layoff. But many companies in Bewley's survey do not offer severance benefits at all. Why not?

Bewley's answer is that many workers don't care about severance pay, and employers are aware of this. Workers do not want to think about being laid off. They tend to be optimistic, thinking that the risk is low, and they focus on the short-term. And they do not want to give a bad signal to prospective employers. Thus the owner of a small dry-cleaning company explains: "We pay no severance. It would be throwing money away. You would get nothing for it. The employees have low aspirations. They all live from paycheck to paycheck." Similar sentiments come from the human resources official in a large supermarket chain: "Severance pay is not a big issue. . . . There is little demand for it from the union. People prefer not to think about layoff. The thinking is very optimistic." The personnel manager of a mid-size company reports: "No one asks about severance when they are being hired. Job applicants ask first about salary, then about health insurance, then holidays, then vacations, and then educational assistance."

Bewley also offers other notable findings. It is commonly thought that employers will not hire over-qualified applicants; but we know little about whether this phenomenon is real and widespread and if so, why it occurs. Bewley finds that employees are indeed reluctant to hire over-qualified people. He also uncovers reasons for this reluctance. Obviously there is a risk that such people will soon quit. A less obvious point is that they are likely to be miserable at work and thus suffer from, and spread, low morale. In the words of the personnel manager of a mid-sized firm: "You cannot fulfill the needs of an over-qualified person. They will be unhappy and will be a problem. They might leave or feel they know too much and not cooperate with other members of their team." A human resource official in a large brokerage and investment banking company spoke more simply: "They won't be happy with the job."

Bewley discovers intriguing and large differences between what he calls primary-sector jobs and secondary-sector jobs. The primary sector contains long-term and full-time employment. By contrast, jobs in the secondary sector are typically part-time, with high levels of turnover. Bewley finds that employers in the secondary-sector are entirely willing to cut pay. They are also happy to hire overqualified people. The reason is that in the secondary sector these practices do not have adverse effects on morale. In that sector, the "confusion caused by part-time schedules and high turnover makes it difficult for workers to get to know one another and to learn one another's past." In addition, there is "less resentment of pay inequities because jobs are seldom taken seriously as careers." Hence employers are freer to cut wages. They also need not worry about the adverse effects of over-qualified workers on general morale.

Bewley's most important claim is that employers depart from conventional economic predictions not because employers are irrational or altruistic (they are not!), but because fairness is something that employees care about, and judgments about fairness operate in accordance with predictable rules. Employees will punish unfair treatment, even at an economic cost to themselves. For their part, employers will behave fairly, partly because of their own sense of fairness, but far more because of fear of offending their employees' sense of fairness. If they want to ensure good work, employers are not going to cut wages or otherwise impose any "insult" on workers.

Bewley also shows the pervasive importance of "loss aversion" to workers' judgments, including their judgments about fairness. Workers hate pay cuts even though they may not much mind when an employer reduces raises or bonuses, or even freezes pay at existing levels. He also shows that workers do not plan against disasters; hence the failure to bargain over severance pay. At the same time, workers care a great deal about how their pay compares to others within the firm. Serious problems will develop if people think that they are doing worse than their coworkers.

Bewley's fascinating analysis does raise two serious perplexities. The first has to do with the reliability of his survey evidence. The second has to do with the implications of his account, about which he says little.

Bewley's surveys are very far from scientific. To a large extent, they are a matter of serendipity. Consider his research methods. He began by arranging "a few interviews with businesspeople" in the interest of "seeking inspiration for theoretical models of wage rigidity." Later he received help in contacting companies from the New Haven Chamber of Commerce, finding thirty-five in manufacturing and service industries, of which sixteen agreed to interviews. Other interviews came as a result of "personal connections," through which Bewley followed up on suggestions from acquaintances, and from "cold calling," where Bewley had a success rate of about forty percent.

There is an immediate difficulty here: we do not know if Bewley's sample is representative. Were the initial thirty-five companies typical of companies in general? Of those thirty-five, were the agreeable sixteen different from the disagreeable nineteen? Did the personal connections introduce further distortions? Were the forty percent who accepted a "cold call" typical of the industry, or were they outliers? We do not know. And it is reasonable to wonder whether employers who are willing to take time to submit to lengthy interviews from economics professors are especially likely to care about the appearance and the reality of fairness.

An additional problem comes from the fact that any survey, even the most representative, measures what people say, not what they do. What people say is likely to be connected to what they do; but memories are imprecise, and perhaps some employers have an incentive, in the interview setting, to give biased accounts of their practices--say, by exaggerating their tendency to be fair to employees. And from the standpoint of good social science, the problem is aggravated by the fact that Bewley's questions were not standardized. Everything came from informal discussions. His presentation consists largely of quotations, taken out of context. Many of these quotations are extremely vivid, but who knows if they are representative?

Such criticisms suggest that Bewley's somewhat impressionistic findings should be taken with salt. But what Bewley finds about employers' practices is entirely consistent with what other evidence we have of their behavior. What Bewley adds is people's own accounts of why they do what they do. Those accounts may be flawed, since people are sometimes not the best explicators of their own motivations; but at the very least, Bewley's book puts the burden of proof on those who seek to offer alternative explanations.

Bewley says very little about the implications of his findings for either businesses or government. On this score there is much room for thought. Bewley mounts a passionate challenge to conventional economic methods and accounts, and he provides a great deal of real-world support for what behavioral economists have been urging. A central point, for those in the private and public sectors alike, is that it is best to avoid, whenever possible, any impression that some action will produce a loss from the status quo. Consumers will respond far more favorably to the idea of a "credit card discount" than to the notion of a "cash surcharge"--even though this is merely a matter of labelling, not of any difference in behavior. If it is necessary to save labor costs, an employer would do better to deny bonuses or raises than to cut wages. This simple point helps to explain why sensible employers include bonuses as part of a pay package and struggle to make it appear that these are not a part of ordinary compensation. The same point also helps to explain why hotels often offer "discounts" in periods when there is little demand for rooms, rather than "price increases" in the most popular months.

Similar lessons apply to government. Suppose that officials are seeking to encourage people to take measures to improve their own health. It is much less effective to say, "if you do this, you will improve your chances of avoiding illness by ten percent," than to say, "if you don't do this, you will decrease your chances of avoiding illness by ten percent." The difference is purely verbal, but the second formulation triggers the pervasive human inclination to avoid losses, and it produces more change for exactly that reason. And citizens, like workers and consumers, care about fairness. They will punish government, possibly at their own expense, if they believe that it is treating people unfairly. People are much more likely to comply with the law, including the tax law, if they are convinced that the government is treating people equally and making no invidious distinctions. At least some of the time, law-breakers are like Bewley's workers, punishing those who exercise authority unfairly.

Bewley also provides new reasons to question the dogmatic anti-paternalism of free-market enthusiasts. In recessions, layoffs are common. But workers are often unprotected through severance pay, partly because they are focussed on the short-term, partly because they are excessively optimistic. In these circumstances, it is reasonable to ask whether there is a market failure, justifying governmental response. Perhaps government should provide workers with information about the occasional need for severance pay. Perhaps government should create incentives to ensure that workers have some protection in the event of a layoff-- supplementing unemployment insurance programs by promoting a form of private insurance via agreements between workers and employers. What we do not know is whether the problems that lead workers not to seek severance pay apply to other workplace benefits, such as health care and leave time to take care of ailing family members.

In his closing plea, Bewley objects that economics "lacks the status of a real science" and "sometimes indulges in fanciful theoretical representations of reality." He complains that many "branches of economics are not anchored in empirical knowledge." His own work fills important gaps, above all because it shows how concerns about fairness and avoidance of losses affect central decisions within the workplace. But Bewley's work is only a start. If we want to build our social and economic policies on actual behavior rather than on "fanciful theoretical representations," we will need much more work in this very promising vein.

Cass R. Sunstein is a contributing editor at The New Republic.

By Cass R. Sunstein