Perhaps the most famous and controversial downsizing is that of AT&T in January 1996. 1995 had been one of AT&T's most profitable years, as the company earned in excess of $4.7 billion. On the first trading day of the new year, Tuesday, January 2^nd, in a memorandum to all employees, AT&T CEO Robert Allen announced that the firm would be firing 40,000 workers over the next three years as a "strategic response to changes in our environment." He categorized the move as one that would make "our businesses more competitive and more responsive to customers." Allen further asserted that the downsizing was "driven by changes in our marketplace, changes in our customer needs, technology and public policy." Two-thirds of the firings would take place within the management of the company. The breakdown of the layoffs was given as follows: 10,000 from "staffing," 12,000 from "operations," and 18,000 from "systems and technology." It had already been announced four months prior that AT&T would be spinning off two companies, Lucent Technologies, and NCR Corp., however, there was little sentiment that these divestitures would be accompanied by reductions in the work force.

The stock market greeted this news with glee. Within minutes of the announcement, the Dow jumped, and closed up 60 for the day, as traders bet that "profit-sweetening job cuts...would remain in vogue among large corporations." Figure IV.1. shows AT&T's return versus the S&P 500 for the five trading days before, and the five trading days following the announcement. Both the value of the S&P and of AT&T on the NYSE are normalized to 100, on the day before the announcement was made. Once AT&T announced its downsizing, its stock jumped $2.625 per share to close at $67.375 for the day. This was a gain of $4.2 billion in just one day, a gain of over 4% of AT&T's market capitalization.

The announcement fueled the flames that downsizing was beginning to generate in America. Many felt that "Main Street" was suffering at the hands of "Wall Street;" in other words, corporations were abusing labor, and precluding it from participating in the economic recovery. Labor market tensions heightened: if jobs were not safe at a bellwether firm such as AT&T, were they safe anywhere? A New York Times story on downsizing proclaimed that "job apprehension has intruded everywhere, diluting self-worth, splintering families, fragmenting communities, altering the chemistry of workplaces...and rubbing salt on the very soul of the country." Newsweek ran a cover story on CEOs of major corporations, calling them "corporate killers" and "hit men" that were earning record salaries (like Robert Allen's $3.362 million-$5.489 million with benefits-announced just days after the downsizings) while putting thousands out of work; Allen's face was plastered on the cover.

AT&T's announced downsizing "seemed like the crest of a mounting wave of corporate cutbacks," and sparked a flurry of angry newspaper articles, decrying the greed of corporate management at the expense of the labor force. The financial press seemed pleased with the decisions, as investors boosted the stock over the first week following the announcement. The New York Times, whose main section ran the article about the effects of corporate greed on defenseless middle-class workers contained a story in the business section entitled "AT&T: A Leaner Company without a Crash Diet." The article praised a "kinder, gentler sort of downsizing," and described the multitude of severance packages and benefits that terminated workers could choose.

Nevertheless, the enthusiasm for AT&T's layoffs soon dwindled as the market realized the costliness of firing workers, the reduction in experienced and skilled labor, and the actual willingness of workers to leave. AT&T had only anticipated for 6,500 to leave voluntarily; however, 12,000 experienced managers opted for the severance package, confident that they could find employment elsewhere, given the increasing demand for skilled management in the technology industry. By mid-March, AT&T began a public relations campaign, taking out full-page ads in 29 national newspapers entitled "Wanted: Good Jobs for Good People," encouraging other employers to hire its terminated workers. AT&T also quietly-yet publicly-revised the number of workers it was planning to eliminate to 18,000, rather than 40,000. It also seemed that the downsizing would be costly, as many workers eligible for generous severance pay chose to leave, while many who would receive stingier packages resisted leaving voluntarily. Realizing that AT&T might mismanage the downsizing, the market punished AT&T by sending it down about 3% for the week; by contrast the S&P 500 rose 12 points to 652 over the same time period. Both CEO Robert Allen and President Alex Mandl resigned during the following months. After some turmoil in the executive management, the firm was given to current CEO Michael Armstrong to lead; the stock rose over 44% in his first few months at the helm.

Yet even these downsized downsizings never took place on the scale that AT&T had promised. Employment numbers obtained from AT&T's investor relations department show that AT&T's employment roster in its continuing operations shrank by about 2,000 at best, and actually rose over calendar year 1996. In a press release exactly one year after the original 40,000 job cuts were announced, AT&T said that it cut about 7,700 workers in its computer communications business, but added jobs in its internet and long distance business. In many cases, workers simply transferred from one department to another.

Perhaps by that time, the market had become convinced that downsizings were not the answer to the market's insatiable need for rapid asset appreciation. The market began to punish downsizers such as the Bank of America and US Airways as more labor cuts were announced. As a matter of irony, in January 1998, two years after former CEO Bob Allen first announced the massive round of 40,000 job cuts, current CEO Michael Armstrong announced that AT&T would be cutting 18,000 jobs as part of a major cost-reduction plan. AT&T also announced its fourth quarter earnings of 81 cents per share, a record for the firm, which beat the estimates of many analysts (71 cents a share). The market reacted by dumping AT&T stock, which fell more than $3 for the day to $61. It seemed that AT&T was up to its old tricks of announcing massive job cuts in accordance with vague corporate "aggressive targets to cut costs, particularly sales, general, and administrative expenses, to make us competitive with our rivals of today as well as of the future." Since most of the downsizing would take place in middle management, the union that represents AT&T's rank-and-file workers responded to the downsizings in quite a different tone of voice than it had two years earlier. On CNN, union spokeswoman Laura Unger said, "I think it's a good thing that they're finally going to be downsizing...because there is a lot of waste."

Perhaps the stock fell since the market had changed its viewpoint on senseless downsizing. Perhaps the market viewed this as a signal that a company's costs and production capacity are excessive relative to the demand expected by management. It remains to be seen if these job cuts will actually occur, unlike the downsizings of 1996 that never happened. Quite ironically, AT&T is still criticized in the popular press as the poster child of senseless and cruel job cuts, although the plan to eliminate workers was never actually carried out.

The striking irregularity that the AT&T case presents is why strong firms choose to downsize. Before the downsizing craze, downsizing was considered a sign of weakness, as firms laid off workers in lean times. If anything, weak firms would attempt to maintain their labor force, afraid to downsize, lest the market realize it was weak, and punish it. What could possibly motivate a strong firm to downsize? The event study failed to attach significance to a number of variables that a casual observer would consider crucial information pertaining to a downsizing, such as a bad earnings release. Therefore, if this information has nothing to do with the return to a downsizing firm, what motivates firms to downsize or not, and what motivates investors to buy or sell stocks of firms that downsize?

The main handicap to using econometrics in the event study, is that such techniques assume that there is only one equilibrium for a given set of information pertaining to good news, merger announcements, and percentage of workers laid off. The equations could not express why two companies with the exact same data had markedly different receptions in the stock market (Woolworth's and Sears, for example). It may be the case that there exist multiple equilibria for the same set of data.

Perhaps the event study did not correctly model the importance of asymmetric information and uncertainty surrounding stock prices. At any given instant, the managers have much better information about the health of the firm than the shareholders. The interaction between shareholders and managers can be modeled as a game, in which the shareholders are not as enlightened about the value of the firm, and try to interpret the manager's publicly-announced decisions (such as a downsizing) as a signal for how the firm is actually faring.

Since investors don't have as clear a picture as to the firm's health as management does, they will attempt to infer from the management's actions what the true value of the firm is. Assume a firm is laying off workers. This could be because the firm expects difficult times ahead, and is reducing its production capacity to accommodate the lean year anticipated. Since profits are expected to sag, the investors will bid down the stock value after the firm announces its layoffs. Or, the management could be expecting good times ahead, and realizes that some of its labor force is deadweight (perhaps workers are more productive than they used to be, so fewer workers are required for the level of output anticipated), so it lays off workers to save money. Here the investors should bid the stock price up, since profits are expected to rise. The management is the only player who knows the future prospects of the firm, and the investors only get to observe the labor policy, and must infer from that choice what the future health of the firm is.

Consider the following game with three players: a firm, its investors, and "Nature." Nature chooses whether or not the market for the firm's products will be powerful (P) or weak (W) over the next few years. The management has a plethora of internal information as to the market conditions (its market share, customer loyalty, product innovation, etc.) and can correctly anticipate its short-term profitability. The investors do not have access to this information. Based on the firm's anticipated health, the management chooses the number of workers, and pays them a market-determined wage. This determines output, and assuming that the firm is a price-taker, it also determines profitability. The choices of the firm are to downsize (D) or maintain the current level of employment (M). The investors observe this move, and then make a decision (within their information set): buy the stock (B) or sell it (S). The order of their moves is given in Figure IV.4.

The stock price in period 0 is only a function of whether or not investors bought or sold the stock, since there is no direct information revealed to them about the strength of the firm. The value of the firm is simply a function of whether or not investors decide to include it in their portfolios, and not a function of the underlying strength of the firm. The actual value of the firm is revealed in period 1, after the game has ended, and is a function of the inherent strength of the firm and of the quality of the management's decisions. For example, if the managers of a weak firm refuse to downsize, this will hurt the actual firm value. Using the notation developed above for strategies, the firm value in period 0 is a function of whether or not investors will buy or sell.

Likewise, the value in period 1, once the strength of the firm is revealed to all players, is a function of the management's strategy and the inherent strength of the firm,

Clearly, the investor's payoff is a function of the current value of the firm (period 0) and the future value of the firm (period 1). The same applies to management, whose payoffs (via stock options and similar profit-related compensation) are dependent of the firm's value:

Without too much of a loss to generality, the payoffs are as follows. The management gets two points for having the stock price bid up. Presumably, they hold stock options, or participate in profit-sharing plans. Also, if the stock is bought, perhaps the board of directors will decide to renew their contracts in the near term. The management will receive zero points if the stock is sold. If the management makes the right strategy move, they will receive an additional point. For example, if the firm is weak, the right strategy move is to reduce capacity and fire workers. The management will receive no points for making the wrong choice (i.e., not laying off workers in a weak economy).

As for the investors, they will receive one point for buying the stock of a strong company. If the whole market has a rosier outlook for the firm, then the stock will perform better, and investors will make a profit. Since investors are risk-averse, they will be rewarded with two points if they sell (or short) the stock of a weak company. Such investors decide to dump a stock in advance of finding out that market conditions are weak. If the investors dump the stock of a firm, and the firm turns out to be in a strong position, they will miss out on the rise in equity value, and receive no points. Likewise, if the investors buy the stock of a weak company, they will receive nothing. Investors also care about how the firm is being run; in this game, good management increases the stock price. If the company downsizes in a weak market or maintains the labor force in a strong market, the investors holding the stock will be rewarded with an additional point. If the investor is holding the stock of a company whose management does not make the right decision, he will receive zero points.

The receipt of points in this game can be considered the discounted value of utility that the player will receive over his lifetime. The payoffs are known with certainty, since the only uncertainty in the model lies in what decisions the players will make, including Nature, which chooses the state of the world. Clearly, players care only about maximizing their expected utility. The preferences of risk and reward are already incorporated in the values that the utility function creates.

A set of behavioral strategies is a SPNE if and only if the utility derived from that strategy is at least as good as the utility of any other strategy, with the other player's strategy given. In terms of utility:

where and are any other possible strategy Players 1 and 2 could respectively undertake.

Player 2's decisions are made within information sets. When a player makes a decision at an information set, he does not know at which of the nodes he is playing. In this case, Player 2 does not know if the market is weak or strong. Nature moves first, then Player 1, then Player 2.

This game has two sub-game perfect Nash Equilibria. Here is the first strategy profile.

Player 1 never fires any workers no matter what, so his employment decision is independent of the strength of the firm. So the probabilities can be written as

Player 1's strategy specifies that Player 2's right information set will be reached with probability 100%. When Player 2 is reached, he uses Bayes' Law to update the probability of which node he is at. This is done as follows.

The expected payoff to the investor of selling the stock is given as the payoff in each state of the world (weak firm or powerful firm) weighted by the Bayesian probability of that state:

The expected payoff to buying is:

The payoff to buying is clearly higher, so Player 2 will always buy if he observes that Player 1 is not downsizing. Player 2 needs to specify a strategy at his left information set to discourage Player 1 from downsizing (particularly a Player 1 facing a weak market).

The Bayesian probabilities of being reached are meaningless. Modern game theory allows us to specify any set of beliefs that justify Player 2's decision to sell the stock of all downsizing firms at his un-reached left information set.

The payoff to selling the stock is

The payoff to buying the stock is

Any x lower than 1/2 will cause Player 2 to sell the stock whenever his left information set is reached. Now that Player 2's strategy has been specified given Player 1's, we need to ensure that Player 1 has no profitable deviations from his strategy. If he finds himself in a weak market, his payoff will be 1 if he downsizes (Player 2 will sell the stock), and 2 if he does not (Player 2 will buy the stock). So Player 1 will not downsize if the market outlook is weak. If Player 1 finds himself in a strong market, he will receive 0 if downsizes (Player 2 will sell the stock) and 3 if he does not (Player 2 will buy the stock). So Player 1 will not downsize if the market is strong. No matter what the market outlook, Player 1 will not downsize given Player 2's strategy. Since we have proven that Player 1's strategy is a best response to Player 2's in all states of the world, and that Player 2's strategy is a best response to Player 1's in all states of the world, this is a sub-game perfect Nash Equilibrium. In this case, downsizing is perceived as a symbol of weakness, so no firm downsizes. Weak firms are able to masquerade as strong firms by maintaining their workforce.

The second and final equilibrium of this game is the one which is empirically observed for firms in the mid 1990s. All firms, strong or weak, downsize. Investors will dump the stocks of firms that refuse to downsize, and buy the stocks of the firms that do reduce their workforce. Here, firing workers is perceived as a sign of financial strength. It might demonstrate that the firm is re-focusing its core businesses, jettisoning any non-essential assets, or reducing production in areas in which its market share is lagging. Firing workers is perceived as a signal that the management is preparing the firm for a rosy economic future. This strategy is also a SPNE.

Player 1's strategy specifies that Player 2's left information set will be reached with probability 100%. When Player 2 is reached, he uses Bayes' Law to update the probability of which node he is at. This is done as follows.

The expected payoff to selling the stock is

The expected payoff to buying is

The payoff to buying is clearly higher, so Player 2 will always buy if he observes that Player 1 is downsizing. Player 2 needs to specify a strategy at his right information set to discourage Player 1 from not firing workers (particularly a Player 1 facing a strong market).

The Bayesian probabilities of being reached are meaningless. Modern Game Theory allows us to specify any set of beliefs that justify Player 2's decision to sell the stock of all downsizing firms.

The payoff to selling the stock is

The payoff to buying the stock is

Any y less than 1/2 will cause Player 2 to sell the stock whenever his right information set is reached. Now that Player 2's strategy has been specified given Player 1's, we need to ensure that Player 1 has no profitable deviations from his strategy. If he finds himself in a weak market, his payoff will be 3 if he downsizes (Player 2 will buy the stock), and 0 if he does not (Player 2 will sell the stock). So Player 1 will downsize if the market outlook is weak. If Player 1 finds himself in a strong market, he will receive 2 if downsizes (Player 2 will buy the stock) and 1 if he does not (Player 2 will sell the stock). So Player 1 will downsize if the market is strong. No matter what the market outlook, Player 1 will downsize given Player 2's strategy. Since I have proven that Player 1's strategy is a best response to Player 2's in all states of the world, and that Player 2's strategy is a best response to Player 1's in all states of the world, this is a sub-game perfect Nash Equilibrium. In this case, downsizing is perceived as a symbol of strength, so all firms downsize. Strong firms ironically masquerade as weak firms by reducing their workforce.

Most importantly, the results of this game are not too sensitive to the arbitrary utility numbers that I assigned to each state of the world for each player. Manipulation of the relative utility values consistent with the "real world" does not significantly alter the results. The implications for such a strategy, in terms of whether or not it actually increases shareholder value in the long run, remain to be seen. Chapter V will address issues relating to how sensitive this result is to the payoffs and game structure I have chosen.

There are no other equilibria to this game. If a strong Player 1 maintains the workforce, so will a weak Player 1, or else the investors will know he is weak, and punish him by selling his stock. Likewise, if a strong Player 1 always downsizes, so will a weak Player 1. If a weak Player 1 maintained the workforce ever, the market would know he is weak, and would sell his stock. It is also impossible for both Players to randomize, since there is no strategy of Player 2's that would ever induce both strong and weak firms to be indifferent about downsizing. This concept will be explored in detail in Chapter VII for the case of United Airlines.

Several fascinating results emerge from this simple model. Firstly, all firms behave the same, regardless of the fact that some of them find themselves in weak markets. Until now, downsizing has been thought of as a signal to the market as to whether the firms are weak or strong. Yet, the only equilibria of this game are ones in which either everyone downsizes, or everyone does not. Downsizing reveals no information whatsoever about the strength or weakness of a firm, so the indication that a firm is planning on cutting most of its labor force will not help the market to update the probability (using Bayes's Law) of whether or not a firm is weak or strong. Absent of this, investors make an arbitrary decision of whether or not they like downsizing, and firms will cater to investors by delivering to them the result they want to see.

Secondly, the two equilibria are completely different. The incentives and payoffs to management and investors are exactly the same in both equilibria, but in one, everyone downsizes and in the other, no one dares to downsize. Both equilibria are perfectly plausible in the Nash sense, and are observed in the stock market. The first equilibrium explains why a weak firm will not try to downsize, although it would be the right business move. By not downsizing, it hides its weakness from its investors. The second equilibrium explains why a strong firm would downsize, even though it has no economic reason to lay off workers. These sub-optimal decisions on the part of firms are a direct result of the lack of information flow between investors and management.

Assume the market adores downsizing firms. The question arises: why can a strong company not simply verbally announce to the market that it is strong, and then avoid having to downsize? For this game, the answer might lie in the fact that actions speak louder than words. Why should an investor believe that a given firm claiming strength actually would be dominant in the market? Investors only care about the actions of the management (such as the employment decision) and not vapid or vague statements about the company's outlook. In these games, seeing is believing. If a firm wants to prove that it is strong, stating it is not enough; it must behave as strong firms are expected to behave. The problem arises, that occasionally, a weak firm will be able to fool people by acting like a strong firm, or a strong firm (for whatever reason) will pretend to be weak. No action specifically needs to be taken in this game on the part of the firm; all it needs to do is announce that it will downsize. Exactly how it downsizes, and over what time period is not as relevant. As the case of AT&T proves, merely announcing a downsizing, without ever carrying it out, is enough to boost the stock price, at least for a while.

Neither equilibrium seems more plausible than the other does. Consider the early 1980s, in which US manufacturers were firing workers as several forces, such as high interest rates, and cheap foreign labor, were causing them to cut production capacity. Here, the economy was in the first equilibrium, in which laying off workers was a sign of weakness. Then, GE's heavily-publicized and massive downsizing of a quarter of its workforce taught the market that firing workers increased profitability, so the world found itself in the second equilibrium, in which investors loved the stocks of companies which downsized, so all firms downsized.

The next chapter will explore the specifics of the game, including altering the payoffs and circumstances to see how different conditions will change the outcome.