My Years With General Motors

by Alfred P. Sloan Jr.

  The details of the new formula were worked out in three days of almost continuous bargaining. The contract was, as I have said, for two years; its novelty made the union unwilling to commit itself for a longer period. The annual improvement factor was set at three cents an hour for every worker covered by the contract. And the base date finally agreed upon for use in calculating the cost-of living formula was 1940—the last year in which prices had been fairly stable.

  There are several points about the General Motors formula that need to be understood. First, as to the annual improvement factor, which is often misinterpreted even by persons familiar with labor questions. Section 101(a) of the contract, which deals with the improvement factor, states "that a continuing improvement in the standard of living of employees depends upon technological progress, better tools, methods, processes and equipment, and a cooperative attitude on the part of all parties in such progress ... to produce more with the same amount of human effort is a sound economic and social objective . . ." In other words, the real source of income is productivity. The union's acceptance of these sensible words was surely a milestone in labor relations.

  But contrary to a widespread assumption, the improvement factor is not linked to a definitely known increase in productivity at General Motors. There is, to my knowledge, no satisfactory technique with which to measure productivity at General Motors, or, in fact, at any corporation which manufactures constantly changing products. And even if an industrial productivity measurement could somehow be provided, it would still not be desirable to relate it in direct proportion to wage increases. Such a policy applied to the economy as a whole would bring about intolerable discrepancies between the wages paid in industries where technological progress is rapidly increasing, and the wages paid where technological progress must be limited—as it is in the so-called service industries. It is my belief that the improvement factor should reflect the long-term productivity increase of the U.S. economy as a whole.

  Over the years, it had been estimated, productivity in the United States had been rising about 2 per cent a year. How good that estimate was, I do not know. In any case we set the improvement factor in the General Motors formula at three cents an hour. This amounted to a 2 per cent annual increase in wage rates : three cents is 2 per cent of $1.49, which was then the average hourly rate at General Motors. In subsequent negotiations the improvement factor has been increased several times. Note that the corporation commits itself to deliver this increase over the period of the contract no matter what actually happens to the productivity of U.S. industry. Even if productivity should decline, for the country as a whole or for General Motors in particular, we would still be obliged to pay the annual improvement factor.

  I have always felt that it was a source of confusion to label the improvement factor a "productivity increase." I prefer to think of it as a group merit raise of sorts, and I suspect that many General Motors employees see it in that light.

  In the end, increased efficiency flows not so much from the increased effectiveness of the workers, but primarily from more efficient management and from the investment of additional capital in labor-saving devices. Some union spokesmen talk as though the entire benefits of increased productivity should go to labor. I do not believe that is sound. New machinery costs money and the additional investment must be justified by a return on that investment. An argument could be made that the consumer, and the economy as a whole, would benefit most if productivity increases were applied entirely to the lowering of prices. Ideally that might be a good thing. But since it is in the nature of people to work better with the incentive of an individual or group gain and to want to bargain over it, it is a good thing to have something to bargain over. And so I conclude that the benefits of productivity increases should be apportioned among the consumer (lower prices or better product), labor (higher wages), and the shareholders (return on investment).

  The improvement factor as it was first applied in the General Motors formula had one curious effect. Under the 1948 and 1950 agreements, the improvement came to exactly the same rate per hour for all workers, floor sweepers and highly skilled tool and die-makers alike. The decision to take the average worker and pay 2 per cent of his rate (that is, three cents an hour) to all our men was clearly a move in the direction of equalitarianism. The effect, of course, was that the improvement factor given the tooland diemakers did not come to as much as 2 per cent a year, while the sweepers were getting an increase of perhaps 3 per cent. Thus from 1948 to 1955, the improvement factor had a tendency to narrow percentage wage differentials. This tendency was corrected in the 1955 contract, which put the improvement factor on a basis of 2.5 per cent—with a minimum of six cents—for all workers.

  While the improvement factor was changed, the escalator formula remained substantially unchanged—even though it, too, has tended to equalize rates of pay over the years. Here again, there was no theoretical reason why the formula could not have provided, simply, for a 1 per cent increase in pay every time the cost of living rose by 1 per cent. Instead, the formula was set so that every worker would get the same amount of money added to his pay when living costs rose. The cost-of-living program was calculated in this manner: First, it was determined that consumer prices had gone up about 69 per cent since the 1940 base date. The hourly rate of the average General Motors worker had gone up only about 60 per cent in that interval. To make up the 9 per cent difference, eight cents an hour was added to the rate. But this increase obviously amounted to more than 9 per cent for low-paid workers, and less than that for higher-paid workers. In its provisions for future increases, the escalator clause had a similarly equalizing effect. We took the average hourly wage and the April 1948 Consumer Price Index—the latest available at that time—and determined to preserve the relationship between them. Dividing the average wage of $1.49 into the index figure of 169.3, we came up with a one-cent increase for each upward movement of 1.14 points in the price index; accordingly, this became the rule for all our workers. But note again that our highest-paid workers really needed more rapid increases to keep up with the cost of living, while a janitor making $1.20 an hour in 1948 would clearly be more than compensated for any inflation. In thinking about the cost-of-living part of the General Motors wage formula, it is important to remember that it is the average wage rate that is really tied to the price index, and that the escalator clause tends to pull all other wage rates toward the average. Whether this equalitarian effect is something good or bad in the long run, I am not prepared to say. I think it is interesting to observe that the many wage formulas adopted by other unions in imitation of our own program almost invariably tend to preserve this equalizing feature.

  On several occasions we have granted skilled-trades employees special increases. These increases have offset the equalitarian effect of the formula as far as the skilled men are concerned. Altogether, special increases for tooland die-men for the period 1950 through 1962 have amounted to thirty-one cents an hour.

  In other ways, too, the original concepts underlying the wage formula have been deflected somewhat by the exigencies of collective bargaining. One recurrent problem has been the "floor" under wages in our cost-of-living agreements. As I had suggested in my original letter to Mr. Jordan, workers would want some limit on wage reductions, even in a period of severe deflation. In the 1948 agreement we specified that no matter how much the cost of living might fall, no more than five cents of the original eight-cent cost-of living raise could be taken away. Again in 1953, 1958, and 1961, the "floor" was raised by an agreement between the union and the corporation. The logic of escalator clauses apparently cannot be extended to periods of severe deflation because workers are always reluctant to take wage cuts.

  The 1953 negotiations, incidentally, present an interesting example of the kind of public pressures which have been steadily exerted on the General Motors wage formula, and which have made it difficult for the formula to operate as Mr. Wilson originally intended. In principle, there sho
uld not have been any 1953 negotiations, because in the five-year contract signed in 1950, the union "unqualifiedly waive [d] the right" to bargain on any issue covered in the contract. But toward the end of 1952, the UAW became dissatisfied with the cost-of-living provisions of the contract. The union feared, as many others did at the time, that the post-Korea inflation had about run its course. If the cost of living declined, its members would lose some, conceivably all, of the special allowance which they were then receiving under the escalator clause. To make matters worse, the Wage Stabilization Board had allowed other groups of workers, that is, in the steel, electrical, and other industries, to receive living-cost increases which were added to their base rates. In other words, a deflation would mean that UAW members would lose some take-home pay, while other union members would not. We agreed with the union that wages at General Motors should not lag behind those in comparable industries. And so we reopened the contract and incorporated nineteen cents of the cost-of-living allowance (which was then up to twenty-four cents) into the permanent base rate. This episode illustrates the difficulties of adhering strictly to the original concept of the wage formula.

  Our wage formula has often been attacked as inflationary. I would agree with Mr. Wilson that the formula itself does no more than protect our employees against inflation. However, the formula is by no means our whole labor contract. Because we have granted many fringe benefits some critics maintain that the cost increases are in excess of productivity and that, therefore, the formula plus the fringe benefits may have inflationary implications.

  There is another factor of importance that must be considered. In discussing the improvement factor, I have said that, in my opinion, it is more of an incentive or a bonus than it is an improvement factor per se. And from that point of view I think the fact that our workers benefit on a definite and prescribed basis, resulting in an increase in their standard of living, gives us a more sympathetic co-operation in the introduction of labor-saving devices and other improvements that flow from technological progress, which on the whole have a healthy influence on the efficiency of the corporation's operations.

  It is undeniable—as of this writing—that the formula has served to bring relative peace and stability to our labor relations. We have had no extended strikes over negotiation of a national agreement since the formula went into effect in 1948.

  The most widely publicized addition to the General Motors labor contract in the last few years has been the provision for supplementary unemployment benefits—a provision often, though not accurately, described as the guaranteed annual wage. As all the major automobile companies approached the 1955 labor negotiations, it was apparent that the union regarded this program as a great milestone in its history, and intended to win it at all costs. A large part of the idea behind this program—that is, the idea of an employer-financed supplement to state unemployment compensation—had already been worked out by the union, apparently in 1954 and 1955. However, the plan proposed by the Ford Motor Company and finally accepted, differed in many respects from the union's specific proposals—and was much more conservatively financed. We agreed to this program shortly after Ford did, although we disagreed with several aspects of the plan. Ultimately, the entire industry conceded the point.

  Actually, we at General Motors had been considering alternative plans of this broad type for about two decades. In December 1934, before the state unemployment-insurance laws were on the books, some thoughts on a private insurance program for the corporation's own employees were outlined. Among the ideas suggested, we endorsed these:

  General Motors subscribes to the principle of accumulating reserves to be paid to employees in periods of involuntary unemployment.

  We also subscribe to the principle of joint contribution to such reserves by both employers and eligible employees.

  We also believe in the justice of a probationary period before any employee becomes eligible.

  I was impressed with the merit of these points and so, I think, were most of my colleagues. However, the sudden growth in federal and state unemployment insurance programs in the mid1930s altered our perspective on the problem. With insurance against unemployment now available, we developed a program designed to allay the hardships caused by our cyclical production. It worked in general this way: Any employee with five years' seniority who was temporarily laid off—in a model change-over, for example—or who was earning less than twenty-four hours' pay in a week, could borrow from the company each week the difference between his earnings and twenty-four hours' pay. No interest was charged. In weeks when he had an income in excess of twenty-four hours' pay, he would repay the loan at the rate of one half of the excess over twenty-four hours' pay. In the case of employees with less than five years' but more than two years' seniority, the corporation made advances up to sixteen hours' pay with a maximum aggregate advance of seventy-two hours' pay. In other words, the earnings of our workers were being spread more evenly over the entire year. The program was discontinued when defense production made it unnecessary.

  In addition to this program of interest-free loans, we began to consider whether we could somehow guarantee a substantial proportion of our workers some minimum number of working hours during a year. The Social Security Act of 1935 included one section which was intended to give employers an incentive to devise such plans. Under this section, employers who guaranteed their workers 1200 hours of work a year were exempted from paying the 3 per cent payroll tax. We seriously considered offering some such guarantee to our workers in 1938. However, Donaldson Brown, then vice chairman of the board, stated the case against it very persuasively. In a memorandum to me dated July 18, 1938, Mr. Brown argued that the guarantee could not be extended to very many workers—or that if it could, these workers could not be guaranteed very many hours. Further, he said:

  The extension of a guaranteed annual number of hours of employment to a given segment of the employees inevitably will tend to freeze the average hours of employment at that level. A plan of the kind would be taken as implying the purpose—in event of declining business—to spread work to the end of averaging hours at the guaranteed level. Union pressure towards this result inevitably would be exerted.

  All of us were dubious about the feasibility of work-sharing in such a large and complex organization as General Motors. Personally, I regarded work-sharing at low levels of hours over long periods as unsound, economically and socially. But in the early post war period, I felt that the corporation would have to devise some kind of guarantee. On May 15, 1946, I expressed my view on supplementary unemployment compensation:

  ... if we could determine what the limitations are we might get ahead of the pressure that is going to be put on us, and determine in our own way and in a factual way, just how far it would be practical to go, which might result in a better relationship between our people and ourselves without the liability of paying for work that was not accomplished.

  On balance, the plan which was finally written into the contract seems to me to represent less of an innovation than its proponents believe it to be. As many economists have pointed out, the plan is merely another extension of unemployment insurance, which has been in effect for more than twenty years, and which has always been financed by the employers. I suspect that the real benefit of the new plan is not simply the degree of protection it will give to workers in slack periods; after all, many workers will always be ineligible for coverage, and many other workers will receive only small payments. Rather, it is that the plan gives our workers a greater feeling of economic security; and perhaps in the long run that is merit enough.

  Before 1933 General Motors had no dealings with labor unions, except for a few craft organizations in the construction field. For this and perhaps other reasons we were largely unprepared for the change in political climate and the growth of unionism that began in 1933. One is inclined to forget that unionization in large industries was not then the custom in the United States. The significance of large-scale unio
nization was not yet clear to us. We knew that some political radicals regarded unions as instruments for the attainment of power. But even orthodox "business unionism" seemed to us a potential threat to the prerogatives of management. As a businessman, I was unaccustomed to the whole idea. Our early experiences with the AF of L unions in the automobile industry were unhappy; the chief issue with these unions became organizational. They demanded that they represent all our workers, even those who did not want to be represented by them. Our initial encounter with the CIO was even more unhappy; for that organization attempted to enforce its demands for exclusive recognition by the most terrible acts of violence, and finally seized our properties in the sit-down strikes of 1937. I have no desire to revive the bitter controversies that arose over these early encounters with labor organizations. I mention them merely to suggest one of the reasons why our initial reaction to unionism was negative.

  What made the prospect seem especially grim in those early years was the persistent union attempt to invade basic management prerogatives. Our rights to determine production schedules, to set work standards, and to discipline workers were all suddenly called into question. Add to this the recurrent tendency of the union to inject itself into pricing policy, and it is easy to understand why it seemed, to some corporate officials, as though the union might one day be virtually in control of our operations.

  In the end, we were fairly successful in combating these invasions of management rights. There is no longer any real doubt that pricing is a management, not a union, function. So far as our operations are concerned, we have moved to codify certain practices, to discuss workers' grievances with union representatives, and to submit for arbitration the few grievances that remain unsettled. But on the whole, we have retained all the basic powers to manage.