Historical Evidence on Profit-Sharing
Short life Schemes

In recent years the arrangement whereby workers share in the profits of the firm in which they are employed has attracted much attention and has become a focus of a lively public policy debate. In an important book in 1984, Weitzman argued that such a system would have important macroeconomic effects, stabilizing the economy and leading to lower average unemployment rates, but both this claim and the microeconomic implications of profit- sharing have proved controversial. It is therefore surprising that the long history of profit-sharing in Britain, going back 120 years, has been ignored in this debate. This paper provides such a historical perspective by examining the history of profit- sharing up to 1913, by which time it had become an established practice.

In the nineteenth century writers such as Charles Babbage, J S Mill and W S Jevons were among those who advocated profit-sharing as a means of raising labour productivity and improving the climate of industrial relations. The sources of increased productivity and the consequent increase in profits claimed by contemporary supporters were thought to arise in several ways. First, there were thought to be direct effects on worker effort. Even where incentive systems already existed, it was suggested that profit-sharing might lead to improved quality of output, better conservation of tools and machinery and reduction of waste in raw materials. Second, it was argued that profit-sharing would reduce supervision costs not only through individual motivation but also by workers monitoring each other's performance. Third, profit-sharing was thought to engender greater loyalty to the firm among existing workers as well as enabling the firm to attract and retain the most productive workers. Fourth, it was expected to improve industrial relations not only by reducing strikes but also, more importantly, by allaying everyday frictions at the workplace. Finally, increased consultation was expected to improve the flow of information in both directions, leading to more efficient working practices.

A series of firms adopted, or experimented with, profit-sharing schemes before the First World War and the issue drew the attention of writers, industrialists and trade unionists. The first significant profit-sharing experiments began in France in the 1840s and in Britain in 1865. Much of the attention of contemporaries and historians has been drawn to two prominent cases, that of Henry Briggs and Co., whose scheme began in 1865 and the South Metropolitan Gas Co., which adopted profit-sharing in 1889. These schemes were intended to counter rising industrial unrest and the spread of trade unionism; many other schemes, which were not directly aimed at countering trade unionism, have attracted less attention.

The importance of profit-sharing is reflected in the series of Reports on profit-sharing, co-partnership and similar arrangements published by the Board of Trade between 1891 and 1920. These provide details of almost all schemes known to have existed and provides much of the material on which we draw in this paper. Some 299 schemes are known to have commenced up to 1912, of which 133 (employing 106,000 workers) remained in existence at that date. Most schemes paid a cash bonus, though in some firms the workers' profit share was credited to a provident fund and others operated employee stock ownership plans.

Did these schemes yield returns to workers sufficient to create non-negligible effects on incentives? Over the decade 1901-11, payment from profits added 5.5% to the wages of eligible employees in reporting companies, the average varying from 4.5% in 1908 to 7.1% in 1906. Thus for the eligible worker profit- sharing typically amounted to over two weeks' extra wages each year. Although these additions to wages were significant, they could easily be eliminated if profit-sharing firms paid correspondingly lower wages. One recent study found that profit- sharing firms in engineering paid 3% of the wage as a share of profits but that total renumeration was lower than in other engineering firms. One third of the trade unions responding to enquiries for the 1894 Report alleged that profit-sharing firms were paying a basic wage less than the union rate. Such employees, however, might not in any case have obtained union rates. A key question is whether these higher wages were reflected in higher productivity. Unfortunately we have little in the way of direct evidence, although some qualitative evidence can be gleaned from the replies to employer questionnaires conducted for the 1894 and 1912 Reports.

The fact that 55% of the schemes started before 1913 had been abandoned by that time is often seen as telling evidence against the viability of profit-sharing. Examining the proportion of failures up to a given point in time is, however, an inadequate method of judging the survival of schemes. Clearly if survival is regarded as an important criterion of success or failure then the factors influencing the duration of schemes should be analysed, but such an analysis has not previously been undertaken for the period before 1914.

The data at our disposal are severely limited but they do allow us to investigate certain hypotheses. One would be that profit- sharing would be more successful where workers have higher levels of human capital and more complex skills, since there is a greater incentive to bind the worker to the firm and close supervision is more difficult. A second question concerns the influence of the type of profit-sharing scheme: while a cash bonus might be expected to provide a direct stimulus to productivity, provident schemes and employee stock ownership plans would be expected to engender greater loyalty to the firm. The size of firm might also be important: profit-sharing is a group incentive, and the free-rider problem might be less serious in small units. Alternatively the greater stability and more formal organizational structure of large firms may favour profit- sharing.

We model the probability of survival for these schemes using a Weibull distribution. We estimate the parameters of the hazard function, or age-specific failure rate, which allows us to investigate whether the probability of failure increases or decreases with the duration of the scheme. Negative duration dependence would arise from unmeasured differences between firms taking up profit-sharing. If, on the other hand, the introduction of profit-sharing had initially beneficial effects which then wore off over time, there would be positive duration dependence. In addition we allow one parameter of the hazard function to depend on variables such as the industry group, the type of scheme and the number of employees in the firm. We also include a dummy variable for the years of "new unionism" and the attendant industrial unrest between 1889 and 1892: this represents a crude test for whether schemes aimed industrial unrest were more likely to fail.

The estimated survival functions suggest relatively weak effects of industry type but a strong positive effect of firm size on the probability of survival. There is some evidence that cash and especially provident schemes raised the probability of survival relative to share-based schemes: perhaps the provident variable is correlated with some element of employer paternalism. Schemes including some share element had a smaller effect. This may indicate that shares had smaller incentive effects but also that these schemes were perhaps less generous than the other types. Firm size had a strong positive effect on duration, suggesting that greater continuity in organization or more formal structure may have outweighed the problem of free riders. Finally the estimates indicate that the hazard functions are upward sloping, i.e. that probability of survival appears to have decreased with the duration of the scheme. This suggests that the positive effects of profit-sharing may have eroded over time.