Employee Ownership Companies Pay Less for Workers' Compensation CostsAn NCEO study found that employee ownership companies have lower workers' compensation insurance rates than comparable non-employee ownership firms. Leslie Hakala authored the study. She began the project as an NCEO research intern and completed it for a thesis requirement at Harvard University. The study was unable to ascribe a specific causal relationship between employee ownership and lower workers' compensation costs, but it did find that these costs declined as employee ownership plans matured.
BackgroundIn 1989, the last year for which we have data, U.S. employers spent over $48 billion on workers' compensation costs. These costs grew at 16.9% per year in the mid-1980s. Cost increases were partly attributable to increased benefits mandated by state workers' compensation insurance reforms. At the same time, as employer provided health care coverage has declined, more employees sought to cover health problems under workers' compensation. Many people believe there has been increased fraud as well.
Workers' compensation programs vary from state to state, but in most programs, insurers attempt to provide employers with an incentive to limit safety problems by developing an experience rating. The ratings compare an individual firm's experience with other firms of its type. If the rating is better than average, insurance premiums will be lower; if it is worse, they will go up.
In this study, we looked only at California firms. In California, employers are assigned a "manual rate," an insurance rate expressed as a percentage of every $100 of payroll. Rates are assigned to all companies based on their industry classification. These rates are then adjusted for companies with a premium above a certain level according to their actual experience. This means smaller and less risky firms are not assigned an experience modification rating. The experience modification rate is set for each year based on three years of past experience, excluding the most recent year (because data are generally not yet available). The experience modification rate is determined by looking at actual experience modified by a size weighting factor. For larger firms, the adjustment may be very small; for smaller firms, actual experience is given a lower weight because a single incident can skew results dramatically. This weighted experience rating now becomes the "experience modification" figure.
Theoretically, the average experience modification factor for any business classification should be 100%. A company with a good record would have a rating under 100%; a bad record would rate higher. These numbers are then multiplied by the manual rate to set the premium. In practice, the average rating is somewhat under 100.
Study DesignThis system provides a good way to assess whether employee ownership has any impact on an important area of corporate operations. Experience modification ratings can be lowered by employees being more careful, by company policies, such as worker safety teams, that find ways to reduce risk, or by workers filing fewer claims. The ratings are not a perfect measure, however. Less weight is given to smaller companies' experience in worker modification rates. Because averages are calculated over three previous years, with a one year lag, there may be few changes in the ratings showing up in the initial years after an employee ownership plan is established. There is also no way to determine whether companies implemented other kinds of cost reducing initiatives coincident or following the establishment of employee ownership, such as setting up safety teams. If these teams would have been set up anyway, then the impact of employee ownership is muddled.
Our data set consisted of 117 California employee ownership firms from NCEO lists. We were able to obtain experience modification ratings for these companies from 1981 through 1993 and statewide averages for 1981 through 1992. We then compared the averages of companies that had a plan in place during any of those years with the statewide averages.
Study ResultsThe results of the study are summarized in the table below. The table shows that employee ownership firms have a significantly lower average rating. If all U.S. companies had employee ownership, workers' compensation costs would be $4 billion per year lower. The chances that these results are random coincidence are virtually zero. But this does not tell us if employee ownership causes the lower numbers.
To evaluate that question, we ran several regressions, statistical tests that can evaluate multiple variables simultaneously. The tests showed two important results. First, the higher the percentage of ownership, the lower the rating. While this effect was not dramatic, a 100% employee owned firm would have a rating seven to eight points below an non-employee ownership firm. Second, the older the plan, the lower the rating. On average, for every two years of plan experience, the experience rating dropped one percent. This helps explain why the difference between employee ownership firms and the rest of the sample increased over time (more firms' ESOPs had matured by the end of the period).
|Year||Statewide Experience Modification Rates||Employee Ownership Company Average Experience Modification Rates|
|1981 ||95.3 ||88.0|
|1982 ||93.9 ||92.2|
|1983 ||97.8 ||94.0|
|1984 ||97.1 ||92.8|
|1985 ||98.7 ||92.8|
|1986 ||98.0 ||94.3|
|1987 ||96.0 ||94.4|
|1988 ||95.0 ||92.0|
|1989 ||92.3 ||86.5|
|1990 ||95.0 ||86.8|
|1991 ||91.2 ||83.1|
|1992 ||94.3 ||87.3|
|1993 ||n/a ||89.4|
On the other hand, the regression analysis showed that employee ownership had no immediate impact in the first few years after the plan was set up, even after adjusting for the time lag in the rating system. To the extent there is any relationship between employee ownership and lower costs, it appears to take some time to set in. Just how that relationship develops requires further investigation. In particular, we need to evaluate whether there is a synergy between ownership, employee participation programs, and performance results, as has been found in other studies.