Research on Employee Ownership and Corporate Performance

Home » Articles » General Topics »

Return to regular version

Over the years, the NCEO has conducted and reported on research on employee ownership and corporate performance. The research comes to a very definite conclusion: the combination of ownership and participative management is a powerful competitive tool. Neither ownership nor participation alone, however, accomplishes very much. The findings apply most clearly to closely held companies. The relationship between ESOPs and corporate performance in public companies is more ambiguous. Limited research on broad-based equity compensation plans have consistently found positive results. Studies on ESOPs and corporate performance in public companies come to more mixed conclusion.

ESOPs and Corporate Performance

The 2000 Rutgers Study

In the largest and most significant study to date of the performance of ESOPs in closely held companies, in 2000 Douglas Kruse and Joseph Blasi of Rutgers University found that ESOPs increase sales, employment, and sales/employee by about 2.3% to 2.4% per year over what would have been expected absent an ESOP. ESOP companies are also somewhat more likely to still be in business several years later. This is despite (or perhaps because of) the fact that ESOP companies are substantially more likely than comparable companies to offer other retirement benefit plans along with their ESOP.

Kruse and Blasi obtained files from Dun and Bradstreet on ESOP companies that had adopted plans between 1988 and 1994. They then matched these companies to non-ESOP companies that were comparable in size, industry, and region. They then looked for which of these companies had sales and employment data available for a period three years before the plan's start and three years after. The sales and employment growth data were then compared for each year for each paired company. They also checked the companies' filings with the Department of Labor to determine which of the companies had other retirement-oriented benefit plans. Finally, they looked to see what percentage of the companies remained in business in the 1995 through 1997 period.

The process yielded 343 ESOP companies and 343 pairs for the overall sample. However, missing data meant that employment data were available only for 254 ESOP companies and 234 pairs, 138 ESOP companies and 77 pairs for sales, and 115 ESOP companies and 65 pairs for sales/employee (some pair companies could be used for more than one ESOP company).

The results showed that ESOP companies perform better in the post-ESOP period than their pre-ESOP performance would have predicted. The table below shows the difference in the pre-ESOP to post-ESOP period for ESOP companies on sales growth, employment growth, and growth in sales per employee:

Difference in Post-ESOP to Pre-ESOP Performance
Annual sales growth	+2.4%
Annual employment growth	+2.3%
Annual growth in sales per employee	+2.3%

It might be assumed that sales per employee would not go up by 2.3% per year since sales and employment growth differences were about the same, but, the researchers explain, the differing compositions of the samples for the measures makes such a simple comparison misleading. The relative growth numbers might seem small at first glance, but projected out over 10 years, an ESOP company with these differentials would be a third larger than its paired non-ESOP match.

The 1986 NCEO Study

The first study to show a specific causal linkage between employee ownership and corporate performance was by Michael Quarrey and Corey Rosen of the NCEO. The study looked at the performance of employee ownership companies for five years before and after they set up their employee stock ownership plans (ESOPs). It indexed out market effects by looking at how well employee ownership companies did relative to competitors in the pre- and post-ESOP periods, then subtracted the difference. For example, if a company were growing 3% per year faster than its competitors in the pre-ESOP period, and 6% per year faster in the post-ESOP period, there would be a +3.0% difference attributable to the ESOP, other things being equal.

The study found that ESOP companies had sales growth rates 3.4% per year higher and employment growth rates 3.8% per year higher in the post-ESOP period than would have been expected based on pre-ESOP performance. When the companies were divided into three groups based on how participatively managed they were, however, only the most participative companies showed a gain. These companies grew 8% to 11% per year faster than they would have been expected to grow, while the middle group did about the same, and the bottom group showed a decline in performance.

Participation alone, however, is not enough to improve performance. A large number of studies show that the impact of participation absent ownership is short-lived or ambiguous. Ownership seems to provide the cultural glue to keep participation going.

The New York and Washington Studies

Economist Gorm Winther and colleagues in New York and Washington State followed up the NCEO study, using the same research design but different samples, one of 25 employee ownership firms in New York State and one of 28 employee ownership companies in Washington State. In both studies, employee ownership per se had little or no impact on corporate performance, but a substantial impact when combined with participative management. In Washington, companies that combined ownership and participation grew in employment 10.9% per year more than would have been expected. Sales grew 6% per year more. The New York results used correlations and cannot be compared directly, but the results were in the same direction. In Washington, majority employee-owned firms that were participatively managed did even better.

The Washington study also found that the synergistic effect of ownership and participation was not diminished even when the control group companies had no employee ownership, but had profit sharing and participation programs.

The GAO Study

In 1987, the U.S. General Accounting Office (GAO) did a before and after study using a similar methodology, but covering 110 firms and focusing on productivity and profitability. The measures the GAO used were controversial because they assumed that employee ownership firms did not increase overall compensation when they set up an ESOP. In fact, it appears that about half of all ESOP companies do increase compensation, and few decrease it. The GAO results are probably too conservative because of this assumption.

The GAO study found that ESOPs had no impact on profits, but that participatively managed employee ownership firms increased their productivity growth rate by 52% per year. In other words, if a company's productivity growth rate were 3.0% per year, it would be 4.5% after an ESOP.

The Impact of ESOPs on Employee Compensation

One reaction to the above findings might be that employee ownership companies do better because they substitute stock for wages or benefits. A 1998 study by Peter Kardas and Jim Keogh of the Washington Department of Community, Trade, and Economic Development, and Adria Scharf of the University of Washington, Wealth and Income Consequences of Employee Ownership, shows that, in fact, employees are significantly better compensated in ESOP companies than are employees in comparable non-ESOP companies. Using 1995 employment and wage data from the Washington State Employment Security Department, and 1995 data on retirement benefits from a survey of companies and from federal income tax form 5500, the study matched up 102 ESOP companies with 499 comparison companies in terms of industrial classification and employment size. In terms of wages, the median hourly wage in the ESOP firms was 5% to 12% higher than the median hourly wage in the comparison companies, depending on the wage level of those being compared. The study found the average value of all retirement benefits in ESOP companies was equal to $32,213, with an average value in the comparison companies of about $12,735. Looking only at retirement plan assets other than ESOPs, the ESOP companies had an average value of $7,952, compared to $12,735 for non-ESOP companies. Given that the typical ESOP is actually about 20% invested in diversified assets other than company stock, employees in ESOP companies would have had about as much in diversified assets as employees would have in all assets in non-ESOP companies. In ESOP companies, the average corporate contribution per employee per year was between 9.6% and 10.8% of pay per year, depending on how it is measured. In non-ESOP companies, it was between 2.8% and 3.0%.

In, "Employee Capitalism or Corporate Socialism? Broad-Based Employee Stock Ownership," E. Han Kim of the University of Michigan and Page Ouimet of the University of North Carolina, Chapel Hill (in submission as of this writing), found that ESOPs owning less than 5% of company shares have a small but positive 0.8% effect on total compensation, while in companies in which the ESOP owns more than 5%, total compensation is 5.2% higher. The more leverage associated with the ESOP, the lower the increase in employee compensation, perhaps because companies exercise restraint on total compensation in the face of greater debt. By contrast, the subsample of companies in which the ESOP was established in conjunction with declining sales had lower total compensation (2.8% for small ESOPs and 6.3% for large ESOPs).

Public Companies and ESOPs

In their 2008 study cited above Kim and Ouiment of the University of North Carolina also found that ESOPs have a positive effect on company value. Using Tobin's Q, a ratio of the company's stock value to its book equity value), they found that ESOPs led to an 8.12% increase in company valuation relative to the industry median. Companies with ESOPs with less than 5% ownership showed a valuation increase of 16% relative to the industry median; companies with larger ESOPs showed neither an increase nor a decrease. The impact of company value is positively correlated with greater leverage, perhaps because companies are keeping overall compensation costs more neutral.

To conduct their study, Kim and Oiument studied public companies with ESOPs between 1980 and 2004 for which they could determine an adoption date and for which there was sufficient before and after data to conduct analysis over a significant time period. They found 756 public ESOPs during this time, but had the needed data for 418.

For the first time in ESOP studies, the researchers used the Standard Statistical Establishment List of the Bureau of the Census. The list provides detailed data on all forms of compensation, including ages, benefits, leave policy, severance pay, etc. From this, the researchers computed a measure of total compensation and benefits. These data were then paired with Compustat data on company performance. They analyzed performance for five years before and 10 years after the ESOP was set up, excluding the year of adoption, with a shorter time line for some of the sample companies for a minority of the companies.

In "The ESOP Performance Puzzle in Public Companies," published in the fall 2006 issue of the Journal of Employee Ownership Law and Finance, Robert Stretcher, Steve Henry, and Joseph Kavanaugh looked at 196 publicly traded U.S. ESOP companies during the years 1998 through 2004. Each ESOP company was matched to a comparable non-ESOP company. This is one of the few public company studies to use the more methodologically rigorous matched-pair technique. The ESOP companies had returns on assets that were higher than the matched non-ESOP companies in all seven years, net profit margins that were higher in all of the five years where comparable data were available and better operating cash flows in three of the five years where data were available. The authors present the data for each year, rather than as a single summary measure. Below, we show our own calculated mean of the difference for the years in question:

Return on assets: +5.5%
Net profit margin: +10.3%
Operating cash flow to assets: +.0.1%

On several other measures, the differences achieved statistical significance in less than a majority of the measured years. Keeping in mind, then, that the differences reported may be the result of chance, we can calculate the means of the differences year-by-year for each of these measures:

Return on equity: +5.6%
Sales growth rate: -0.8%
Market-to-book ratio: +0.8%
Debt ratio: -2.9%
Operating cash flow to sales: +0.4%
Operating cash flow per employee: +5.7%

In "When Labor Has a Voice in Corporate Governance" by Olubunmi Faleye, Vikas Mehrotta, and Randall Morck (National Bureau of Economic Research Working Paper, No. 11254, 2005), the authors looked at 225 public companies with 5% or more of the company's stock held by an ESOP. They found that productivity, Tobin's q (a ratio of the market value of a firm's securities to the replacement costs of its tangible assets), long-term investment, operating risk, and growth were all worse in employee ownership companies than non-ESOP public companies. The median Tobin's Q for ESOP companies, for instance, was .0975 and for non-ESOP companies 1.026. The median annual sales growth was 5.7% for ESOP companies compared to 8.7% for non-ESOP companies. Total factor productivity for ESOP companies was -.032 compared to .5 for ESOP companies. The study looked at 100 ESOPs and 115 companies with other plans; plans had to be in place for at least five years prior to 1995. Data were gathered for 1995 through 2001. Generally, the ESOP companies performed less well than the non-ESOP comparison, as detailed in the table below:

The authors argue that "labor voice" causes the underperformance, as workers would vote their shares to favor policies that underspend on investment, research, and development so as to increase current wages. They do not specify just how participants in these plans could possibly do that, given that they would normally only vote for board members and auditors, often do not vote at all, rarely vote with much unanimity and, in any event, are typically voting for a slate of directors that has no meaningful opposition.

There is also an important question of whether the employee ownership plans are overrepresented in underperforming industries. The authors do not control for this variable, although we know that the airline, steel, and banking industries were heavily overrepresented among public company employee ownership plans during the study period. The authors do control for prior economic performance (the five years prior to the employee ownership plan), but in the airline and steel industries, the entire industry suffered severely shortly after these plans were instituted. The plans were created, in both cases, as the industries were still in reasonable financial health, but facing potentially severe future problems, problems that, in fact, led to the demise of many companies in both sectors. In both cases, the plans held ownership stakes much larger than the normal plans.

Nonetheless, the findings appear robust enough to lend support to the argument that just setting up an employee ownership plan does not improve performance and can harm it if not coupled with an ownership culture.

A 1999 study by Hamid Mehran of Northwestern University for Hewitt Associates found that ESOPs in 382 publicly-traded companies increased the return on assets (ROA) 2.7% over what would otherwise have been expected. The study looked at the companies' financial returns for two years prior to the plan's implementation and four years after. Each company was compared to industry norm ROA figures for both periods. Mehran also found that for the 303 ESOP companies surviving the entire four year post-ESOP study period, ROA was 14% higher than the comparison group scores, while for the 382 companies as a group, ROA was 6.9% higher for the four-year period. Over 60% of the companies experienced an increase in their stock price in the two-day period following public announcement of the ESOP, with the average increase for all companies at 1.6%. This suggests that the stock market now reacts positively to ESOPs, a change from the pattern in the 1980s when ESOP announcements were often seen as an indicator that a company was trying to prevent a hostile takeover.

Other studies look at before and after results, with mixed conclusions. Donald Collat, in a 1995 study, found that public companies that did not set up their ESOPs in response to a takeover threat saw their operating margins improve 2.1% per year compared to their pre-ESOP performance. The study looked at companies for three years before and after the ESOP, indexing for market effects. Takeover threat ESOPs, however, saw a decline of 3.3%. In a 1996 study, Mary Ducy, Zahid Igbal, and Aige Akhigbe found that ESOP companies show a decline in operating cash flow of .2% to 2.1% in post-ESOP performance, also using a three years before, three years after measure, and again indexing for market effects. While these are the most thorough of several studies on public company ESOPs, others come to a similarly mixed conclusion.

Finally, a 1998 study by Margaret Blair, Douglas Kruse, and Joseph Blasi found that companies that are publicly traded and at least 20% or more owned by an ESOP are more organizationally stable than comparable non-ESOP companies. Looking at companies between 1983 and 1996, the study found that 74.1% of the ESOP companies remained as independent operations while only 37.8% of the comparison companies did (these figures changed to 59.3% and 51.1% for the period 1983 through 1997, however). None of the ESOP companies went bankrupt, but 25% of the comparison companies did.

These mixed results are probably explained by three factors. First, a 1997 NCEO study found that public companies generally seem to view employee ownership solely as a benefit plan, not part of an explicit organizational culture, as many closely held companies do. Second, ESOPs in public companies tend to own a much smaller percentage of company stock than ESOPs in closely held companies. Finally, in many cases, public company ESOPs simply replaced existing plans where the company contributed company stock to a 401(k) plan. Now the company used an ESOP to make this contribution instead. Hence, the "before" was really not much different from the "after," so not much could be expected to change.

Broadly Granted Stock Options and Corporate Performance

In "Executive and Broad-Based Stock Options: Evidence From U.S. Panel Data," Rutgers University Working Papers (2005). James Sesil and and Yu Peng Li looked at 291 companies with broad-based option and executive option plans for which a start date could be identified for both plans (these dates could be different). They then looked at productivity, growth, and capital intensity changes for three years before and after the plans' start dates.

The table below provides the results for productivity, growth, and capital intensity in terms of the mean changes before and after the plans. Generally, the executive-only plans would have been established well before the broad-based plans.

Mean productivity, growth, and capital intensity changes in three years before and after plan introduction

Variable	Broad-based plan, before adoption	Broad-based plan, after adoption	Executive-only option plan, before adoption	Executive-only plan, after adoption
Output/employee in dollars	$265,575	$318,925	$231,188	$403,071
Employment	19,190	20,446	14,433	30,365
Capital intensity (million $)	$1,497.0	$1,915.0	$1,428.4	$4,563.9

The table indicates that the impact of the two kinds of plans was positive in each case, with more of a productivity impact from executive plans. The researchers found that the impact of broad-based plans occurs primarily immediately after the plan's introduction then levels off over five years. Executive-only plans, however, show continual improvement over the period. The researchers suggest, however, that this may be because executive plans tend to award grants annually or more often, while broad-based plans often make awards either only once or at periods less than annually. They did not assess how frequently this pattern occurs, but anecdotally, the NCEO would estimate that only about half the broad-based plans would provide annual and bi-annual grants during this time frame (they have become more common in recent years). Sesil and Li suggest that the continual reinforcement of regular grants may be critical to the incentive effect.

In "Broad-Based Stock Options Before and After the Market Meltdown" (Rutgers Working Papers, 12006) by James Sesil and Maya Krumova, the authors used the same data set for the other studies in this series to evaluate companies with broad-based options in 1998 and 2000. The study looked at two issues: 1) would stock options be less effective in times of declining share prices than rising prices? and 2) does the effectiveness of broad-based stock options depend on the size of the company (specifically, do employees react more positively in smaller companies where their individual efforts have a more visible effect)?
To analyze this issue, data were analyzed only for companies that provided stock options to 50% or more of their non-management employees and were in business from 1995 through 2002. From this, two datasets were created, one of 463 companies for 1995-1997, a period of rising stock prices, and one of 367 companies for 2000-2002, a period of falling prices. Companies were then classified as small (under 500 employees), medium (500-5,000), or large (over 5,000). Using data from Compustat, the researchers matched the broad options companies with comparable companies in their 2-digit SIC codes. Because the researchers could not know when the plans were established, they used a technique called a random effects estimator to correct for any bias these omitted data might introduce (this is a statistical technique used with panel data in a time series that allows corrections for certain missing data).

In the 1995-1997 period, they found that companies with broad-based options had productivity levels 20% to 33% higher than comparable firms. The smallest companies and largest companies registered at about a 20% differential; medium sized at 33%. In 2000-2002, medium and large-sized companies retained these differentials; the small company differential declined just over than 1%.

The results indicate that the declining stock market did not undermine the impact of broad options. Moreover, contrary to popular perception that the incentive effects of options should be lower in larger companies (because individual employee efforts seem to matter less), company size does not seem to be consistently related to performance.

The largest study on this topic was a 2000 study by Douglas Kruse, Joseph Blasi, and Jim Sesil of Rutgers University, and Maya Krumova of the New York Institute of Technology, using data provided by the NCEO. The study was published in our book Stock Options, Corporate Performance, and Organizational Change (now out of print).

The study sample was drawn from the 1998 NCEO Current Practices in Stock Option Plan Design study. That study sent surveys to 1,360 companies that were identified as possibly having broad-based option plans, which we defined as plans in which more than 50% of full-time employees would actually receive options. We received 141 responses. For the Rutgers study's purposes, 105 companies provided usable data. The authors used a before and after approach to the data to reduce or eliminate sampling bias issues.

Results were based primarily on the 91% of the sample companies that were publicly traded. Data were gathered on productivity, return on assets, Tobins Q (a complex financial measure of return on assets that produced similar results to the return on assets measure and is not reported here), and total shareholder return. These were then compared to all non-broad based stock option companies in their industries of similar size (the full sample group) and to paired comparisons of matched non-broad based stock option companies (the paired sample).

Because few companies had discrete plan start dates early enough to perform a comprehensive before and after analysis, the researchers, as a substitute, analyzed companies in the period 1985-87 and 1995-97, reasoning that few, if any, of the companies had option plans in the earlier period and most had them in the later period. Comparisons were made with non-stock option companies for the two periods and the difference subtracted. In effect, the earlier period results provide a baseline to measure the performance in the later period. If a stock option companies had productivity 3% greater than its peers in the earlier period and 6% greater in the later period, than it could be argued that the plan improved relative performance on this measure by 3%.

The study found that productivity rates did improve with the institution of a plan. The difference between productivity scores from the for the overall sample from the pre-plan period (1985 to 1987) to the post-plan period (1995 to 1997) was 14.8% when the comparison group was all non-option companies and 16.8% when looking just at paired comparisons. Sampling error can be strongly rejected as an explanation for these results.

Return on assets showed a similar pattern. Here the stock option companies showed an improvement of 2.5% on ROA relative to the full sample in the post-plan period compared to the pre-plan period. When just paired comparisons are used, the improvement was 2.05%. Again, sampling error is very unlikely to have caused these results.

Total shareholder return, however, showed no statistically significant difference in the relative performance during the two periods, meaning any measured change could simply reflect random sampling error. The researchers thus believe that the any value consequences of dilution caused by broad-based options seems counterbalanced by increased productivity.

Looking simply at how the companies did in the period 1992 to 1997, without trying to adjust for market effects, a similar pattern emerged. Productivity growth was 1% per year greater and return on assets 5.8% greater, but shareholder return was not statistically distinguishable.

Conclusion

Researchers now agree that "the case is closed" on employee ownership and corporate performance. Findings this consistent are very unusual. We can say with certainty that when ownership and participative management are combined, substantial gains result. Ownership alone and participation alone, however, have, at best, spotty or short-lived results.