The Performance of Workers’ Cooperatives1
Workers’ cooperatives are businesses owned and managed by their employees. Labor-
managed businesses have existed since the 1830s, yet they remain one of the least well-
known parts of the cooperative movement outside the specialized research community.
The image of worker cooperatives has been marred by preconceived ideas that businesses
run by their employees cannot work and must be rare, very small affairs that survive only
in special industries. In this chapter, I present an overview of the key findings of
international economics studies on labor-managed firms’ performance and examine some
of the implications for cooperative practice.
There exist many more worker cooperatives than most people think, even though
the employee-owned firm is not a very common business form. For example, there are
more than 25,000 worker cooperatives in Italy, several thousand in Spain, some 2,000 in
France, and hundreds in many countries around the world. Worker cooperatives are found
in most industries, including very capital-intensive ones as well as services, and
1 In P Battilani and H Schroeter (eds) The Cooperative Business Movement, 1950 to the
Present. New York: Cambridge University Press, 2012, ch. 8, pp195-221.
Special thanks to Alberto Zevi, Lanfranco Senn, and other participants in the Conference on “The
Cooperative Movement 1950–2010 . . . and Beyond” (Bocconi University, Milan, October
2010), as well as to Mónica Gago, who provided insightful comments on an earlier version of
traditional as well as high-technology sectors. Detailed comparative data available for a
few countries also show that worker cooperatives tend to be larger on average than other
firms, the vast majority of which are very small (for example, some 80 % of all firms
with at least one employee have fewer than ten employees in France and in the United
States, but the figure is only 55 % for French worker cooperatives).2 The largest
cooperative group owned by its workers – the Mondragon Cooperative Corporation in the
Spanish Basque Country – employs some 85,000 people around the world. Worker
cooperatives have often operated for considerably longer than a century, and a number of
firms created in the late nineteenth century are still trading today. This descriptive
evidence alone would suggest that workers’ cooperatives are capable of performing
reasonably well in market economies.
Unfavorable preconceptions about worker cooperatives come from the fact that
cooperatives practice a form of economic democracy that many observers regard as
unlikely to work. The concept of labor-managed firms turns on its head a fundamental
feature, the hierarchy, of what most people think of as a firm. For this reason, the
performance of worker cooperatives has been a thorny issue for more than a century
2 The figures for France are for 2007 (see Fakhfakh, et al., 2012); the U.S. figure refers to 2004
(see U.S. Census Bureau at www.census.gov/epcd/www/smallbus.html, accessed on April 24,
2011). Comparative average sizes for Italian worker cooperatives and other firms are presented
in Pencavel, et al. (2006). Burdίn and Dean (2009) show that 64 % of all firms in Uruguay
have fewer than six employees, but only 9 % of worker cooperatives. A likely reason for the
greater size of worker cooperatives is that it takes several people to form a cooperative. Italy,
France, and Uruguay legally require worker cooperatives to have certain minimum sizes,
though in all three cases the required minimum has been decreasing in recent decades. There
exist few very large worker cooperatives, but it is unclearr there is a larger proportion of very
large firms among firms in general than among worker cooperatives.
among economists. Employee-run businesses are a minority form of firm, and many
economists have thought the reason must be that labor-managed firms are less efficient,
and generally less viable, than other firms. As a result, much of economists’ interest has
focused on the comparative efficiency of worker cooperatives relative to that of
conventional capitalist firms. Fortunately, a few studies have looked at factors that make
certain worker cooperatives more successful than others.3
The period we are examining – 1950 to 2010 – happens to correspond to that of
the modern economics literature on labor-managed firms. In that time, a large share of
that literature has been concerned with the performance of workers’ cooperatives, several
aspects of which relate to their efficiency. There are primarily two threads to that body of
research. In the last three decades, a number of empirical papers have examined worker
cooperatives’ performance relative to that of conventional firms, investigating, for
example, whether cooperatives are more or less productive than other businesses.
Another set of studies has focused on cooperatives’ overall ability to survive, that is, to
achieve a measure of institutional sustainability. Several theorists have argued that
perverse incentives built into the labor-managed firm model inevitably lead to the firm’s
demise, whether by underinvestment or degeneration to the capitalist form. These
hypotheses have generated a small body of empirical work examining the survival record
of workers’ cooperatives.
3 Because the profit-maximizing, investor-owned firm is the reference model to which the
cooperative is compared, worker cooperatives are effectively compared to all types of private
for-profit firms in these studies, though it might be interesting to compare worker cooperatives
with more specific groups of conventional businesses that may share some characteristics with
worker cooperatives, such as self-employment (as with family businesses) as well as with other
types of cooperatives.
Perhaps more fundamentally, the theoretical models of the labor-managed firm
that predict underinvestment and degeneration have implications for the choice of
constitutional structure for workers’ cooperatives. In particular, these theories point to the
central role of capital ownership and profit allocation arrangements in determining the
success or failure of this business form and can be discussed in the light of existing
Comparatively little economics research can be found regarding the factors that
make some worker cooperatives more successful than others with the same basic
structure. The studies concerned have focused on the features that make worker
cooperatives different from conventional firms, in order to assess whether the special
features of labor-managed firms hamper or instead enhance their performance. Thus, the
studies look at the proportion of employees with formal rights to participate in decisions
and the share of profit they will receive, for example, and examine their effect on the
firm’s performance. I will cover this aspect of the literature first, along with comparative
empirical studies of labor-managed and conventional firm performance that use the same
methodological approach. Although the underinvestment and degeneration hypotheses
have generated a large theoretical literature and many ad hoc discussions of empirical
observations, much confusion remains regarding these hypotheses, and little rigorous
empirical work on them has been done. I will summarize the theoretical hypotheses and
extend the discussion of these issues to different types of evidence: thus, I will consider
comparisons of the institutional arrangements to be found among workers’ cooperatives
in Italy, Mondragon (Spain), and France (Alzola, et al., 2010) before providing some
evidence on survival and employment.
In order to accommodate evidence from countries that have multiple forms of
firms fully owned by their employees, and to focus my institutional discussion of
ownership and profit allocation arrangements, I will use a broad definition of a worker
cooperative based on cooperative principles. Unless otherwise specified, a worker
cooperative (or a labor-managed firm) in this chapter is a firm owned and managed by its
employees, where the bulk of the capital is owned by employees (whether individually or
collectively), all employees are eligible to apply for membership and a majority are
members, and each member has one vote. Beyond this, the firms we will be looking at
may have tradable or nontradable shares, collectively owned capital, and so on.4 As we
will see, the details of these financial arrangements have important implications for
institutional viability and firm survival.
For conventional firms, performance is commonly measured by financial success
– for example, profit or return on assets. However, in worker cooperatives, pay is
endogenous (Pencavel, 2001) and not analytically distinct from profit,5 whereas in
4 I will not cover the case of Yugoslavia, where firms were socially owned and managed by their
employees. More generally, the chapter will look primarily at industrialized countries, with a
few references to the empirical studies that compare employee-owned firms to other ownership
forms in transition economies.
5 Depending on the tax regimes applying to pay and profit, members of worker cooperatives may
choose to increase pay or to distribute more profit to themselves in a given year (pay increases
do not necessarily have the same permanent character in worker cooperatives as in
conventional firms, as worker members may decide to cut pay in subsequent years – see below,
p. 28). A study of large and medium-sized Italian cooperatives carried out by Centrostudi
Legacoop shows that in manufacturing and construction, their accounting profit would have
been 13.0 % higher in 2007 if pay increases approved by the AGM once operating surplus was
known had been included in the profit appearing on the balance sheet. Among social
cooperatives, the figure was 49.9 % (Centrostudi Legacoop, 2009).
conventional firms pay is a cost bearing negatively on returns. Profitability is therefore
not an appropriate measure of performance for employee-owned firms and is not
comparable across the two ownership categories. Productivity, on the other hand, is a
measure of performance more strictly related to the economic notion of productive
efficiency, can be compared across firm types, and is an appropriate measure to test
theoretical hypotheses that predict that labor-managed firms will be more (or less)
efficient than conventional firms. Here I will look primarily at total factor productivity,
which takes into account both labor and capital inputs. The objectives of cooperators, key
theoretical hypotheses on labor-managed firms, and the policy debates also suggest that a
broader view of performance is appropriate. Firm survival is a measure implied by the
hypotheses that underinvestment and/or degeneration will lead to the demise of all labor-
managed firms. In addition, both investment and job creation or preservation are
especially interesting in today’s recessionary context and relate to externalities to
individual firms’ behavior that are relevant to public policy. As Craig and Pencavel
(1993) have shown, it is likely that employment as well as pay is an objective pursued by
members of a labor-managed firm. I will therefore look at investment and employment in
addition to total factor productivity and firm survival.
For the most part, the empirical studies I will refer to use cross-sectional data (i.e.,
data on many firms but only in one year) or short panels (i.e., data covering the same
firms for several years) because consistent time-series and long panels are rare. This
means that time dynamics can only occasionally be examined. However, the literature I
review has different strengths. It covers a number of countries and types of worker-
managed firms. Several of the studies use large samples of firms in a range of industries
(especially recently) so that the issues can be put in the perspective of the practices of
hundreds of businesses operating in a variety of contexts. In addition, the strong
econometric tradition in this area means that considerable attention has been paid to
controlling for possible confounding factors, reverse causality issues, and so on, so that
the bulk of the evidence is solid.
The section on institutional sustainability looks at studies investigating factors
that increase the productivity of workers’ cooperatives and estimations of the
comparative productivity of labor-managed and conventional firms. The section on job
preservation and survival looks at institutional sustainability issues, including
underinvestment and degeneration to the capitalist form. The evidence regarding
employment and firm survival is presented later, and conclusions are drawn at the end of
A substantial body of literature developed from the late 1970s to test the proposition that
several forms of employee participation that are practiced in workers’ cooperatives had
positive effects on productivity. Although that literature has been dominated by studies of
employee participation in conventional firms, several papers, especially early on,
examined the effects of different levels of worker participation among worker
cooperatives, thus providing us with tests of whether performance is improved or
hampered by different practices in some of the areas that are crucial for worker
cooperatives. Another small group of studies compared the productivity of conventional
and labor-managed firms with the help of data sets including both types of firms.
Factors Increasing Productivity among Worker Cooperatives
The theory behind the hypothesis that employee participation increases productivity is
well known and can be summarized as follows.6 In conditions of asymmetric information,
uncertainty, and incomplete contracts, employee involvement in decision making
improves the quality of information flows and decisions and may contribute to retaining
employees by providing a “voice” alternative to the “exit” option and by internalizing
employees’ interests in decisions. This, in turn, may make it easier to implement
decisions. Participation in decisions also may contribute to fostering intrinsic motivation
(see, e.g., Frey and Jegen, 2005) by increasing employees’ perceptions of being valued
and treated with dignity as well as their sense of autonomy at work.
Participation in the economic returns of the firm, whether by receiving profit-
related bonuses and/or dividends on capital shares (and, where relevant, capital gains),
makes employees’ income (and possibly wealth) dependent on good firm performance.
This is thought to provide incentives to work harder and better, to share information with
management and coworkers, and to invest in human capital and train others. The
collective nature of returns participation in employee-owned firms also may encourage
cooperation and team work. Having participation in both decision making and economic
returns should further increase organizational effectiveness and productivity by providing
incentives to make decisions consistent with firm profitability, and by offering employees
6 See, e.g., Blinder, 1990; Ben-Ner and Jones, 1995; Bonin, et al., 1993; Kruse and Blasi, 1997;
Dow, 2003; and Addison, 2005 for reviews.
opportunities to release relevant information as well as a way to check moral hazard on
the part of managers in decisions that affect employees.7
Against these optimistic hypotheses, it has been argued that the collective nature
of the incentives provided by profit participation promotes free-riding rather than
increased effort, although in the context of a firm, where the game among employees is
normally repeated, a cooperative equilibrium may emerge (FitzRoy and Kraft, 1987).8 It
has also been contended that managers’ incentives are diluted by employee ownership
and profit sharing, and that conflicts, slow and ill-advised decision making, and
coordination problems may beset employee-run firms.
Several early studies tested these hypotheses by estimating production functions
on data from worker cooperatives that practiced different degrees of participation in
decisions, profit, and capital ownership. In particular, a series of studies used three data
sets from the U.K., France, and Italy, respectively (see, in particular, Jones, 1982;
Defourny, Estrin, and Jones, 1985; and Estrin, Jones, and Svejnar, 1987). The British data
set concerned some 150 long-established U.K. worker cooperatives in the printing,
clothing, and footwear industries, observed every five years in 1948–68. The second data
set covered around 550 French worker cooperatives in manufacturing, construction, and
services observed in 1978–79. The third one included annual information on 150 Italian
7 “Moral hazard” refers to cases in which management may make decisions consistent with their
own interests (and/or, in conventional firms, with investors’ interests) but detrimental to other
8 Anecdotal as well as statistical evidence actually point to increased peer pressure in
participatory firms (Kruse, et al., 2004).
manufacturing and construction worker cooperatives observed in 1976–80.9 The general
approach was to augment the production function by inserting variables measuring the
level of each form of participation, so that each of those effects could be estimated while
taking into account the employment and capital levels of the firm as well as its industry
and other relevant controls, such as the age of the firm. The extent of participation in
decision making was measured with the proportion of cooperative members among
employees (or, in some of the estimations for the U.K., the proportion of workers on the
board of directors). The average amount of profit allocated to each worker (or, in the
Italian data set, profit per employee) measured the level of profit sharing, and the average
individually owned capital stake per worker represented the level of participation in
ownership (in all three sets of cooperatives, only limited dividends were paid on capital,
and membership shares were paid back at nominal value when the member left the firm;
the bulk of capital was accumulated in collective ownership). Some of these studies also
controlled for the level of collectively owned capital. Various functional forms for the
production function were tested for. In the later studies in the series, the estimation
methods took into account both the simultaneous determination of the input and output
levels and the possibility that levels of participation were endogenous (so that estimated
effects might be biased by reverse causality if, for example, more productive firms paid
higher profit-sharing bonuses, or if more workers were inclined to join a prosperous
cooperative). These issues were appropriately handled with Instrumental Variables
estimation and firm-specific fixed effects where possible, though the nature of the French
data set limited the availability of good instruments.
9 The whole series of early studies using these data sets is reviewed in Conte and Svejnar (1990)
and in Doucouliagos (1995).
In the three countries, the studies found that increased profit sharing was
associated with higher total factor productivity, though some of the estimated effect still
might have been due to reverse causality, in the French case in particular. Increased
participation in decision making, in the form of a higher proportion of employees being
cooperative members, was found to increase productivity in both France and Italy, but not
in the U.K. cooperatives. However, in the U.K. case, there was some evidence that a
greater proportion of workers on the board was associated with higher productivity,
suggesting once again that greater participation in decision making improves
performance. The level of individually owned capital per worker was found to improve
productivity in the French and Italian cases but not in the British one, where the average
stake per worker was very low. In the British cooperatives, however, higher individual
capital stakes were associated with higher productivity when there were high proportions
of workers on the board and of members among the workforce, suggesting some
complementarity between participation in decisions and in ownership, as suggested by
the theory. When the level of collectively owned capital per employee was included in
the equation, it was found to be unrelated to productivity in the U.K. and French
cooperatives, and negatively associated with productivity in Italy, though this last result
was sensitive to specification. A later study by Estrin and Jones (1995), using the French
data set, explicitly modeled the decision to join in an open membership cooperative and
estimated equations determining the membership rate and individual capital stake in the
cooperative jointly with the production function, in order to remedy the reverse causality
problems that potentially biased earlier estimations of the effects of these two forms of
participation on productivity. The findings of this study confirmed the earlier results,
showing that both increased participation in the governance of the cooperative and
greater capital commitment on the part of members are associated with productivity
increases, independently of reverse causality effects.
The pivotal role of participation in decisions is confirmed by several studies of
North American employee-owned firms, including, in particular, very early studies of the
plywood cooperatives of the U.S. Pacific Northwest (reviewed, e.g., in Conte, 1982, and
in Conte and Svejnar, 1990) and a more recent study of the compared performance of a
90 % employee-owned firm set up as an ESOP (employee stock ownership plan) with
matched conventional firms (Ros, 2003). These studies investigated the possible effects
of participation in decision making (measured by identifying firm practices or by looking
at workers’ perceptions of the extent of their participation collected with employee
surveys) and capital ownership or profit sharing on employees’ attitudes, including
commitment, motivation and job satisfaction, and/or effort in cooperatives.10 Their
findings suggest that participation in decision making is associated with greater employee
commitment, satisfaction, motivation, and effort. In contrast, ownership or profit
participation may have narrower effects and some of these effects could be dependent on
the presence of participation in decision making. These findings seem to imply that, at
least in cooperatives, the hypothesized complementarity between participation in
governance and profit may be verified for the effects of participation in profit or
ownership, which may require participation in governance in order to obtain, but not
necessarily for those of participation in decisions, which may stand alone. This
10 See also Kruse and Blasi (1997) for a review of the evidence on these issues in part-employee-
owned firms from studies in psychology and sociology as well as economics.
interpretation of the evidence is consistent with the findings on the U.K. cooperatives
reviewed previously and is echoed in Pencavel (2001).
The finding that participation in decision making increases efficiency as well as
job satisfaction is confirmed by a more recent study investigating nearly 1,000 firms in
Spain, including about 60 worker cooperatives (Bayo-Moriones, et al., 2003).
Interestingly, however, the favorable estimated effect did not extend to measures of
employee behavior, like absenteeism or industrial action, examined in the study; but the
study did not consider forms of participation that are important to cooperatives, such as
employee representation on the board of directors.
The evidence to date, therefore, is remarkably consistent in showing that the key
feature of worker cooperatives, increased worker participation, never causes performance
to deteriorate in these firms, contrary to many theoretical predictions. Across countries,
firm samples, and methodologies, studies find that greater participation in governance is a
factor of increased productive efficiency in worker cooperatives, both in itself and
perhaps in boosting the incentive effects of participation in ownership and/or profit. More
large-scale studies of these issues with panels of worker cooperatives would be useful in
order to evaluate how solid the very early results are in relation to individually and
collectively owned capital, for example, and to look at the different aspects of
participation in worker cooperatives more systematically.
Participatory practices are not always comparable across firm types. For example,
capital shares often have different characteristics and confer different rights in
cooperatives and conventional firms. Several studies have instead taken advantage of the
availability of comparative data to focus on the compared productivity of worker
cooperatives and conventional firms overall, without looking at the details of
Compared Productivity of Worker Cooperatives and Conventional Firms
The studies of Italian, British, and French cooperatives we have just looked at were
reviewed, along with studies of the productivity effects of participatory practices in
conventional firms, in a meta-analysis by Doucouliagos (1995). His key finding is that
employee participation in decision making and in profit have greater positive productivity
effects in worker cooperatives than in conventional firms. He notes, however, that in the
case of participation in governance, this is probably because there is more participation in
cooperatives, whereas the greater effect found for profit sharing in cooperatives is
independent of the size of the bonus.11 It is possible that the greater productivity effect of
profit sharing in cooperatives is due to the presence of greater participation in decision
making in these firms. In contrast, the modest positive effects of employee ownership are
not found to be statistically different in cooperatives and in other firms. This is consistent
with the findings of Bayo-Moriones and colleagues (2003) that the same level of
participation in decision making had the same effect in worker cooperatives and in
conventional firms, and that the fact of being a cooperative did not affect productivity in
and of itself independently from governance participation. Ros (2003) similarly found
that once the level of participation was controlled for, the firm’s being employee-owned
had no extra effect on employee effort.
11 The meta-analysis pools together results of studies measuring governance and other forms of
participation in different ways, and in which the average level of participation, if it is
measured, varies (see Doucouliagos, 1995).
These results suggest that we should expect cooperatives to be more productive
overall than conventional firms, both because they involve higher levels of participation
in governance and because this feature may make profit sharing, which is also present in
many conventional firms, more effective in raising productivity for cooperatives than for
conventional firms. Even if they work more productively, however, cooperative members
may well decide not to work as hard as they can. If the cooperative’s objectives function
is the utility of the representative or the median member, and members have a normal
income-leisure trade-off,12 output need not be maximized at a given level of employment
and capital, even if employment does not enter the objectives function of the
cooperative.13 This implies that worker cooperatives, even if they are more productively
efficient than conventional firms for the same input levels, may not appear more
productive or even appear less productive if the labor input is measured as the level of
employment (or even as the number of hours worked, if we cannot measure effort).
Five studies have estimated production functions on comparative samples
including both workers’ cooperatives and conventional firms from the United States
(Berman and Berman, 1989, and Craig and Pencavel, 1995), Italy (Estrin, 1991, and
Jones, 2007) and France (Fakhfakh, et al., 2012). Four of these studies used samples of
cooperatives and conventional firms matched by size and industry and/or technology.
Berman and Berman’s data comprised 144 observations on 37 plywood plants in the U.S.
Pacific Northwest, including ten cooperatives, seven former cooperatives, and 19
12 That is, if they dislike effort and like leisure as well as income.
13 This is the model proposed by Jensen and Meckling (1976) for the owner-manager of a
conventional firm, who maximizes her utility but not profit because she prefers to spend some
of the potential profit in getting benefits in kind from the firm.
conventional plants. The plants were observed at five-year intervals during the period
1958–77. Craig and Pencavel used an unbalanced panel of 170 observations on 34
plywood mills in the same region (seven cooperatives, seven unionized conventional
mills, and eight nonunionized ones) observed every two years in 1968–86. Estrin’s Italian
sample included 49 cooperatives and 35 private firms in light manufacturing in Tuscany
and Emilia Romagna, matched by industry and size and observed annually in 1981–85.
Jones’s research included 26 cooperatives and 51 conventional firms in construction in
the same Italian regions, observed annually in 1981–89. The fifth study, on French firms,
used data on two representative samples of conventional firms with 20 employees or
more merged with information on all the worker cooperatives in the same size band. The
resulting data sets were an unbalanced panel of 431 cooperatives and 6,456 conventional
firms in construction, manufacturing, and services (seven industries) with about 19,600
annual observations in 1987–90; and an unbalanced panel of 166 cooperatives and 2,266
conventional firms in four manufacturing industries in 1989–96, with about 15,300
Three of the studies estimated Cobb-Douglas production functions and the other
two (Jones, 2007, and Fakhfakh, et al., 2012) used translog specifications. Most of the
estimations used Instrumental Variables/Random effects in order to take into account the
simultaneous determination of input and output levels. Fakhfakh and colleagues (2012)
also used System Generalized Moments Method (System-GMM) estimation for some
specifications, which provides the most robust treatment of endogeneity issues (e.g., if
cooperatives exist or survive mostly in subindustries that are most favorable to
When worker cooperatives and conventional firms were constrained to have the
same production function, four of the studies (Berman and Berman, 1989; Estrin, 1991;
Craig and Pencavel, 1995; and Fakhfakh, et al., 2012) found no significant difference in
total factor productivity between the two groups of firms, although Fakhfakh and
colleagues found some evidence that cooperatives may be more productive in certain
industries. The fifth study (Jones, 2007) found differences that were not consistent across
specifications and estimation methods. However, cooperatives and conventional firms do
not have the same production function. Four of the studies tested for this and found the
functions estimated for the two groups of firms to be statistically significantly different;
and Jones (2007) found significant firm-specific fixed effects that may reflect
technological differences between the groups. As Estrin (1991) put it, the two types of
firms organize production differently. This is consistent with the hypothesis that the
effects of the different types of participation practiced in cooperatives are reflected not
only in greater output at all input levels with the same factor elasticities14 (disembodied
effects) but also in different elasticities (embodied effects).15
In order to find out whether one group is more productive, Craig and Pencavel
(1995) computed the output predicted by the functions estimated for cooperatives and for
other firms at the mean point of each group’s sample. They found that the predicted
output was higher at both mean sample points with the function estimated for
14 Factor elasticity refers to the percentage change in output associated with a 1 % increase in one
of the inputs (capital or labor) only.
15 Fakhfakh, et al. (2012) use the properties of the translog production function, in which factor
elasticities and marginal products vary with input levels, to show that differences in estimated
elasticities between the two groups are not simply explained by differences in factor demands,
and can therefore be attributed to the effects of participation on the production function.
cooperatives. Fakhfakh and colleagues (2012) computed, for each sample firm, the output
levels that would be predicted with the parameters estimated for the cooperatives and
with the parameters estimated for the conventional firms, and tested whether the two
levels were the same on average for each sample and in each industry. In all industries
and with both data sets, the output predicted for worker cooperatives with their current
inputs was the same with both sets of parameters or higher with the parameters estimated
for cooperatives. However, with both data sets, there were several industries in which, as
in Pencavel and Craig, the predicted output for conventional firms was higher with the
parameters estimated for the cooperative sample. In other words, if conventional firms
organized production in the same way as the cooperatives, they might produce more with
their current average input levels in these industries.
Few economists expected the explosion of employee ownership that marked the
transition to market systems in former centrally planned economies in the 1990s. In many
transition countries, mass privatization programs resulted in an unprecedented incidence
of employee ownership, with many firms in which nonmanagerial employees owned the
majority of capital (Earle and Estrin, 1998). However, it has been shown that majority
employee ownership in the transition tended not to be associated with corresponding
levels of employee control, or participation in governance, and that as a rule control
remained in the hands of managers (see Jones, 2004). In addition, employees’ shares may
not have been very liquid and profitability may have been low or nonexistent. This
pattern may explain the findings of research on employee ownership in these economies.
Employee-owned firms have been included in studies examining the effect of
privatization on total factor productivity. Endogeneity issues are crucial in this area, since
privatization may have targeted better- or worse-performing firms in the first place. In
their review of the empirical literature on the effects of privatization, Estrin and
colleagues (2009) identify seven studies that estimate the effects of employee ownership,
among other ownership forms, on total factor productivity and handle endogeneity
robustly. Of these studies, six found employee ownership to have no statistically
significant effect on total factor productivity in Central and Eastern Europe and
Confederation of Independent States (CIS) countries, and one study on Estonia found a
Altogether, the evidence is again remarkably consistent across countries, types of
labor-managed firms, economic circumstances and time periods, and with
methodologically robust studies. Worker cooperatives are never found to be less
productive than conventional firms17 and may be more productive. The key factor
explaining this productivity seems to be members’ involvement in governance, which
boosts productivity in and of itself as well as by improving the incentive effects of
participation in the economic returns of the firm. Consistently equal or greater total factor
productivity is a key element for competitiveness.
16 Although it is hardly a triumph for worker ownership, this evidence is strikingly at odds with
the predictions regarding the effects of employee ownership in the transition literature. As
Estrin, et al. (2009) remark, their findings are more favorable to employee ownership than
reviews that included studies in which endogeneity was not adequately taken into account in
the econometric analysis.
17 Megginson and Netter (2002) review studies that find employee ownership to have negative
effects on performance, but the group of papers they review includes studies that do not deal
appropriately with endogeneity (Estrin, et al., 2009) so that these effects probably pick up
lower prior performance among firms that were privatized with employee ownership.
One of the recurring questions regarding worker cooperatives has been why the firm type
is so rare in market economies, especially if it is at least as productive as conventional
firms, as the evidence suggests. For a long time, the widely accepted answer was that
labor-managed firms are not institutionally viable – it was argued that there are incentives
built into the structure of the organization that make it unsustainable. Two models, in
particular, have dominated the debate: the underinvestment/self-extinction hypothesis put
forward, with variants, by Furubotn and Pejovich (e.g., 1970) and Vanek (1977), and
degeneration to the capitalist form, which was analyzed by Ben-Ner (1984) and Miyazaki
(1984). Both hypotheses were put forward to explain phenomena that had been observed
among labor-managed firms, and both provided key insights into the crucial importance
of capital ownership and profit allocation arrangements to the institutional viability of
workers’ cooperatives. As we will see, existing types of labor-managed firms have
provided different solutions to both potential problems. It is, therefore, of interest to
summarize each hypothesis and examine its implications in the light of cooperative
practice and evidence.
This hypothesis generated much discussion and critique until the mid-1990s (see Uvalić,
1992, and Dow, 2003, for reviews). At the heart of the hypothesis is the truncation of
property rights associated with collective capital ownership. When members of a
cooperative with collectively owned capital leave the firm, they cannot receive a share of
the present value of future profits generated by investment their work has helped finance,
as owners of capitalist firms can by selling their shares. If it relies on internal finance, an
income-maximizing, labor-managed firm where capital is owned collectively will,
therefore, have an incentive not to invest, or to invest only in projects with inefficiently
high short-term returns. The firm may even consume the collectively owned capital
instead, and “self-extinguish,” bringing its scale down to one member if its technology is
characterized by constant returns to scale, or to an inefficiently small scale (i.e., with
increasing returns) otherwise. This is a much simplified presentation of the hypothesis,
and a number of assumptions that are necessary for the model to work (e.g., regarding the
lifetime of capital equipment, the opportunity cost of capital, etc.) have been thoroughly
questioned. Descriptive evidence suggests worker cooperatives with collectively owned
capital assess investment projects with similar time horizons as conventional firms
(Bartlett, et al., 1992; Robinson and Wilson, 1993). However, the insights of the theory
are, first, that it may be tempting for cooperative members not to increase capital that will
go to future generations of workers and, second, that far from accumulating capital,
members who have access to capital accumulated by previous generations may instead be
tempted to demutualize and appropriate that capital if they are allowed to do so. Indeed,
the demutualization of cooperatives of other kinds, where demutualization was authorized
by legislation, such as that of many British building societies, suggests that the temptation
can be real.
The solution favored by many economists is to have a market for membership
rights, so that shares are tradable, and if that market is reasonably efficient members can
receive a share of the present value of future returns when they leave the cooperative.18
18 As Estrin and Jones (1992) note, however, many economists in the past also have argued in
favor of collective capital ownership, which may in particular strengthen cooperative
Such a market may be difficult to organize, since, as Putterman (1984), Dow (2003), and
others have argued, membership rights in the case of a labor-managed firm are tied with
particular skills, and so on. In practice, many employee-owned firms that resulted from
privatization were organized in a way that provided tradable rights to the returns to
capital and control, though not exactly membership in the “bundled” cooperative sense,
where both capital ownership and membership of a one-member,-one-vote firm were
merged in membership shares. In such firms, the stock was purchased by the employees
and its value depended on the valuation of the company. Examples include many
employee-owned firms in transition countries and the worker-owned bus companies that
resulted from privatization in the 1990s in the U.K. A common pattern among such
companies is that after a few years, especially if the firm is successful, worker members
sell the company to a conventional owner. This process was observed even among the
American plywood cooperatives of the Pacific Northwest, where there was a limited
market for membership shares, in that shares were advertised in local newspapers but
membership was subject to acceptance by the existing members and seemed underpriced
It is, of course, debatable whether this kind of institutional instability represents a
problem – after all, if cooperators are successful entrepreneurs and can retire comfortably
thanks to the success of the cooperative, there is nothing wrong there. It is, however, a
potential issue for the movement as a whole, and also for public authorities if the
cooperatives have received any tax concessions or subsidies. The key element here is the
advantages, such as commitment to certain values, and in this way decrease the risk of
value of the share, which is both the solution to the investment incentive and the source
of the incentive to sell off. Shares with a high value also require setting up arrangements
for prospective members to pay for their shares by installments, in order to preserve an
open membership, and for the payment of the shares of departing members over a period
of time, in order to avoid potential decapitalization problems when a whole generation
retires (Berman, 1982). Protection against selling out may be afforded by systems in
which employees’ shares are held in trust, as in some U.S. ESOPs, and generally by
systems that allow the capital of members that sell shares and/or leave the firm to remain
in the firm.
An alternative solution to the underinvestment issue, which was proposed by
Vanek (1977), is one that actually has been in operation at least since the Second World
War in worker cooperatives in France and Italy and also was adopted in the Mondragon
system (Alzola, et al., 2010) – mandatory collective capital accumulation. In these
cooperatives, part of the capital is owned individually, but attracts limited returns and, in
Italy and in France, the individual membership shares are paid back at their nominal
values. Another part of the capital (often the bulk of the capital) is owned collectively and
may not be split among the members of the cooperative, even if the firm is wound up – in
that case, the net assets devolve to another cooperative, a cooperative institution, or a
charity (this provision is sometimes called an “asset lock”). This setup ensures
institutional stability19 but creates potential underinvestment incentives. However, in all
19 It has been said that some of the old-style British cooperatives, in which collective assets could
not be split except in the case of dissolution, were wound up for the purpose of appropriating
accumulated capital. However, the figures presented by Jones (1982) suggest that the demise of
three cases, the law or, in Mondragon’s case, the cooperative group’s own articles of
association specify that a minimum percentage of profit has to be plowed back into the
firm every year, and adds to the portion of capital that is collectively owned (Alzola, et
al., 2010). The little rigorous empirical evidence there is on this issue suggests there is no
underinvestment in these cooperatives. Estrin and Jones (1998) estimated an investment
equation on a balanced panel of 270 French worker cooperatives observed in 1970–79.
The equation is estimated robustly by GMM in first differences in order to deal with
potential serial correlation and heteroskedasticity and includes time fixed effects. These
authors found that the share of collective capital in the firm’s assets had no effect on
investment, but that investment might be financially constrained by the limitations on
that part of the U.K. cooperative movement may have been due at least as much to the absence
of new cooperative creations.
It is often remarked that the institutional stability provided by a full asset lock relies on a
forced sacrifice on the part of those cooperators that leave the firm. However, it also allows
new cooperators to enjoy the use of capital accumulated by previous generations – in this
conception, the cooperative is a kind of public good to be used by successive generations of
employees. In itself, this system may create other incentive problems, in that members of
cooperatives that have very large reserves accumulated by earlier generations may be tempted
by the complacency of rentier behavior (as Zevi put it) expecting the money to work for them.
In the Mondragon system, incoming members pay a nonrefundable fee toward the collectively
owned capital (Alzola, et al., 2010). This may ensure greater commitment on the part of
members, as may other systems proposed in Zevi (2005) to keep stable resources in the firm
while offering members appropriate incentives. More generally, Conte and Ye (1995) suggest
that intergenerational financial arrangements of the kind already practiced by Mondragon can
resolve underinvestment issues. Mondragon, the Italian, and the French worker cooperatives
also all provide for individually owned capital accumulation over the years an individual
member is employed in the firm (Alzola, et al., ibid). In any case, a full asset lock need not
prohibit the firm from leaving the cooperative form – this could be allowed, for example,
provided the owners of the new firm buy back the collectively owned capital.
access to equity finance (individually owned capital). Gago and colleagues (2008)
estimated investment functions by GMM on a 16-year unbalanced panel (1989–2004)
comprising some 190,000 observations on conventional firms and 1,900 on worker
cooperatives in French manufacturing. Their preliminary findings indicated that French
worker cooperatives did not invest less than conventional firms, all else being equal, nor
were the cooperatives more financially constrained than other firms. These findings are
consistent with the investment equation estimated by Pencavel and colleagues (2006) on
a long panel (1982–94) of some 2,000 worker cooperatives and 150,000 conventional
firms in Italy, which indicates that investment in cooperatives and conventional firms
reacts in the same way to product market shocks.
Although French worker cooperatives do not underinvest, it is unclear that this is
entirely due to the existence of a mandatory plow-back rule. Navarra (2009) finds that the
60 worker cooperatives in the Italian province of Ravenna on which she has annual data
for the period 2000–05 (in addition to interview and employee survey data for one third
of the cooperatives in 2007), systematically plow back a considerably larger share of
profit than the required minimum. This is consistent with figures presented in Alzola et
al. (2010) that indicate that the bulk of worker cooperatives’ profit is plowed back in
Italy, in part due to Italian regulation and requirements to benefit from tax concessions.
Anecdotal evidence suggests that profit plow-back is a little lower in French
cooperatives, but still considerably higher than the legally required minimum. Fakhfakh
and colleagues (2012) compared the mean annual rate of growth of fixed assets (i.e.,
investment) in worker cooperatives and conventional firms, using their two data sets (one
covering 1987–90 and seven industries in manufacturing, construction, and services, and
the other covering 1989–96 and manufacturing) with information on some 7,000 French
firms, about 500 of which were worker cooperatives. With both data sets, they found that
annual investment was always at least as large in the cooperatives as in conventional
firms (in three out of seven industries in the first data set, cooperative investment was
statistically significantly higher than in conventional firms, and there was no difference in
any of the other industries studied). They also found no evidence that the cooperatives
systematically produced with increasing returns to scale nor that they produced at a
smaller scale than conventional firms.20 Zevi (2005) argues that, far from having an
inefficiently short time horizon, the members of worker cooperatives are first concerned
with job security at a decent level of pay and thus plow-back profit in order to ensure the
growth that will guarantee them jobs. This concern for long-term job stability, and in
some cases, for the continued existence of the firm, may actually give cooperative
members a longer time horizon than many conventional firms where managers are
subject to short-term capital market pressures.
Degeneration to the Capitalist Form
Among the critiques of the underinvestment model, it was pointed out that a worker
cooperative that employed nonmembers who did not share in profits would not
underinvest (see discussion in Stephen, 1982). However, the use of hired employees has
been at the center of another issue of institutional sustainability – degeneration to the
20 There is a larger proportion of medium-sized and large firms, and a lower proportion of small
and very small ones, among workers’ cooperatives than among conventional firms in France
(Fakhfakh, et al., 2012) as in Italy (Pencavel, et al., 2006) and Uruguay (Burdίn and Dean,
2009), a phenomenon already reported by Ben-Ner (1988) for France, the U.K., and Sweden in
the early 1980s.
capitalist form. The process modeled by Ben-Ner (1984) functions roughly as follows.21
If an income-maximizing labor-managed firm is allowed to hire nonmember employees
who do not get a share of the firm’s profit, members will have an incentive to replace
retiring and resigning members by nonmembers. A nonmember employee will produce
the same marginal revenue product as a member would but will only be paid a wage,
leaving more profit to share among the remaining members. Little by little, the
cooperative will have a lower and lower proportion of members and an increasing
proportion of nonmembers among its labor force. It will eventually become a
conventional capitalist firm in which a minority of members exploit the majority of the
workforce.22 Ben-Ner notes that this process may not operate or may even be reversed if
members are more productive than nonmembers because they participate in profit and
decisions. Pencavel (2012) also notes that degeneration may not happen if nonmembers
perform work that is distinct from members and remains necessary because of a
complementarity between the two types of work.
The degeneration models were inspired by empirical observation, and clear
evidence of it is presented, for example, by Russell (1995) for Israeli worker cooperatives
and by Jones (1982) for early U.S cooperatives. Pencavel (2001, 2012) discusses some
recent U.S. cases. Both profit sharing among members and shares that attract dividends or
can be sold at a higher price than they were bought for provide incentives for
21 Miyazaki (1984) proposes a different framework to explain degeneration, which applies only to
systems in which unemployed members of the cooperative remain members.
22 This is a very limited definition, for the purposes of this chapter. There is a more qualitative
process of degeneration that has to do with democracy among members, which I am not
degeneration (Estrin and Jones, 1992). If capital shares become expensive, it will also
become easier to exclude new members (Russell, 1995). The proportion of
nonmanagerial workers owning shares in firms that were employee-owned immediately
after privatization dramatically decreased in the majority of employee-owned firms in
transition countries (Jones, 2004). Kalmi (2004) shows that in the case of Estonia, this
reduction in the proportion of employee-owners was achieved by means of a
degeneration process of the type analyzed by Ben-Ner, where managers excluded new
employees from share ownership. The obvious solution is to prohibit the habitual hiring
of nonmember employees; however, a strict policy of this type may be too rigid. Most
cooperatives in France, for example, have new employees go through a probation period
of, say, six months before they get a permanent contract and apply for membership.
Reviewing arrangements in some U.S. forestry cooperatives, Pencavel (2001) also noted
the imbalances potentially introduced in voting patterns for decisions involving short-
term versus long-term trade-offs by admitting to full membership rights employees who
have just joined the firm and not fully paid their membership fee. In addition, hiring
nonmembers may make it possible to pay employees with special skills substantially
more than members.
The practices of French, Mondragon, and Italian worker cooperatives, all of
which allow employment of nonmembers, provide different solutions to the problem
(Alzola, et al., 2010) even though they may not have been adopted for this reason. French
worker cooperatives, by law, split the share of profit allocated to labor (as opposed to
mandatory allocations to collective capital and the share of profit paid out as dividends on
individually owned capital shares) among members and nonmember employees on the
same terms (CG-SCOP, 2003). These terms, in keeping with cooperative principles,
typically consist in a profit bonus proportional to the individual worker’s pay or hours
worked in the cooperative (as with the “cooperative divi” or patronage-based payment).
Together with the limited dividends paid on capital shares that do not appreciate in value,
this setup eliminates the incentives for degeneration identified by Ben-Ner and Miyazaki.
Estrin and Jones’s (1992) findings confirm that French worker cooperatives do not
exhibit degeneration, even though the percentage of members among the workforce
varies over the life of a cooperative.23 Although Italian worker cooperatives can in
principle pay a share of profit to members only (again as patronage payments, as opposed
to dividends paid on capital shares) legal caps on the amount that can be paid to
individual workers in this way and tax concessions attached to plow-back mean that
cooperatives seem to have policies of plowing back most profits and/or offering profit
sharing to members and nonmembers alike (Zevi, 1982; Alzola, et al., 2010). These
policies are also consistent with a central concern for growth perceived as a way to ensure
job security – an objective that would unite members and nonmembers (Zevi, 2005). The
solution adopted by the Mondragon group is to limit the percentage of nonmembers
allowed in the workforce to a predefined maximum (Alzola, et al., 2010). The maximum
percentage of hired employees allowed for worker cooperatives to enjoy tax benefits in
Uruguay seems to have had a similar effect, and Burdίn and Dean (2009) do not find
evidence of degeneration when they estimate a membership rate equation for all the
23 That percentage is likely to drop, in particular, during periods of growth, before new employees
become members. The French setup, however, gives employees few incentives to become
members, and many cooperatives have resorted to adopting a rule that requires all employees
to apply for membership after a certain time with the firm.
worker cooperatives of that country observed in 1996–2005. Mondragon also has recently
created a category of “temporary members” who have temporary membership rights and
duties, in order to be able not to offer new members the absolute job security normally
attached to membership during the recession, without degenerating.
Judging from the rules in force in many worker cooperatives, it therefore seems
that under certain sets of rules used by the most successful Western European
cooperatives, labor-managed firms are unlikely to disappear by underinvesting and can
avoid degeneration. The little available evidence on cooperative survival will be
presented in the next section. Increasingly, available evidence also points to worker
cooperatives’ concern with employment stability, a point that might explain the pattern of
capital accumulation over and above the legal minimum (Zevi, 2005). In addition to her
evidence on capital accumulation, Navarra (2009) also presents evidence in support of
her argument that accumulating collectively owned capital is a form of collective
insurance. Accumulating collective resources in this way allows cooperatives that
consider employment stability a priority to provide more stable income to their members.
Evidence on worker cooperatives’ pay and employment adjustments to the business cycle
will also be presented in the next section.
Job Preservation and Survival
Pay and Employment Adjustments
Part of the reason why the economics literature on labor-managed firms focused so much
effort on institutional stability for several decades comes from the model of the income-
maximizing Illyrian firm. In this model, the need to maximize income per member leads
to the well-known “perverse supply response,” in which output price increases lead the
cooperative to cut employment (see discussion in, e.g., Bonin and Putterman, 1987). The
perverse supply response disappears if employment is included in worker cooperatives’
objectives, whether as one of the arguments in a utility function or in the form of a labor
supply constraint. As Craig and Pencavel (1993) show with data on the U.S. plywood
cooperatives, the cooperatives behave as if both income and employment are relevant to
their objectives. This is confirmed by Burdin and Dean’s (2010) work on worker
cooperatives from Uruguay.
Labor-managed firms are able to adjust pay in downturns in order to preserve
employment, because the same people – the members – will decide the allocation of
future profits, so that a commitment to increase pay later when market conditions
improve, which would not be credible coming from a conventional employer, is incentive
compatible in a cooperative. Similarly, worker cooperatives can increase pay in upturns
knowing they may decide to cut pay again should market conditions worsen. Using their
data set on U.S. plywood cooperatives and conventional firms, Craig and Pencavel (1992)
show that employment and hours worked in the cooperatives are uncorrelated with
movements in output prices, while there is an almost unit elasticity of pay with respect to
the output price. Conventional firms do the opposite. Cooperatives also increase
production in response to an increase in output price, though by less than conventional
mills. In other words, the worker cooperative does not respond perversely and adjusts pay
rather than employment in response to changing market conditions. Cooperative members
bear financial risk rather than employment risk. Pencavel and colleagues (2006)
estimated wage and employment (and capital – see p. 22 above) equations by fixed
effects and Instrumental Variables (in first differences) respectively using a matched
employer-worker panel data set covering 13 years of information on some 2,000 workers’
cooperatives, 150,000 conventional workplaces, and about 13,000 individual workers per
year in Italy. The worker cooperatives are found to adjust pay rather than employment to
demand shocks, whereas conventional firms adjust employment both in response to wage
changes and to demand shocks. In Italy, as in the U.S. plywood cooperatives,
employment is more stable in the cooperatives. Similar results were obtained by Burdίn
and Dean (2009) with monthly data on the entire population of firms in Uruguay in 1996–
2005, on which they estimated pay and employment equations by IV (in first differences)
and fixed effects respectively. Worker cooperatives were found to adjust members’ pay
more than conventional firms in response to output price changes (though not
nonmembers’ pay) but not employment (whether for members or nonmembers, which
suggests the bulk of the risk is borne by members’ pay). Conventional firms were found
to cut employment in response to a rise in pay, whereas members’ pay and employment
move in the same direction for cooperatives, and cooperatives adjusted employment less
and more slowly to recession.
The available evidence, which is quite robust, is therefore once again remarkably
consistent. It indicates that worker cooperatives adjust pay (at least for members) rather
than employment to changing market conditions and generally preserve jobs better.
Navarra (2009) suggests the need to ensure against market downturns motivates the high
rate of profit plow-back. Accumulated collective capital will thus be drawn on by the
cooperatives to weather unfavorable market conditions, in order to avoid the pay cuts that
might otherwise be necessary to preserve jobs. These findings are consistent with earlier
evidence from Italy and Spain (Bartlett, 1994). Descriptive evidence on Italy in the 1970s
(Zevi, 1982) and France in the 1980s and 1990s (Fakhfakh, et al., 2012) also shows
worker cooperatives preserving or even creating jobs in years in which conventional
firms in the same industries cut jobs. In transition countries, the studies reviewed by
Estrin and colleagues (2009) find that employee ownership has no effect on employment
(contrary to predictions that employee owners were going to keep too high levels of
employment), but these studies also find that privatization was generally associated with
increases in employment rather than cuts as predicted.
It would be interesting to examine separately what happens to members and
nonmembers when market conditions deteriorate. Mondragon offers complete job
security to members (except for the recently created category of temporary members)
who are redeployed in other cooperatives of the group if necessary. At various times in
the group’s history, pay has been cut in order to preserve jobs. Nonmembers have been
massively laid off in the recession that followed the financial crisis of 2008, when
Mondragon cut 10,000 jobs, but members’ pay also has been cut (e.g., in Eroski for two
years in a row). Other workers’ cooperatives may also have cut nonmembers’ jobs. Media
reports on the John Lewis Partnership’s response to the recession in late 2009 and early
2010suggested that members’ jobs may be cut.24 Both Mondragon and John Lewis are
very large employee-owned organizations, in which monitoring managers may
sometimes be difficult for members. In addition, John Lewis is co-managed by senior
management and other employees, so that managers have greater statutory power over
24 See, e.g., reports in The Guardian by Julia Finch in September-October 2009 (available on
http://www.guardian.co.uk/business/2009/sep/30/john-lewis-call-centre-jobs, accessed on April 27, 2012)
and by Julia Finch and Zoe Wood in March 2010 (available on
http://www.guardian.co.uk/business/2010/mar/11/john-lewis-staff-share-bonus, accessed on April 27,
governance than in classic workers’ cooperatives. An interesting question for research
would be whether there is an order of priority between absorbing possible losses with
collectively owned capital, cutting nonmembers’ jobs, and cutting members’ pay and
members’ jobs – and if so, whether that order, and employment policy generally, is
related to the governance of the cooperative.
Worker cooperatives in the U.K., France, and Italy have often exhibited considerable
longevity, with a number of firms surviving for more than a century (Jones, 1982; Estrin
and Jones, 1992; Pérotin, 2004). However, little comparable evidence exists regarding
failure rates, and as yet very few econometric studies have looked at the conditions under
which worker cooperatives survive or die. In France, annual death rates averaged 10 %
for worker cooperatives and 11 % for conventional firms in 1979–2002 (Pérotin, 2006)
but were 11 % for both groups of firms over the 1979–98 period. Ben-Ner (1988) shows
death rates of one-third less for worker cooperatives in France (6.9 % in 1976–83) and
the U.K. (6.3 % in 1976–81) than among conventional firms (10.0 % in 1980–83 in
France and 10.5 % in 1974–82 in the U.K.). Overall, the patchy evidence reviewed by
Dow (2003) suggests that labor-managed firms probably survive better than conventional
firms. The only comparative estimate of a causal hazard function model for worker
cooperatives and conventional firms is Burdίn’s (2010), which confirms this with a data
set comprising 22,315 firms including 243 worker cooperatives observed from April
1996 to December 2005 in Uruguay. His Cox proportional hazard estimates show the
cooperatives to have lower hazard rates, all else being equal, than conventional firms. In
addition, he finds the cooperatives have lower hazard rates than conventional firms
specifically in industries where rates of worker supervision and of labor turnover are
high, suggesting labor-managed firms have a specific advantage in these industries, but
relatively higher hazards in industries with higher inequality (which may reflect greater
skills heterogeneity). No difference between the hazards of cooperatives and conventional
firms is found in industries with high investment rates.
The riskiest years in a worker cooperative’s life seem to be the early years, as
with conventional firms. However, hazard functions estimates for Israel, Atlantic Canada,
and France suggest that the riskiest year for worker cooperatives may not be the first, as
with conventional firms, but may happen later, after two to five years, so that
cooperatives are characterized by a “liability of adolescence” (Staber, 1993; Russell,
1995; Pérotin, 2004).25 The mean survival hazard estimated on the basis of a Cox
proportional hazard model was 9.2 years for Israeli cooperatives but 18 for those of
Atlantic Canada. Pérotin (2004) constructs nonparametric hazard curves for the 2,740
worker cooperatives created in France in 1977–93, 1,660 of which closed down during
the period, and finds that in the first eight years or so of a firm’s life, cooperatives created
from scratch have the highest hazard rates, followed by rescue employee takeovers of
failing firms, followed by cooperatives formed by an employee buy-out of a sound
conventional firm. However, in the few years that follow, the order is reversed, with the
highest hazards found among conversions of sound firms, followed by rescues, and last
by cooperatives created from scratch. The origin of the firm may therefore affect the
25 This pattern could result from financing problems experienced by young cooperatives relying
entirely on profit plow-backs at a time when growth is needed.
timing of the failure risk at least as much as its level. Studies on other countries and with
data covering longer time periods may or may not confirm this in the future.
Following the sociological and economics literatures, Burdίn, (2010), Staber
(1989), and Russell (1995) all focus on external factors such as the dynamics of
organizational demography and the business cycle in explaining hazards, and Pérotin
(2006) on these factors to explain death rates. Both Staber (1989) and Pérotin (2006) find
that the number of existing worker cooperatives affects their failure risk, though the
findings of the two studies are not comparable and suggest effects in opposite directions.
Finally, Pérotin (2006) estimates equations explaining the annual number of firms closing
down in France for conventional firms and workers’ cooperatives in 1981–2002 and finds
the two equations are not statistically different. In particular, deaths among both types of
firms respond in the same way to the business cycle, which suggests that fears that
worker cooperatives disappear when market conditions are good (Ben-Ner, 1988) are
unfounded. It is unfortunate that there is very little research to date on the relationship
between individual cooperatives’ characteristics, such as their start-up size or capital
intensity or growth rates, and survival, which might tell us whether, for example, the
widespread notion that worker cooperatives’ difficulty in accessing external finance is a
serious liability is verified over time.
This overview of the empirical evidence on the performance of worker cooperatives
suggests both that worker cooperatives perform well in comparison with conventional
firms, and that the features that make them special – worker participation and unusual
arrangements for the ownership of capital – are part of their strength. Contrary to popular
thinking and to the pessimistic predictions of some theorists, solid, consistent evidence
across countries, systems, and time periods shows that worker cooperatives are at least as
productive as conventional firms, and more productive in some areas. The more
participatory cooperatives are, the more productive they tend to be. The temptation to
consume capital accumulated by previous generations, demutualize, sell out successful
cooperatives to conventional owners, or degenerate by restricting membership (about
which the theoretical literature has had such useful insights), all have solutions that were
adopted by different types of worker cooperatives around the world, assisted by
legislation. That legislation has not protected workers’ cooperatives, but rather enabled
them to avoid perverse incentives (just as legislation protects minority shareholders’
rights in conventional firms, for example). And the little we know about the survival
record suggests that these solutions work.
Among the possible solutions are measures like asset locks and collective
accumulation of capital that have been looked at with suspicion by generations of
economists. Such measures do not seem to hamper productivity by dampening incentives
– some of the same cooperatives that have adopted these particular measures are found to
be more productive (as the French cooperatives) or to preserve jobs better (as the Italian
cooperatives) than conventional firms. This, to me, seems to imply that we have given too
much importance, in this literature, to issues of income over issues of job security and,
more broadly, empowerment in worker cooperatives. Employment in a labor-managed
firm is not the same thing as employment in a conventional one. In a labor-managed firm,
members participate in the decisions that affect their unemployment and income risks.
They are considerably better protected against the moral hazard potentially attached to
management decisions over investment, strategy, or even human resource policies. This
may explain why participation in governance is so important to the performance of
workers’ cooperatives (though these results have to be updated) rather than the monetary
incentives we have focused on for so long. It is also a fact that workers’ participation in
profit and in decisions makes it possible for worker cooperatives to adjust pay rather than
employment in response to demand shocks.
In this sense, there is no trade-off between cooperative principles and economic,
or indeed social, performance, though not necessarily in the naive sense of a “win-win”
business case for participation – profit may not be higher in more participatory
cooperatives, but the firms may produce more and preserve their members’ jobs better.
One of the things that has become apparent in the course of this overview is that we have
very little empirical economics work looking at what makes certain cooperatives more
successful than others with the same structure.
The recent empirical literature has focused, correctly, on establishing comparative
results that systematically put the cooperatives in the context of all other firms, and a lot
more of this type of research remains to be done, as large representative data sets have
only recently become available. For example, worker cooperatives may need to be
compared with more specific segments of the firm population and other types of
cooperatives. However, now that a lot of the groundwork has been done, we also need to
compare worker cooperatives among themselves again, to look at those cooperative-
specific features and to investigate those differences that may tell us more about the way
forward. We need to know how cooperative specificities relate to success. One area about
which we know little as yet is that of cooperative expansion, subsidiaries, and external
growth. A lot has been happening in this area, which has raised important issues, for
example, in Mondragon when the membership decided to bring into the cooperative fold
noncooperative subsidiaries that had been acquired by external growth. Cooperative
expansion, whether by creating new firms or subsidiaries or by external growth, has long
been identified as an issue that is potentially more difficult and more important to tackle
than cooperative survival (Pérotin, 2006). In this respect, numbers alone should make it
clear that the Italian case has a lot to teach to other countries. Comparative research
investigating different types of growth and examining the role of specialized support
structures like the Italian consorzi may help us find out in particular whether cooperative
specificities can help to handle this challenge.
Addison, John T. (2005). “The Determinants of Firm Performance: Unions, Works
Councils and Employee Involvement / High-Performance Work Practices,” Scottish
Journal of Political Economy, 52(3): 406-450.
Alzola, Izaskun, Saioa Arando, Fathi Fakhfakh, Fred Freundlich, Mónica Gago, Virginie
Pérotin, and Alberto Zevi. (2010, July). Are Labour-Managed Firms All the Same? A
Comparison of Incentives for Growth, Democracy and Institutional Sustainability in the
Constitutions of Worker Cooperatives in Mondragon, Italy and France. Paper presented
at the 15th World Congress of the International Association for the Economics of
Bartlett, Will. (1994). “Employment in Small Firms: Are Cooperatives Different? Evidence from
Southern Europe,” in J. Atkinson and D. Storey (eds.), Employment: The Small Firm and
the Labour Market (New York: Routledge), pp. 256–87.
Bartlett, Will, John Cable, Saul Estrin, Derek C. Jones, and Stephen C. Smith. (1992).
“Labor-Managed and Private Firms in North Central Italy: An Empirical Comparison,”
Industrial and Labor Relations Review, 46(1): 103–18.
Bayo Moriones, José Alberto, Pedro Javier Galilea Salvatierra, and Javier Merino Díaz de
Cerio. (2003). “Participation, Cooperatives and Performance: An Analysis of Spanish
Manufacturing Firms,” in Takao Kato and Jeff Pliskin (eds.), Advances in the Economic
Analysis of Participatory and Labor-Managed Firms. Vol. 7, The Determinants of the
Incidence and the Effects of Participatory Organizations (Oxford: Elsevier Science), pp.
Ben-Ner, Avner. (1988). “Comparative Empirical Observations on Worker-Owned and
Capitalist Firms,” International Journal of Industrial Organization, 6: 7–31.
Ben-Ner, Avner. (1984). “On the Stability of the Cooperative Form of Organization,” Journal
of Comparative Economics, 8(3): 247-260.
Ben-Ner, Avner, and Derek C. Jones. (1995). “Employee Participation, Ownership and
Productivity: A New Theoretical Framework,” Industrial Relations, 34(4): 532–54.
Berman, Katrina V. (1982). “The United States of America: A Co-operative Model for Worker
Management,” in F. Stephen (ed.), The Performance of Labour-Managed Firms (London:
Macmillan), pp. 74–98.
Berman, Katrina V., and Matthew D. Berman. (1989). “An Empirical Test of the Theory of the
Labor-Managed Firm,” Journal of Comparative Economics, 13: 281–300.
Blinder, Alan S. (ed.). (1990). Paying for Productivity (Washington,D.C.: The Brookings
Bonin, John, Derek C. Jones, and Louis Putterman. (1993). “Theoretical and Empirical
Studies of Producer Cooperatives: Will Ever the Twain Meet?” Journal of Economic
Literature, 31: 1290–1320.
Bonin, John, and Louis Putterman. (1987). Economics of Cooperatives and the Labor-
Managed Economy (New York: Harwood).
Burdίn, Gabriel. (2010). Survivability of Worker Cooperatives Compared with Capitalist Firms
and Its Determinants: Evidence from Uruguay. Paper presented at the 15th World
Congress of the International Association for the Economics of Participation, Paris.
Burdίn, Gabriel, and Andrés Dean. (2010). Revisiting the Objectives of Worker Cooperatives:
An Empirical Assessment. Paper presented at the 15th World Congress of the
International Association for the Economics of Participation, Paris.
Burdίn, Gabriel, and Andrés Dean. (2009). “New Evidence on Wages and Employment in
Worker Cooperatives Compared with Capitalist Firms.” Journal of Comparative
Economics, 37: 517–33.
Centrostudi Legacoop. (2009). Aspetti quantitativi e qualitativi della cooperazione italiana
(Milan, Italy: Geco).
CG-SCOP. (2003). Guide juridique des SCOP (Paris: Scop Edit).
Conte, Michael A. 1982. “Participation and Performance in US Labour-Managed Firms,” in F.
Stephen (ed.), The Performance of Labour-Managed Firms (London: Macmillan), pp.
Conte, Michael A., and Meng-Hua Ye. (1995). “An Overlapping Generations Model of
Investment in Labor-Managed Firms,” Advances in the Economic Analysis of
Participatory and Labor-Managed Firms, 5: 159–73.
Conte, Michael A., and Jan Svejnar. (1990). “The Performance Effect of Employee Ownership
Plans,” in A. S. Blinder (ed.), Paying for Productivity: A Look at the Evidence
(Washington, DC: Brookings Institution Press), pp. 143–72.
Craig, Ben, and John Pencavel. (1995). “Participation and Productivity: A Comparison of
Worker Cooperatives and Conventional Firms in the Plywood Industry,” Brookings
Papers: Microeconomics: 121–74.
Craig, Ben, and John Pencavel. (1993). “The Objectives of Worker Cooperatives,” Journal of
Comparative Economics, 17: 288–308.
Craig, Ben, and John Pencavel. (1992). “The Behavior of Worker Cooperatives: The Plywood
Companies of the Pacific Northwest,” American Economic Review, 82(5): 1083–1105.
Defourny, Jacques, Saul Estrin, and Derek C Jones. (1985). “The Effects of Workers’
Participation on Enterprise Performance: Empirical Evidence from French Cooperatives,”
International Journal of Industrial Organization, 3: 197–217.
Doucouliagos, Chris. (1995). “Worker Participation and Productivity in Labor-Managed and
Participatory Capitalist Firms: A Meta-Analysis,” Industrial and Labor Relations Review,
Dow, Gregory K. (2003). Governing the Firm: Workers’ Control in Theory and Practice
(Cambridge, U.K.: Cambridge University Press).
Earle, John S., and Saul Estrin. (1998). “Workers’ Self-Management in Transition Economies,”
in Will Bartlett and Milica Uvalić (eds.), Advances in the Economic Analysis of
Participatory and Labor-Managed Firms, 6: 3–28.
Estrin, Saul. (1991). “Some Reflections on Self-Management, Social Choice, and Reform in
Eastern Europe,” Journal of Comparative Economics, 15: 349–66.
Estrin, Saul, Jan Hanousek, Evžen Kočenda, and Jan Svejnar. (2009). “The Effects of
Privatization and Ownership in Transition Economies,” Journal of Economic Literature,
Estrin, Saul, and Derek C. Jones. (1998). “The Determinants of Investment in Employee-
Owned Firms: Evidence from France,” Economic Analysis, 1(1): 17–28.
Estrin, Saul, and Derek C. Jones. (1995). “Workers’ Participation, Employee Ownership and
Productivity: Results from French Producer Cooperatives,” in Derek C. Jones and Jan
Svejnar (eds.), Advances in the Economic Analysis of Participatory and Labor-Managed
Firms, 5: 3–24.
Estrin, Saul, and Derek C. Jones. (1992). “The Viability of Employee-Owned Firms: Evidence
from France,” Industrial and Labor Relations Review, 45(2): 323–38.
Estrin, Saul, Derek C. Jones, and Jan Svejnar. (1987). “The Productivity Effects of Worker
Participation: Producer Cooperatives in Western Economies,” Journal of Comparative
Economics, 11(1): 40–61.
Fakhfakh, Fathi, Virginie Pérotin, and Mónica Gago. (2012). “Productivity, Capital and Labor
in Labor-Managed and Conventional Firms,” Industrial and Labor Relations Review,
FitzRoy, Felix, and Kornelius Kraft. (1987). “Cooperation, Productivity and Profit Sharing,”
Quarterly Journal of Economics, 102(1): 23–35.
Frey, Bruno S., and Reto Jegen. (2005). “Motivation Crowding theory,” Journal of Economic
Surveys, 15(5): 589-611.
Furubotn, Eirek G., and Svetozar Pejovich. (1970). “Property Rights and the Behaviour of the
Firm in a Socialist State: The Example of Yugoslavia,” Zeitschrift für Nationalökonomie,
Gago, Mónica, Virginie Pérotin, and Fathi Fakhfakh. (2008). “Is the Investment Decision
Different in Cooperatives?” Paper presented at the 14th IAFEP World Congress, Clinton,
Jensen, Michael, and William Meckling. (1976). “Theory of the Firm: Managerial Behavior,
Agency Costs and Ownership Structure,” Journal of Financial Economics, 3: 305–60.
Jones, Derek C. (2007). “The Productive Efficiency of Italian Producer Cooperatives: Evidence
from Conventional and Cooperative Firms,” in Sonja Novkovic and Vania Sena (eds.),
Advances in the Economic Analysis of Participatory and Labor-Managed Firms. Vol. 9,
Cooperative Firms in Global Markets: Incidence, Viability and Economic Performance,
Jones, Derek C. (2004). “Ownership and Participation: A Review of Empirical Evidence for
Transition Economies,” in Virginie Pérotin and Andrew Robinson (eds.), Advances in the
Economic Analysis of Participatory and Labor-Managed Firms. Vol. 8, Employee
Participation, Firm Performance and Survival, pp. 171–209.
Jones, Derek C. (1982). “British Producer Cooperatives, 1948–1968: Productivity and
Organizational Structure,” in D. C. Jones and J. Svejnar (eds.), Participatory and Self-
Managed Firms (Lexington, MA: Lexington Books), pp. 175–98.
Jones, Derek C. (1982). “The United States of America: A Survey of Producer Co-operative
Performance,” in F. Stephen (ed.), The Performance of Labour-Managed Firms (London:
Macmillan), pp. 53–73.
Jones, Derek C., and Jan Svejnar (eds.). (1982). Participatory and Self-Managed Firms
(Lexington, MA: Lexington Books).
Kalmi, Panu. (2004). “Exclusion from Employee Ownership: Evidence from Estonian Case
Studies,” in Virginie Pérotin and Andrew Robinson (eds.), Advances in the Economic
Analysis of Participatory and Labor-Managed Firms. Vol. 8, Employee Participation,
Firm Performance and Survival, pp. 35–65.
Kruse, Douglas, and Joseph Blasi. (1997). “Employee Ownership, Employee Attitudes, and
Firm Performance: A Review of the Evidence,” in Daniel J. B. Mitchell, David Lewin,
and Mahmood Zaidi (eds.), Handbook of Human Resource Management (Greenwich, CT:
JAI Press), pp. 113–51.
Kruse, Douglas, Richard Freeman, Joseph Blasi, Robert Buchele, Adria Scharf, Loren
Rodgers, and Chris Mackin. (2004). “Motivating Employee Owners in ESOP Firms:
Human Resource Policies and Company Performance,” in Virginie Pérotin and Andrew
Robinson (eds.), Advances in the Economic Analysis of Participatory and Labor-
Managed Firms. Vol. 8, Employee Participation, Firm Performance and Survival, pp.
Maietta, Ornella Wanda, and Vania Sena. (2004). “Profit Sharing, Technical Efficiency
Change and Finance Constraints,” in Virginie Pérotin and Andrew Robinson (eds.),
Advances in the Economic Analysis of Participatory and Labor-Managed Firms. Vol. 8,
Employee Participation, Firm Performance and Survival, pp. 149–67.
Megginson, W., and J. M. Netter. (2002). “From State to Market: A Survey of Empirical Studies
on Privatization,” Journal of Economic Literature, 39: 321–89.
Miyazaki, Hajime. (1984). “On Success and Dissolution of the Labor-Managed Firm in the
Capitalist Economy”, Journal of Political Economy, 94(5): 909-931.
Navarra, Cecilia. (2009). Collective Accumulation of Capital in Italian Worker Cooperatives
between Employment Insurance and “We-Rationality”: An Empirical Investigation.
EURICSE Working Paper.
Pencavel, John. (2012). “Worker Cooperatives and Democratic Governance,” forthcoming in
Grandori (ed.), Handbook of Economic Organization (Edward Elgar).
Pencavel, John. (2001). Worker Participation: Lessons from the Worker Co-ops of the Pacific
Northwest (New York: Russell Sage Foundation).
Pencavel, John, Luigi Pistaferri, and Fabiano Schivardi. (2006). “Wages, Employment and
Capital in Capitalist and Worker-Owned Firms,” Industrial and Labor Relations Review,
Pérotin, Virginie. (2006). “Entry, Exit and the Business Cycle: Are Cooperatives Different?”
Journal of Comparative Economics, 34: 295–316.
Pérotin, Virginie. (2004). “Early Cooperative Survival: The Liability of Adolescence,” in
Virginie Pérotin and Andrew Robinson (eds.), Advances in the Economic Analysis of
Participatory and Labor-Managed Firms. Vol. 8, Employee Participation, Firm
Performance and Survival, pp. 67–86.
Putterman, Louis. (1984). “On Some Recent Explanations of Why Capital Hires Labor,”
Economic Inquiry, 22: 171–87.
Robinson, Andrew, and Nick Wilson. (1993). “Co-operatives vs. Private Firms: An Empirical
Comparison of Co-operatives and Private Small Firms in Britain,” International Business
Review, 2(3): 281-296.
Ros, Agustin J. (2003). “Do ESOPs Motivate Employees? Worker Effort, Monitoring and
Participation in Employee Stock Ownership Plans,” in Takao Kato and Jeffrey Pliskin
(eds.), Advances in the Economic Analysis of Participatory and Labor-Managed Firms.
Vol. 7, The Determinants of the Incidence and the Effects of Participatory Organizations,
Russell, Raymond. (1995). Utopia in Zion: The Israeli Experience with Worker Cooperatives
(Albany, NY: State University of New York Press).
Russell, Raymond, and Robert Hanneman. (1995). “The Formation and Dissolution of Worker
Cooperatives in Israel, 1924–1992,” in Raymond Russell, Utopia in Zion: The Israeli
Experience with Worker Cooperatives (Albany, NY: State University of New York
Press), pp. 57–95.
Staber, Udo. (1993). “Worker Cooperatives and the Business Cycle: Are Cooperatives the
Answer to Unemployment?” American Journal of Economics and Sociology, 52(2): 129–
Staber, Udo. (1989). “Age-Dependence and Historical Effects on the Failure Rates of Worker
Cooperatives: An Event-History Analysis,” Economic and Industrial Democracy, 10: 59–
Stephen, Frank. (1982). “The Economic Theory of the Labour-Managed Firm,” in F. Stephen
(ed.), The Performance of Labour-Managed Firms (London: Macmillan), pp. 3–26.
Uvalić, Milica. (1992). Investment and Property Rights in Yugoslavia (Cambridge, U.K.:
Cambridge University Press).
Vanek, Jaroslav. (1977). The Labor-Managed Economy (Ithaca, NY: Cornell University Press).
Zevi, Alberto. (2005). “Il finanziamento delle cooperative,” in E. Mazzoli and S. Zamagni, Verso
una nuova toria della cooperazione (Bologna, Italy: Il Mulino), pp. 293–332.
Zevi, Alberto. (1982). “The Performance of Italian Producer Cooperatives,” in D. C. Jones and J.
Svejnar (eds.), Participatory and Self-Managed Firms (Lexington, MA: Lexington
Books), pp. 239–51.