PRELIMINARY. DO NOT REPRODUCE OR QUOTE WITHOUT AUTHORS CONSENT                   Productive Efficiency in the Mondragon Co...
I INTRODUCTION         There has been a long-standing interest by diverse scholars in diverse mattersconcerning the Mondra...
Pencavel, 1995). There is also a growing literature that usually focuses on more modestexamples of employee ownership, inc...
ownership.       Our method is that of insider econometrics case study (for a review of this andclosely related methods se...
II. The Case and Institutional Framework       Our case is the Eroski chain which is headquartered in Northern Spain. TheE...
stores were added to the Eroski chain during the period there were no exits.1        To enrich our understanding of the in...
the structure and functioning across stores in the three categories.2        These differences are most apparent for worke...
usually set above, the market rate --in 2008 it was 8%. Over time, as Eroski makesprofits, these individual stakes grow as...
with the usual ratio of top-bottom not exceeding 5:1.6 As such top managers tend toreceive lower earnings than do managers...
ineligible to attend the AGM or serve on the Board. As such the scope and nature of amember s opportunities to participate...
So far as wage setting and employment are concerned in stores with conventionalownership, all workers (including non-membe...
or majority EOFs. Second are studies of firms with more modest employee ownershipand/or more limited employee involvement....
expected to be the case in PCs).        These pioneering theoretical papers have elicited a voluminous amount ofresponses ...
data on both conventional firms and cooperatives.14 Amongst these perhaps mostnoteworthy is Craig and Pencavel (1995) who ...
While economic theory is ambiguous and empirical findings are unclear, based onthat literature together with our knowledge...
commitment to teams becomes undermined (Milgrom and Roberts, 1995, Ben-Ner andJones, 1995, and Kato and Morishima, 2002) T...
expending more discretionary effort and working harder and smarter than workforces inconventional stores, these feelings o...
able to see first hand the benefits of a more participatory environment and one thatprovides more job security than did th...
is a GESPA store, 0 otherwise. In addition to labor (L), store space is often consideredcrucial capital input (K) in retai...
The first-difference model is adopted for two reasons. First and most important,our field research at Eroski (repeated int...
percent level, confirming our prior expectation that young stores grow faster than.       As shown in the second column of...
How do COOP stores outperform other stores?        The recent literature on the High Performance Work System (HPWS) points...
opposed to only 0.02 percent for GESPA Hypermarket stores. Likewise, the averageINVOLVE was 0.44 percent for COOP City Sup...
employee involvement opportunities, we create STAKE i = average stake of employeeowners (monthly average of store i during...
recruitment are often an integral part of the High Performance Work System.Unfortunately our data provide no information o...
V. CONCLUSION       Recent years have seen a massive growth in employee ownership around theworld. A substantial volume of...
security; and working in firms with earnings differences that are substantially morecompressed than in comparable firms. W...
model are apparently especially well developed among Eroski stores, such as theexistence of high membership ratios and ave...
REFERENCESAkerlof, George A. 1982. "Labor Contracts as Partial Gift Exchange." Quarterly Journal        of Economics, Vol....
(June), pp. 197-217.Doucouliagos, Chris. 1995. "Worker Participation and Productivity in Labor-Managed        and Particip...
Worker Production Cooperatives." Industrial and Labor Relations Review, Vol.        31, No. 1 (October), pp. 18-30.Jones, ...
Sharing? Employee Outcomes under Employee Ownership, Profit Sharing, and       Broad-based Stock Options." National Bureau...
Supermarket                                              Hypermarket                                                      ...
Table 2 Sales Growth and Ownership Types: Insider Econometric EvidenceDependent variable=DlnQit                           ...
Table 3 Sales Growth and HRM for Hypermarket: Additional Insider Econometric EvidenceDependent variable=DlnQit            ...
Table 4 Sales Growth and HRM for Supermarket (City only): Additional Insider Econometric EvidenceDependent variable=DlnQit...
Upcoming SlideShare
Loading in...5
×

Productive efficiency in the Mondragon Cooperatives: Evidence from an econometric case study

1,700

Published on

Artículo de Saoia Arando, Mónico Gago, Derek C. Jones, Takao Kato, con el motivo del congreso internacional de economía social celebrado en EOI Sevilla y en colaboración con Goldsmiths College.
27/28_05_2010

Published in: Technology, Business
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total Views
1,700
On Slideshare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
30
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Transcript of "Productive efficiency in the Mondragon Cooperatives: Evidence from an econometric case study"

  1. 1. PRELIMINARY. DO NOT REPRODUCE OR QUOTE WITHOUT AUTHORS CONSENT Productive Efficiency in the Mondragon Cooperatives: Evidence from an Econometric case study Saioa Arando*, Monica Gago*, Derek C. Jones**, and Takao Kato*** For Presentation at the ASSA Meetings in Atlanta, January 5, 2010AbstractBy using new panel data we provide the first econometric study of efficiency for a member of thefamous Mondragon group of coops. Eroski is a retail distribution chain and, most unusually,based on employee ownership, there are two distinct types of hypermarkets namely cooperatives,with significant employee ownership and Gespas which have modest employee ownership. Forsupermarkets there are also conventional firms (SA) which have no employee ownership. Ourpanel data are monthly observations from February 2006 through May 2008, a total of about9,800 observations for supermarkets and about 2,150 for hypermarkets.By estimating first difference models we find that hypermarket stores with cooperative ownershipgrow sales significantly faster than do Gespa stores. For supermarkets the picture is morenuanced. We find that for urban supermarkets, those with cooperative ownership are moreproductive than Eroski stores that are conventionally owned. However for suburban stores wefind that conventional owned stores grow faster than both coops and Gespa.These findings, which are supplemented with more qualitative evidence, are consistent with thosewho argue for the existence of powerful incentive mechanisms for coop members who workunder institutional arrangements that differ from those facing workers in other firms, includingtheir having: a large financial stake in the firm; substantial employee involvement; unusual jobsecurity; and working in firms with earnings differences that are substantially more compressed.We attribute the failure to find an effect for more moderate combinations on employee ownershipand employee involvement to most of the factors that underpin the coop advantage being absentfrom Gespa stores.Key words: employee ownership, producer cooperatives, labor managed firm, productive efficiency,Mondragon* ETEO, Mondragon Unibertsitatea, 20560 Oñate, SPAIN. Email: sarando@eteo.mondragon.edu, mgago@eteo.mondragon.edu,** Dept. of Economics, Hamilton College, Clinton NY 13323; e-mail: djones@hamilton.edu*** Dept. of Economics, Colgate University, Hamilton, NY. tkato@colgate.eduJones and Kato are grateful to support from NSF SES-052117 and Jones is also appreciative of support from anIkerbasque fellowship. The authors are extremely grateful to a number of individuals working for the case firm forinterviews, comments and access to the data. The data used in this study are proprietary and were obtained only bysigning a confidentiality agreement. Hence the authors are unable to release these data. The paper has benefitted fromcomments from Fred Freundlich and Jeff Pliskin. 1
  2. 2. I INTRODUCTION There has been a long-standing interest by diverse scholars in diverse mattersconcerning the Mondragon cooperatives (e.g., Johnson and Whyte, 1977, Bradley andGelb, 1982). Since these early studies, the Mondragon group has continued to grow -- bysome estimates the MCC firms as a group represent the seventh largest consortium inSpain and, since its inception in the early 1950 s, total employment has also grown tomore than 80,000 in 2007. Moreover, facts such as no job losses ever having beensustained by coop members, including during the present economic crisis, help explainwhy other studies of Mondragon have continued to appear regularly (e.g., Joshi andSmith, 2008). At the same time, a conspicuous characteristic of the vast bulk of thepublished literature on Mondragon coops is that there have been few if any appliedstudies that use standard hypothesis-testing methods. Thus on one of the key questions ofinterest to economists, the efficiency of the Mondragon coops compared withconventional firms, most published work has been restricted to comparisons usingefficiency indicators that are quite aggregated, for example comparisons at an industrialor overall group level between coops and conventional firms (e.g., Thomas and Logan,1982.) The limited nature of the research on Mondragon coops contrasts with the moregeneral literature concerning employee owned firms. The first econometric studies of theperformance of firms in which there is substantial/majority employee-ownership firms(EOFs), often considered as producer co-operatives (PCs) or labor managed firms(LMFs) appeared more than 30 years ago (e.g., Jones and Backus, 1977.) Work hascontinued to flow including influential studies of the U.S. plywood coops (Craig and 2
  3. 3. Pencavel, 1995). There is also a growing literature that usually focuses on more modestexamples of employee ownership, including the shared capitalism literature for the US(e.g., Kruse, Freeman, and Blasi, 2008), European countries (e.g., Pendleton andRobinson, 2008 for the U.K), transition economies (e.g., Estrin et al, 2009) and Japan(e.g., Jones and Kato, 1995.) However, the issues that are examined in the literature onemployee owned firms are still not definitively settled. Thus while the meta analysis byDoucouliagos (1995) concludes that the balance of evidence demonstrates betterperformance by PCs compared to participatory capitalist firms, other assessments,including Bonin et al. (1993:1305) and Dow (2003:184) are not so sanguine. Concerningthe performance of PCs, this ambivalence is particularly apparent when evaluation isrestricted to studies that endeavor to make comparisons between PCs and conventionalfirms within the same industry (for a review, see, e.g., Dow, 2003:184.) By providingnew evidence for a Mondragon case we also contribute to many of these debates. Our study is facilitated because we are fortunate to have access, for the first time,to primary data for the population of stores in the Eroski retail chain, with Eroski by farthe largest employer within the Mondragon group today. By using these new panel datawe provide the first econometric study of efficiency for any Mondragon coop. Moreover,growth in the chain has resulted in there being large numbers of individual stores thathave three distinct categories of employee ownership, ranging from significant (thecoops), through limited (known as Gespa) to zero (conventional stores). Coops also havelarge numbers of workers who are not (yet) members and in Gespa stores many workerschoose not to become members. These panel data enable us to contrast the impact onefficiency of both cooperative and limited employee ownership with conventional 3
  4. 4. ownership. Our method is that of insider econometrics case study (for a review of this andclosely related methods see Ichniowski and Shaw, 2003 and Jones and Kato, 2007). Theoverwhelming bulk of the literature on EOFs, especially majority EOFs, has adopted afirm-level approach. While this is a valuable approach, as is widely recognized there arepotential problems with this empirical strategy including issues surrounding measurementerror, endogeneity and omitted variables. By adopting a case study approach we are ableto more thoroughly investigate the ramifications of important institutional realities incoops, such as the co-existence of member and non-member workers, issues whichtypically are not afforded central attention in larger firm-level studies. Thus by providingwhat is apparently the first econometric case study of a firm with substantial employeeownership, potentially we provide a valuable complement to the body of findings derivedfrom firm level studies, as well as a contribution to a literature that is of growingimportance. The plan of this paper is as follows. The next section highlights key institutionalfeatures of our case while in the third part we briefly review key themes in the theoreticalliterature and a review of previous empirical work. This is followed by a description ofour data. In the main sections of the paper we describe the first difference approach that isthe basis of our estimating framework as well as our econometric findings. In thepenultimate section we provide diverse kinds of additional evidence that mainly relates tothe mechanisms that underpin our key empirical findings. In the final section we offerconcluding remarks and discuss the implications of our study. 4
  5. 5. II. The Case and Institutional Framework Our case is the Eroski chain which is headquartered in Northern Spain. TheEroski group is a diversified company with different activities including sport and leisureoutlets though its core businesses are supermarkets and hypermarkets, which are thefocus of our investigation. Both supermarkets and hypermarkets stores sell similar items,although there is some variation in the range of items sold, since the outlets are ofdifferent size ranging from floor space of 1950 sq.m. to 12, 853 in hypermarkets and 162to 2500 sq. m. for supermarkets. Smaller supermarkets carry a product assortmentessentially in the food area that is a subset of the product mix offered by larger stores.Each hypermarket is divided into three basic divisions-- food, clothing and domesticgoods. A key strategy is to sell rather standard products to a wide range of customers withall items in stock on display, and self-service is the main form of service in mostdepartments. For the bulk of employees the main tasks are to receive goods, shelve items,and maintain the appearance of their department. In departments such as specializedfoods it appears that customers are apt to call on the expertise of sales clerks more oftenthan in other departments such as basic foods. The overwhelming bulk of these retail outlets are in Spain (2398 of 2441), withthe remainder in France and Andorra, though amongst those 2398 units it is the 109hypermarkets and 705 supermarkets that are at center stage. Total employment in 2007was 50, 587 and on the basis of total retail space this made Eroski the third largest retailchain in Spain in that year. During the study period (2005-2008) the Eroski chaincontinued to grow including pursuing a strategy of acquisitions. However, while many 5
  6. 6. stores were added to the Eroski chain during the period there were no exits.1 To enrich our understanding of the institutional realities at the case we readvarious materials that the company provided and we also made repeated visits to the firmheadquarters where we have had extensive discussions with senior personnel. Many ofthese personnel had made frequent and recent visits to branch stores and thus hadintimate knowledge of these branches. We have also spent considerable time on fieldtrips including multiple visits to outlets of this company in the Basque region as well asdifferent areas of Spain; during each of these visits we interviewed the store manager andemployees. The retail chain began operations in Northern Spain in 1969. Most of thecooperatively owned stores are in the Basque region and in these cooperatives employeeshave substantial employee ownership. To sustain growth in the chain, in 1997 Eroskibegan to open stores in other parts of Spain. Most of these stores are not coops but insome, known as Gespa, employee-members have ownership stakes, though they are moremodest than in coops. Since 1997 in its quest to become a leader in retailing throughoutSpain, the case has begun to acquire conventional stores. Unsurprisingly there aresubstantial differences among these acquired stores in terms of pre-acquisitionperformance. To control for these differences represents an important challenge in ourempirical work since all of these acquisitions continue to have no employee ownership.Hence, from the perspective of the extent of employee ownership, there are three distincttypes of store. In turn, these ownership differences result in considerable differences in 1 To allow for adjustment costs, in our study we will focus on hypermarkets and supermarketsthat have been part of the Ersoki chain for at least 6 months. 6
  7. 7. the structure and functioning across stores in the three categories.2 These differences are most apparent for worker-members in Eroski coops who,compared to workers elsewhere in Eroski, have unusual opportunities to participate inboth ownership and decision-making. Equally it is important to realize that, usually, thereare non-member workers in the workforce. While many of these non-members areprospective members on probation, they also include workers on temporary contracts. Inthe main it appears that workers on temporary contracts are quite low skilled, in suchpositions as cashiers. This non-member workforce in Eroski coops, averages about 25%of the total workforce. Effectively all coop workers who work under permanent contracts are expected tobecome members and, as such, significant worker owners. While there is no fixedprobationary period, at some point (usually not less than six months) the immediatesupervisor of the candidate for membership, after soliciting opinions of other coopworkers, makes a recommendation with the ultimate decision concerning membershipbeing made by the store s manager. In selecting prospective members, a key requirementis the willingness and ability of the candidate to commit to a substantial capitalcontribution, a sum that currently is about 6000 Euros which in 2009 amounted to about30% of the average annual remuneration in an Eroski store3. While this initial stakeremains substantially individually owned, about 20 % of it is allocated to collectivelyowned reserves. The member s stake receives an interest rate that is related to, though 2 We should also note that Eroski, as are other MCC coops, is supported by a web ofinstitutions (see, e.g., Thomas and Logan, 1982 and Joshi and Smith, 2008). 3 However the capital contribution requirement is not as onerous as it appears since it can bespread over 5 years. In addition, possibilities exist for some new members to use previously paidunemployment premia towards these stakes, since coop members are guaranteed job security. 7
  8. 8. usually set above, the market rate --in 2008 it was 8%. Over time, as Eroski makesprofits, these individual stakes grow as distributions from surpluses are credited to theseindividuals capital accounts.4 Coop members also have opportunities to participate in decision-making in wayssubstantially beyond those available to workers in other Eroski stores. Thus worker-members are able to attend the AGM (Annual General Meeting), though the large numberof members requires that this is done on a representative basis. More importantly perhaps,worker-members are able to be elected to the Governing Council (the Board of directors)and the Social Council (the body responsible for determining many matters of interest toworkers, such as working conditions.). At the same time, the potential influence ofworker members in Eroski coops is necessarily circumscribed since, unlike inmanufacturing coops, there is another large group of members in Eroski, namelyconsumer-members. The Governing Council comprises equal numbers of representativesfor worker- and consumer-members. The understandings and implicit policies concerning job security andremuneration are potentially of crucial importance to the functioning of coops. A keyincentive for workers to want to become and to remain as coop members is job securityno coop members have ever been laid off.5 In addition, coops have wage structures thatare much more compressed and more flexible than in firms outside the group. Thus thenorm is for coop members in non-managerial positions to receive a premium of at least20% over their outside counterparts. Also the internal wage differences are compressed, 4 For more detail on these features of Mondragon coops, see Thomas and Logan, 1982 andBradley and Gelb, 1982. 5 There have been occasional instances of store closures. In such instances members arealways offered comparable employment in nearby Eroski stores. 8
  9. 9. with the usual ratio of top-bottom not exceeding 5:1.6 As such top managers tend toreceive lower earnings than do managers in conventional retail stores (by some estimatesabout 30% below outside rates.) Turning to the Eroski stores known as Gespa, as in coops individuals can becomemembers and have individual ownership stakes though membership in Gespa involves acapital stake that is about half as large as in a coop, about 3,000 Euros. This representsabout 25% of the average annual earnings for workers in a Gespa store. However, andunlike in coops, not all permanent workers in Eroski stores are required/expected tobecome members. In particular, when the stores were acquired, existing workers were notobliged to become members. However, new workers who are offered permanent contractsare expected to become members. These individual membership stakes earn a dividendwhere the interest rate is determined in a similar way to individual stakes of coopmembers, and is usually a little over the market rate. Over time, as Eroski makes profits,these individual stakes grow as distributions from surpluses are credited to individualscapital accounts. However reflecting lower initial levels and shorter average life thefirst Gespa began in 1997, whereas coops have existed since 1969-- in 2009 the averageindividually owned stake in an Gespa was substantially less than in Eroski coops.Membership in Gespa, as with membership in coops, provides what is effectively 100%job security no Gespa members have ever been laid off, and in the few instances ofGespa store closures, members have always been offered alternative employment nearby.However, while Gespa members are able to be elected to the Social Council, they are 6 There are some strategic positions for which the gap is higher, sometimes approachingwithin business differences of 8:1. Still these differences are much more compressed than incomparable capitalist retail chains. 9
  10. 10. ineligible to attend the AGM or serve on the Board. As such the scope and nature of amember s opportunities to participate in control and membership in a Gespa issubstantially below that for members in coops. Indeed in many interviews we heardviews expressed that Gespa membership was widely regarded as a second class form ofmembership. From the perspective of employee ownership, all stores with conventionalownership in the Eroski chain, and unlike other Eroski stores, do not provideopportunities for employee ownership or special structures through which employees canparticipate in decision making. At the same time it is important to emphasize that all ofthese stores represent acquisitions of what were capitalist firms. As such many workers inthese stores had worked for the previous capitalist owners. Now they experience workingas part of a cooperative chain one feature of which is all workers in all stores beingsubject to key features of the same set of HR policies. The language contained in variousinternal company documents and the associated institutional arrangements stronglysuggests that by working in a store within a coop chain, the lot of these continuingworkers might be expected to have improved --arguably they are subject to betterworking conditions and better treatment by managers than previously. For example, theworkers in these conventional stores are encouraged to participate in meetings and thereare policies concerning meetings between employees and supervisors and annualdevelopment discussions even though they cannot, as in other stores in the group, becomeowner-members. Also there is a raft of policies that encourage training and skillformation. 7 7 The need for such policies emerges from our discussions with managers at the chain whoemphasized that the firm s way of operating, or a key part of its competitive strategy, was that 10
  11. 11. So far as wage setting and employment are concerned in stores with conventionalownership, all workers (including non-members in coops and Gespa and workers in theseconventional firms,) receive no less than the wage rates that are set out in the collectiveagreement that is negotiated by the retail workers trade union and which applies to allretail workers. Based on the preceding discussion we believe that our case is a good one in whichto test propositions on the impact of varying degrees of employee ownership andparticipatory practices on organizational performance. The extent of these two sets ofpractices varies both across the three types of stores and within stores for differentcategories of workers. The varying combinations of these practices also have implicationsfor other policies, e.g. base earnings, wage differences and job security. Hence we caninvestigate hypotheses concerning the joint impact on business performance ofdifferences in ownership and participation for the different types of store. In the nextsection to flesh out some of these hypotheses we will review key themes in differenttheoretical literatures.III Theory and Previous empirical work Since ours is not a theoretical contribution, here we merely review some of thecentral themes in theoretical and empirical work emerging from three broad sets ofliterature. First are studies of the economic theory of the LMF which investigates LMFsemployees discretionary effort mattered for company performance. Employees are needed for morethan just being there, and high turnover of employees is not desirable. This is reflected in thecompany s written HRM strategy which emphasizes skill development of employees and themanagement capabilities of supervisors, and also career development and job rotation. These issueswere also raised in our discussions with management members, suggesting genuine commitment to theHRM strategy. This perspective applies to all types of stores. 11
  12. 12. or majority EOFs. Second are studies of firms with more modest employee ownershipand/or more limited employee involvement. Third is the extensive behavioral literaturethat examines interrelated themes such as management-employee relations, social and giftexchange, employee incentives and the implications for expected organizationalperformance. This review of theoretical and empirical evidence is interwoven withreferences to our discussion (in section II) to key institutional features at Eroski so thatwe end up with specific predictions for expected comparative performance for differenttypes of stores.8 For majority EOFs, the economic theory of the LMF yields conflictingpredictions about the productivity effects of worker participation in ownership and is thusinconclusive concerning the expected comparative performance of EOFs andconventional firms.9 Much of the early influential theoretical work (e.g. Vanek, 1970)argued that co-operative firms would generate very strong incentives for labor resultingin high technical efficiency of labor. By contrast other studies were more pessimisticconcerning the expected performance of PCs. Alchian and Demsetz (1972) and Jensenand Meckling (1979) argue that productivity will be lower in a cooperative becauseefficient monitoring of workers requires the monitor to be the claimant on the firmsprofits and that the cost of monitoring increases with the number of monitors. Anotherinfluential paper is Holmstrom (1982) who argues that effort level is expected to be besetwith free-rider problems and thus sub-optimal when work takes place in teams (as is 8 One general point to note is that studies of technical efficiency within retailing, are thin onthe ground. For coops, an early investigation of British retail coops is Jones (1987). So far as studiesof the impact of HR on business performance in retailing is concerned prior work include Ben-Ner etal. (1999) and Jones, et al.,( 2006, 2009). 9 See Bonin, Jones, and Putterman (1993), Jones and Pliskin (1991a) and Dow (2003) forsurveys. 12
  13. 13. expected to be the case in PCs). These pioneering theoretical papers have elicited a voluminous amount ofresponses and theoretical objections. Thus Macleod (1988) shows how, in a repeatedgame framework, effort supply in LMFs need not be below that in conventional firm.Others point to different benefits of PCs. Thus cooperatives are expected to be moreproductive than conventional firms because incentives (financial participation), peergroup pressure (horizontal monitoring) and the close identification of cooperativemembers with the firm will elicit greater effort from workers (Jones and Svejnar, 1985;Fitzroy and Kraft, 1987). Subsequent work (reviewed in Bonin et al. 1993; Dow, 2003)also recognized that predictions are more nuanced once it is recognized that real worldLMFs may not have 100% membership. Turning to the empirical literature, often the relative performance of conventionalfirms and producer cooperatives has been estimated by comparing subsample means ofmeasures such as value added per worker using data on both conventional firms andcooperatives.10 Most econometric evidence has been obtained from samples exclusivelyof producer cooperatives.11, 12,13 A few studies have estimated production functions using 10 For example, George (1982). 11 Some exceptions are Jones (1987), Lee (1988), Berman and Berman (1989), Estrin (1991)and Craig and Pencavel (1995). 12 For example, see Defourney, Estrin, and Jones (1985), Estrin, Jones, and Svejnar (1987),and Jones and Svejnar (1985). 13 Using estimates of how the productive efficiency of firms varied with respect to measuresof financial and decision making participation, the authors of these studies estimated the efficiency ofa typical cooperative relative to a firm with no worker participation. Since the samples of cooperativesoften exhibited considerable variation over both firms and time in the degree of worker participation,the estimated productivity effects might be reliable. However, other things remaining the same, onewould prefer a sample of both conventional firms and cooperatives since the variance of the predictionerrors is lower for observations that are similar to those in the sample than for atypical ones. 13
  14. 14. data on both conventional firms and cooperatives.14 Amongst these perhaps mostnoteworthy is Craig and Pencavel (1995) who carefully gathered data for plywood coopsand conventional firms in the Pacific Northwest in that industry. The authors estimateseparate Cobb Douglas production functions for several types of firms including coopsand conventional firms. They find that coops are between 6 and 14% more efficient thanthe principal conventional firms though there is little difference between the efficiency ofthe unionized and classical mills. More recently Jones (2007) also assembles data forconventional firms and PCs in the same industry, namely the Italian construction sector.However no evidence is found that coops are more efficient than conventional firms.15 Insum, it would seem that a reasonable conclusion based on the research to date formajority EOFs is that there is no strong evidence that either cooperatives or conventionalfirms have a sizeable and persistent significant edge in performance over otherorganizational form. Equally it is apparent that there is a need for more targeted research.For example, the most frequently cited comparative study nowadays is probably that ofCraig and Pencavel (1995). However, while the quality of the data the authors use ismost impressive arguably the robustness of the findings are somewhat diminished by therelatively small size of the data set (170 observations for 34 mills), and the use of aproblematic measure of capital in the production function estimates. Also not all studieshave taken into account key features of the institutional realities, such as the variation inmembership amongst firms. 14 George (1982), Jones (1987, 2007), Conte and Svejnar (1988), Lee (1988), Berman andBerman (1989), Estrin (1991) and Craig and Pencavel (1995). But only the papers by Jones, Lee,Estrin, Craig and Pencavel (1995) have focused on the relative technical efficiency of cooperatives. 15 For Mondragon coops, the only econometric study is Martin (2000). This unpublishedstudy reports findings for a few firms (in machine tools) that do not yield a clear picture. 14
  15. 15. While economic theory is ambiguous and empirical findings are unclear, based onthat literature together with our knowledge of the institutions at Eroski (as discussed insection II), we make the following prediction concerning the performance of Eroski co-ops. Even though the formal arrangements in Eroski coops are somewhat short of what isenvisaged in the pure theory case of the LMF, Eroski coops do provide high levels ofownership and participation as well as substantial job security for members. Front lineworkers benefit from receiving wage premia compared to non-members. We expect thatthese arrangements will lead to a more committed and motivated workforce who wouldexpend more discretionary effort and work harder and smarter than the workforces inGespa or conventional stores so that cooperatives are expected to display much higherlevels of efficiency. Another source of performance advantage of coops might stem fromthe possibility that high levels of peer monitoring in coops would lead one to expect cooplabor forces to have fewer layers of supervisors than in conventionally owned stores. The second set of literature we examine concerns firms with minority EO and/ormore limited employee involvement. But again the theoretical predictions that emergefrom that literature are ambivalent. Some theoretical and empirical literature suggests thatan individual change in organizational design, such as modest employee ownership, isexpected to be sufficient to produce sustained benefits to the firm.16 By contrast otherliterature argues that, for sustained benefits, complementary measures are needed. Anindividual initiative when introduced alone may be insufficient to lead to persistent gains.For example, employees might need more sharing of enterprise rewards through financialparticipation, such as employee stock ownership to accompany teams lest their 16 See for example reviews in Blinder (1990) and in Blair and Kochan (2000). 15
  16. 16. commitment to teams becomes undermined (Milgrom and Roberts, 1995, Ben-Ner andJones, 1995, and Kato and Morishima, 2002) The empirical literature often reports thatthe productivity-enhancing effects of individual practices introduced alone may be short-lived since initiatives lack a complementary mechanism, such as also delegating power tofront-line workers (e.g., for quality circles, Levine, 1995 and Jones and Kato, 2007). Bycontrast, and broadly speaking, the empirical literature also finds evidence in support ofsuch complementarities (e.g. between employee ownership and employee involvement,such as the shared capitalism studies of Freeman et al, 2008, or for Japan Kato andMorishima (2002) or for UK firms, Pendleton and Robinson (2008).17 While economic theory for firms with minority ownership and control isambiguous and empirical findings are unclear, our knowledge of the institutions at Eroski(as discussed earlier in section II) and that literature leads us to make the followingprediction concerning performance in Eroski Gespas. While Gespas do provide areasonable level of employee ownership and job security for members, the extent ofparticipation is rather modest. Also, front line workers in Gespa stores do not enjoy thewage premia that coop members receive. Also Gespa workforces tend to suffer frommore divisions than do coop workforces because of lower membership ratios not allpermanent workers are required to be members. For reason such as these, the synergiesbetween ownership and participation, especially compared to coop members, are mutedand we would not expect Gespa stores to do nearly as well as coops. Furthermore, whilesome of the institutional arrangements might be expected to result in worker members 17 However, it is also important to recognize that there do not appear to be any studies thathave investigated these propositions within a business that is a cooperative and part of an elaborate setof institutions outside the firm. 16
  17. 17. expending more discretionary effort and working harder and smarter than workforces inconventional stores, these feelings of being second class members might underminesuch forces. Hence we have no strong expectations concerning Gespa performancecompared to conventional stores. Lastly we turn to research mainly from behavioral economics that focuses onmanager-employee relations and ideas such as fairness, social exchange and giftexchange (e.g. Akerlof, 1982; Kahneman et al, 1986; Fehr and Schmidt, 1999.) A pointof departure for much of that mainly experimental literature is that employees havediscretion in their effort choices since contracts designed to regulate employmentrelations are necessarily incomplete. In investigating what effort levels might be expectedto emerge in situations where employers and employees interact frequently over extendedperiods, one focus has been on the role that fairness concerns might play (e.g. Fehr,Goette and Zehnder, 2008). Arguably considerations surrounding fairness are apt to beespecially pertinent within a cooperative firm. Here we focus on the implications of keyfindings from the fairness literature for employees in stores in Eroski that areconventionally owned.18 The main argument we make recognizes that many workers in theseconventionally owned stores that are part of the Eroski group previously worked in thestores when they were not co-op owned. Now they are subject to employmentarrangements that arguably they find to be much fairer than in the past. Such workers are 18 Ideas of fairness clearly carry over to coops and Gespas. Thus in cooperatives, where thecontext is one in which worker members know their managers well, and are aware that wagedifferences are compressed, we can see how ideas of fairness produce more flexibility andreciprocation and high levels of effort among member-workers. Similarly, concerns about being a second-class member might help to undermine effort levels in Gespa stores. 17
  18. 18. able to see first hand the benefits of a more participatory environment and one thatprovides more job security than did their previous employment. In other words theirfairness perceptions shifted when they began to work for the coop, even underconventional ownership it was a case of entering a new firm if not a new workplace(Kahneman et al, 1986.) Consequently we predict that discretionary effort levels in suchfirms will be higher than in competitors.19IV. INSIDER ECONOMETRIC EVIDENCEDo COOP stores outperform other stores? To capture the performance effects of differences in ownership structure, weestimate the following first-difference model:For Hypermarket stores,(1) DlnQit = LDlnL it + cCOOPi + DMARKETit + YEAROPENEDi + additional controls+D itFor Supermarket stores,(2) DlnQit = LDlnL it + cCOOPi + gGESPAi + DMARKETit + YEAROPENEDi + additional controls + D itwhere indicates the first difference between month t and t-1; Qit is output (real sales) instore i in month t; Lit is employment (measured by the number of full-time equivalentworkers) in store i in month t; COOPi is a dummy variable taking a value of 1 if store i isa coop store, 0 otherwise; and GESPAi is a dummy variable taking a value of 1 if store i 19 By drawing on other studies that focus on incentives within organizations (e.g.,Prendergast, 1999) of different groups within firms, such as the literature for law and medical firms(e.g., Gaynor and Gertler, 1995 or Gaynor and Pauly, 1990) or the more general treatment byEncinosa et al. (2005) one would end up with a similar prediction. 18
  19. 19. is a GESPA store, 0 otherwise. In addition to labor (L), store space is often consideredcrucial capital input (K) in retail service production (see, for instance, Jones, Kalmi andKauhanen, 2006). For all Eroski stores during the time period under study, however,month to month variations of store space are zero and hence in our first-difference model,DlnKit = 0. To make sure that the estimated coefficients on COOPi and GESPAi are capturingthe pure employee ownership effects, we include a number of controls. First and perhapsmost important, a store located in a rapidly growing market with rising population andaverage household income will naturally grow its sales faster. To control for suchdifferences in each store s market condition, we include DMARKETit where MARKETitis monthly market index in month t for the area which store i serves. The monthly marketindex is provided by the Spanish National Statistical Institute which is affiliated with theMinistry of Economy and Tax. Second, we have been told by our informants at Eroski that due to the standardlifecycle model of retail stores, younger stores tend to grow faster than older stores. Tocontrol for such a lifecycle effect, we also include YEAROPENEDi = the year store i wasopened. Finally, we also include constant (to capture an Eroski-wide time trend which iscommon to all Eroski stores regardless of its ownership types), monthly dummy variables(to capture seasonality of retail sales), and year dummy variables (to control for year timeeffects) as additional controls.20 20 We also consider a full set of interaction terms involving monthly dummy variables andyear dummy variables. Our main results change little with the use of such a full set of interactionterms though there is slight efficiency loss. These as well as all other unreported restuls are avaiableupon request from the authors. 19
  20. 20. The first-difference model is adopted for two reasons. First and most important,our field research at Eroski (repeated interviews with our key informant as well asmultiple managers) led us to believe that sales growth is indeed a primary business goalof Eroski, resulting in sales growth (which is equal to DlnQit) as the dependent variable.Second, first-difference models are often used in the modern empirical literature tocontrol for unobserved heterogeneity. As such, unobserved heterogeneity of stores iscontrolled for by our use of the first-difference model insofar as such heterogeneity istime-invariant. Table 2 reports the OLS estimates of Eq. (1).21 For the Hypermarket segment, asshown in the first column, the estimated coefficient on COOP is positive and significantat the 5 percent level, confirming that COOP stores grow faster than GESPA stores. Thesize of the estimated coefficient suggests a plausible growth rate advantage of COOPstores over GESPA stores, i.e., on average COOP stores grow faster than GESPA storesby 0.2 percentage points. The estimated coefficients on the control variables are ofexpected signs and significant. Specifically, the estimated coefficient on DlnLit is positiveand significant at the 1 percent level, and the size of the coefficient implies that the outputelasticity of labor in the underlying Cobb-Douglas production function is a little over 0.5.We also find the estimated coefficient on DMARKETit to be positive and significant atthe 1 percent level, confirming that a store located in a rapidly growing market withrising population and average household income will naturally grow its sales faster, andhence that it is important to control for the market conditions of the store location.Finally, the estimated coefficient on YEAROPENEDi is positive and significant at the 5 21 All estimates will be clustered at firm levels. 20
  21. 21. percent level, confirming our prior expectation that young stores grow faster than. As shown in the second column of the table, for the Supermarket segment, wefind no evidence for the growth rate advantage of COOP stores (the estimated coefficientis very small and highly insignificant). Our failure to find any significant difference ingrowth between COOP and GESPA stores for the supermarket segment did not surpriseour key informant at Eroski who pointed out the large heterogeneity of stores in thesupermarket segment as a possible reason for the finding. He then suggested us to focuson a subgroup of the supermarket segment, City Supermarket stores (a group ofsupermarket stores located in downtown as opposed to suburb). Apparently having better customer service employees is particularly important for those City Supermarketstores which are still somewhat reminiscent of intimate, small neighborhood groceries. Amajor strength of COOP stores lies in the fact that COOP helps employees develop asense of ownership and hence makes them more committed employees who are willingand capable of providing closer and helpful attention to their customers. Such a COOPadvantage is probably more relevant to City Supermarket stores than other supermarkets.In short, we are more likely to detect a positive performance effect of COOP in thesubgroup of City Supermarkets. The last column of Table 2 reports the results for the City Supermarket group. Asexpected, the estimated coefficient on COOP is positive and statistically significant at the5 percent level. The size of the estimated coefficient implies a considerable 0.7percentage-point growth rate advantage enjoyed by COOP City Supermarket stores overconventional City Supermarket stores (note that there is no GESPA store in this subgroupof the supermarket segment). 21
  22. 22. How do COOP stores outperform other stores? The recent literature on the High Performance Work System (HPWS) points tothe real possibility of an establishment boosting its performance by adopting a variety ofcomplementary new work practices (often called High Performance Work Practices) andtapping into the ability of frontline workers to produce valuable local knowledge throughtheir collective efforts; and dealing with local shocks autonomously through collaborationamong themselves. Such diverse HPWPs can be grouped into the following three keyelements of the HPWS.22Opportunities: First, in the HPWS, front-line workers will be given opportunities toexert discretionary effort, acquire useful local knowledge, and share it with their co-workers, and higher-level engineers and managers. The importance of providing suchopportunities is self-explanatory. After all, a key objective of the HPWS is to tap intofrontline workers discretional effort and ability to produce valuable local informationand deal with local shocks. Without such opportunities, there will not be any performancegain. To measure the extent of such employee involvement opportunities, we constructfrom our data INVOLVE i = proportion of scheduled work hours spent on joint labor-management meetings (monthly average of store i during the time period under study). As shown in Table 1, for both the Hypermarket and Supermarket segments,COOP stores allow for much more employee involvement than other stores. For instance,the proportion of scheduled working hours spent on joint labor-management meetings permonth (INVOLVE) was on average 0.24 percent for COOP Hypermarket stores as 22 See, for instance, Kochan and Osterman, 1994, Appelbaum, et. al. 2000 and Boning,Ichniowski and Shaw, 2007. 22
  23. 23. opposed to only 0.02 percent for GESPA Hypermarket stores. Likewise, the averageINVOLVE was 0.44 percent for COOP City Supermarket stores as opposed to negligible0.002 percent for conventional City Supermarket stores. As such, it is plausible thatCOOP stores outperform other stores in part by providing their employee owners withmore opportunities to produce useful local knowledge and respond effectively to localshocks.Incentives: Providing workers with such opportunities to produce useful local knowledgeand share it with management is not sufficient. Obviously if the interest of workers is notaligned with that of the firm, workers will have little incentive to put force effort andproduce performance-enhancing local information and share it with management. Theinterest alignment between workers and the firm is fostered by two types of humanresource management policies: (i) financial participation schemes (such as employeestock ownership, profit sharing, gainsharing, and broad-based stock option) by which thefinancial wellbeing of workers is more tied to the final wellbeing of the firm; and (ii)information sharing mechanisms through which management shares importantinformation with workers, and fosters their loyalty and commitment to the firm.23 To gauge the strength of such incentives for workers to take advantage of 23 In addition, job security can be an important necessary condition for the High PerformanceWork System. For instance, local knowledge accumulated through collaboration of frontline workersis often firm-specific in nature, and its value will be considerably lower outside of the firm. As such,in the absence of the practice of long-term employment, frontline workers will have less incentive toaccumulate such firm-specific human capital. Furthermore, even if workers obtain valuable localknowledge, they may not share it in the absence of long-term employment. For example, performance-enhancing local knowledge discovered by frontline workers may result in labor-saving technologicalchange. Imagine that a frontline worker has just discovered a way to perform his/her job more quicklyand thus afford performing his/her co-worker s job as well. This may result in a loss of his/her co-worker s job or even worse his/her own job (which is now performed by his/her co-worker). Unlesssome degree of job security is credibly assured, the frontline worker will have an incentive not toreveal such performance-enhancing local information with management. For the importance of jobsecurity in the participatory employment system such as the Japanese system, see for example Levine(1995) and Carmichael and MacLeod (1993). 23
  24. 24. employee involvement opportunities, we create STAKE i = average stake of employeeowners (monthly average of store i during the time period under study). Table 1 showsthat employees in COOP Hypermarket stores have greater stake in their firm than GESPAemployees with the average stake in the firm worth 33,000 Euros for employees in COOPHypermarket stores and only 2,500 Euros for employees in GESPA Hypermarket stores(unsurprisingly for City Supermarket stores, employees in conventional stores have nostake in the firm). Again, the HPWS literature suggests that employees in COOP storeshave stronger incentives to take advantage of employee involvement opportunities;produce valuable local knowledge; and respond quickly and effectively to local shockswithout invoking lengthy formal involvement of supervisors. It follows that COOP storesoutperform other stores. In addition to STAKE i, our data provide yet another dimension of the overallstrength of incentives, MEMBER i= proportion of workers who are COOP or GESPAmembers (monthly average of store i during the time period under study). While STAKEcaptures the intensity of incentives (how big a deal it is for the average employee ownerto help her store outcompete its competitors), MEMBER measures the scope ofincentives (what proportion of the total labor force in the store has some stake in thefirm). As shown in Table 1, COOP stores tend to have higher proportions of employeeswith some stake in the firm.Ability: Finally, even if frontline workers are given an opportunity to produce valuablelocal knowledge and share it with management AND have the appropriate incentive to doso, such useful local information may never be generated or shared widely in the firm inthe absence of appropriate ability and skill of workers. As such, careful screening and 24
  25. 25. recruitment are often an integral part of the High Performance Work System.Unfortunately our data provide no information on the amount of HPWS-relevant training(such as problem solving; team work; customer relations). Instead the data allow us toconstruct TRAININGi = proportion of scheduled hours spent on training in general(monthly average of store i during the time period under study). Table 1 shows asomewhat mixed picture. For the Hypermarket segment, COOP stores on average devoteless time to training than GESPA stores whereas for the City Supermarket segment,COOP stores on average devote more time to training than conventional stores. Overall,the aforementioned measurement issue makes it somewhat difficult to interpret the resultson training. Specifically, we estimate the following first-difference model:(3) DlnQit = LDlnL it + hHPWPi + DMARKETit + YEAROPENEDi + additional controls + D itFor HPWPi, as discussed above, we consider INVOLVEi, STAKEi, MEMBERi, andTRAININGi. Table 3 shows the OLS estimates of Eq. (3) for the Hypermarket segment.The results are overall supportive of the HPWS with all estimated coefficients ofexpected positive sign and two of them being significant at the 1 percent level(INVOLVEi, and STAKEi). As shown in Table 4, the estimates of Eq. (3) for a subgroup of City Supermarketstores are less precise and mostly insignificant although the estimated coefficient onMEMBERi is close to significant at the 10 percent level.24 24 We also estimated a fully nested version of Eq. (3) with all four HPWP variablesconsidered simultaneously. The results turned out to be quite robust to the use of such a fully nestedspecification although the estimates are slightly less precise due to multicollinearily as expected. 25
  26. 26. V. CONCLUSION Recent years have seen a massive growth in employee ownership around theworld. A substantial volume of theoretical and empirical evidence has appeared thatinvestigates the performance of such firms, but it is inconclusive concerning thecomparative performance of EOFs and conventional firms, as well as the comparativeperformance of firms with majority and minority EO. The evidence also shows howemployee ownership has assumed a wide variety of forms, including the producercooperative in which majority control and ownership is vested in the workforce. Amongstsuch labor managed firms, one example that has attracted close and sustained attention byresearchers are the Mondragon cooperatives. In this paper, we are fortunate to use thefirst micro-econometric evidence for a Mondragon cooperative and, since stores fall intothree distinct ownership categories, we are able to contribute to some of these debates. By estimating first difference equations we find: (i) hypermarket stores withcooperative ownership outperform Gespa stores; (ii) for supermarkets, stores withcooperative ownership outperform other Eroski stores but there are no differences inperformance between Gespa stores and SA stores. We also find that there are differencesby type of store. These findings, which are supported by a series of robustness checks and are alsosupplemented with additional evidence that bears on the mechanisms that help to explainwhy coops are better-performers, are consistent with those who argue for the existence ofpowerful incentive mechanisms for coop members who work under institutionalarrangements that differ from those facing workers in other firms, including their having:a large financial stake in the firm; substantial employee involvement; unusual job 26
  27. 27. security; and working in firms with earnings differences that are substantially morecompressed than in comparable firms. We attribute the failure to find an effect for moremoderate combinations on employee ownership and employee involvement to absence inGespas of many of the factors that underpin the coop advantage. Some of our findings appear to be contrary to some recent field experiments (andalso lab experiments.) For example (Gneezy and List 2006) report how paying generouswages has no effect on worker productivity. In accounting for the differences betweensuch findings and those reported in this study we would point to the potential key role oforganizational structure in this case the existence of cooperative institutions. Fewexperiments seem to have been able to take this crucial dimension into account.Similarly, we interpret our findings of higher efficiency for workers employed onpermanent contracts in conventional stores (compared to workers on temporarycontracts), as consistent with propositions surrounding the role of fairness in thebehavioral literature. In considering the implications of our findings we note that there are limitations tothe approach--econometric case studies cannot easily address concerns about selectivityand external validity (Jones et al., 2006). However, the available data do enable us to useeconometric methods that introduce reasonable controls for individual heterogeneity.25 We note that the bulk of the earlier empirical literature on the impact of the LMFis for manufacturing and not retailing firms such as our case. Moreover, compared tofirms examined in other studies, many potentially important features of the cooperative 25 Although, as noted earlier, other studies in this area that have been able to do so havetypically found that results are unaffected by, for example, the use of fixed effects estimators. 27
  28. 28. model are apparently especially well developed among Eroski stores, such as theexistence of high membership ratios and average ownership stakes that are far bigger thanin other coops outside Mondragon. The particular configuration of coop featuresavailable at Eroski might help to explain our finding more powerful effects ofcooperation here than has been found in other studies. Relatedly we note that for retailingfirms there is an emerging body of literature that investigates different combinations ofinnovative work practices (such as financial participation and teams.) In the main thatliterature finds positive effects flowing from such practices and that the impact onperformance is higher when innovative practices are especially well-developed inretailing firms. Clearly there are many more issues that future research of Mondragonfirms can fruitfully address. 28
  29. 29. REFERENCESAkerlof, George A. 1982. "Labor Contracts as Partial Gift Exchange." Quarterly Journal of Economics, Vol. 97, No. 4 (November), pp. 543-69.Alchian, Armen A. and Harold Demsetz. 1972. "Production, Information Costs, and Economic Organization." American Economic Review, Vol. 62, No. 5 (December), pp. 777-95.Appelbaum, Eileen, Thomas Bailey, Peter Berg, and Arne L. Kalleberg. 2000. Manufacturing advantage: Why high-performance work systems pay off. Ithaca and London: Cornell University Press, ILR Press.Ben-Ner, Avner and Derek C. Jones. 1995. "Employee Participation, Ownership, and Productivity: A Theoretical Framework." Industrial Relations, Vol. 34, No. 4 532-54.Ben-Ner, Avner, Fanmin Kong, and Stacie Bosley. 1999. "Workplace Organization and Human Resource Practices: The Retail Food Industry." University of Minnesota, Retail Food Industry Center, Working Paper 00-01.Berman, Katrina V. and Matthew D. Berman. 1989. "An Empirical Test of the Theory of the Labor-Managed Firm." Journal of Comparative Economics, Vol. 13, No. 2 (June), pp. 281-300.Blair, Margaret M. and Thomas A. Kochan, eds. 2000. The new relationship: Human capital in the American corporation. Washington, D.C.: Brookings Institution Press.Blinder, Alan S., ed. 1990. Paying for productivity: A look at the evidence. Washington, D.C.: Brookings Institution.Bonin, John P., Derek C. Jones, and Louis Putterman. 1993. "Theoretical and Empirical Studies of Producer Cooperatives: Will Ever the Twain Meet?" Journal of Economic Literature, Vol. 31, No. 3 (September), pp. 1290-320.Boning, Brent, Casey Ichniowski, and Kathryn Shaw. 2007. "Opportunity Counts: Teams and the Effectiveness of Production Incentives." Journal of Labor Economics, Vol. 25, No. 4 613-50.Bradley, Keith and Alan Gelb. 1982. "The Replication and Sustainability of the Mondragon Experiment." British Journal of Industrial Relations, Vol. 20, No. 1 (March), pp. 20-33.Carmichael, H. Lorne and W. Bentley MacLeod. 1993. "Multiskilling, Technical Change and the Japanese Firm." Economic Journal, Vol. 103, No. 416 (January), pp. 142- 60.Conte, Michael A. and Jan Svejnar. 1988. "Productivity Effects of Worker Participation in Management, Profit-Sharing, Worker Ownership of Assets and Unionization in U.S. Firms." International Journal of Industrial Organization, Vol. 6, No. 1 (March), pp. 139-51.Craig, Ben and John Pencavel. 1995. "Participation and Productivity: A Comparison of Worker Cooperatives and Conventional Firms in the Plywood Industry." Brookings Papers on Economic Activity, Vol., No. 121-60.Defourney, 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, Vol. 3, No. 2 29
  30. 30. (June), pp. 197-217.Doucouliagos, Chris. 1995. "Worker Participation and Productivity in Labor-Managed and Participatory Capitalist Firms: A Meta-analysis." Industrial and Labor Relations Review, Vol. 49, No. 1 (October), pp. 58-77.Dow, Gregory. 2003. Governing the firm: Workers control in theory and practice. Cambridge, the U.K.: Cambridge University Press.Encinosa, William E., III, Martin Gaynor, and James B. Rebitzer. 2007. "The Sociology of Groups and the Economics of Incentives: Theory and Evidence on Compensation Systems." Journal of Economic Behavior and Organization, Vol. 62, No. 2 (February), pp. 187-214.Ernst Fehr, Lorenz Goette, Christian Zehnder. 2008. "A Behavioral Account of the Labor Market: The Role of Fairness Concerns." IZA Discussion Paper 3901 (forthcoming in: Annual Review of Economics, 2009)Estrin, Saul. 1991. "Some Reflections on Self-Management, Social Choice, and Reform in Eastern Europe." Journal of Comparative Economics, Vol. 15, No. 2 (June), pp. 349-66.Estrin, Saul, Jan Hanousek, Evzen Kocenda, and Jan Svejnar. 2009. "The Effects of Privatization and Ownership in Transition Economies." Journal of Economic Literature, Vol. 47, No. 3 (September), pp. 699-728.Fehr, Ernst and Klaus M. Schmidt. 1999. "A Theory of Fairness, Competition, and Cooperation." Quarterly Journal of Economics, Vol. 114, No. 3 (August), pp. 817-68.FitzRoy, Felix R. and Korenelius Kraft. 1987. "Cooperation, Productivity, and Profit Sharing." Quarterly Journal of Economics, Vol. 102, No. 1 (February), pp. 23-35.Freeman, Richard, Douglas Kruse, and Joseph Blasi. 2008. "WORKER RESPONSES TO SHIRKING UNDER SHARED CAPITALISM." National Bureau of Economic Research, Inc, NBER Working Papers: 14227.Gaynor, Martin and Paul J. Gertler. 1995. "Moral Hazard and Risk Spreading in Partnerships." RAND Journal of Economics, Vol. 26, No. 4 (Winter), pp. 591-613.Gaynor, Martin and Mark V. Pauly. 1990. "Compensation and Productive Efficiency of Partnerships: Evidence from Medical Group Practice." Journal of Political Economy, Vol. 98, No. 3 (June), pp. 544-73.George, D. 1982. "Workers Co-operatives in Denmark." Managerial and Decision Economics, Vol. 3, No. 205-12.Gneezy, Uri and John A. List. 2006. "Putting Behavioral Economics to Work: Testing for Gift Exchange in Labor Markets Using Field Experiments." Econometrica, Vol. 74, No. 5 (September), pp. 1365-84.Holmstrom, Bengt. 1982. "Moral Hazard in Teams." Bell Journal of Economics, Vol. 13, No. 2 (Autumn), pp. 324-40.Ichniowski, Casey and Kathryn Shaw. 2003. "Beyond Incentive Pay: Insiders Estimates of the Value of Complementary Human Resource Management Practices." Journal of Economic Perspectives, Vol. 17, No. 1 (Winter), pp. 155-80.Jensen, Michael C. and William H. Meckling. 1976. "Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure." Journal of Financial Economics, Vol. 3, No. 4 (October), pp. 305-60.Johnson, Ana Gutierrez and William Foote Whyte. 1977. "The Mondragon System of 30
  31. 31. Worker Production Cooperatives." Industrial and Labor Relations Review, Vol. 31, No. 1 (October), pp. 18-30.Jones, Derek C. 1987. "The Productivity Effects of Worker Directors and Financial Participation by Employees in the Firm: The Case of British Retail Cooperatives." Industrial and Labor Relations Review, Vol. 41, No. 1 (October), pp. 79-92.Jones, Derek C. 2007. "The Productive Efficiency of Italian Producer Cooperatives: Evidence from Conventional and Cooperative Firms," in Sonja Novkovic and Vania Sena, eds, Cooperative Firms in Global Markets Volume 10 of the book series: Advances in the Economic Analysis of Participatory & Labor-Managed Firms. Amsterdam: Elsevier/JAI, pp. 3-28.Jones, Derek C. and Jeffrey Pliskin. 1991. "The Effects of Worker Participation, Employee Ownership and Profit Sharing on Economic Performance: A Partial Review," in Russell Raymond and Rus Veljko, eds, Ownership and Participation: International Handbook of Participation in Organizations Vol.2. Oxford: Oxford University Press, pp. 43-63.Jones, Derek C. and David K. Backus. 1977. "British Producer Cooperatives in the Footwear Industry: An Empirical Evaluation of the Theory of Financing." Economic Journal, Vol. 87, No. 347 (September), pp. 488-510.Jones, Derek C., Panu Kalmi, and Antti Kauhanen. 2006. "Human Resource Management Policies and Productivity: New Evidence from an Econometric Case Study." Oxford Review of Economic Policy, Vol. 22, No. 4 (Winter), pp. 526-38.Jones, Derek C., Panu Kalmi, and Antti Kauhanen. 2009. "How Does Employee Involvement Stack Up? The Effects of Human Resource Management Policies on Performance in a Retail Firm." Industrial Relations (forthcoming).Jones, Derek C. and Takao Kato. 1995. "The Productivity Effects of Employee Stock- Ownership Plans and Bonuses: Evidence from Japanese Panel Data." American Economic Review, Vol. 85, No. 3 (June), pp. 391-414.Jones, Derek C. and Takao Kato. 2007. "The Impact of Teams on Output, Quality and Downtime: An Empirical Analysis using Individual Panel Data." IZA Discussion Paper No. 2917 (forthcoming in Industrial and Labor Relations Review).Jones, Derek C. and Jan Svejnar. 1985. "Participation, Profit Sharing, Worker Ownership and Efficiency in Italian Producer Cooperative." Economica, Vol. 52, No. 28 (November), pp. 449-65.Joshi, Sumit and Stephen C. Smith. 2008. "Endogenous Formation of Coops and Cooperative Leagues." Journal of Economic Behavior and Organization, Vol. 68, No. 1 (October), pp. 217-33.Kahneman, Daniel, Jack L. Knetsch, and Richard Thaler. 1986. "Fairness as a Constraint on Profit Seeking: Entitlements in the Market." American Economic Review, Vol. 76, No. 4 (September), pp. 728-41.Kato, Takao and Motohiro Morishima. 2002. "The Productivity Effects of Participatory Employment Practices: Evidence from New Japanese Panel Data." Industrial Relations, Vol. 41, No. 4 487-520.Kochan, Thomas and Paul Osterman. 1994. Mutual Gains Enterprise: Forging a Winning Partnership among Labor, Management, and Government (Hardcover): Harvard Business School Press.Kruse, Douglas, Richard Freeman, and Joseph Blasi. 2008. "Do Workers Gain by 31
  32. 32. Sharing? Employee Outcomes under Employee Ownership, Profit Sharing, and Broad-based Stock Options." National Bureau of Economic Research, Inc, NBER Working Papers: 14233.Lawler, Edward E. III. 1986. High-Involvement Management: Participative Strategies for Improving Organizational Performance. San Francisco: Jossey-Bass.Lee, B. 1988. Productivity and Employee Ownership: The Case of Sweden. Stockholm: Sweden: Trade Union Institute for Economic Research.Levine, David I. 1995. Reinventing the workplace: How business and employees can both win. Washington, D.C.: Brookings Institution.MacLeod, W. Bentley. 1988. "Equity, Efficiency, and Incentives in Cooperative Teams," in Derek C. Jones and Jan Svejnar, eds, Advances in the economic analysis of participatory and labor-managed firms. Volume 3. Greenwich, Conn. and London: JAI Press, pp. 5-23.Martin, Terence. 2000. "The Impact of Worker Ownership on Firm-Level Performance: A Comparative Study." Unpublished Ph.D. Thesis, Yale University.Milgrom, Paul and John Roberts. 1995. "Complementarities and Fit: Strategy, Structure, and Organizational Change in Manufacturing." Journal of Accounting and Economics, Vol. 19, No. 2-3 (April), pp. 179-208.Pendleton, Andrew. 2006. "Incentives, Monitoring, and Employee Stock Ownership Plans: New Evidence and Interpretations." Industrial Relations, Vol. 45, No. 4 (October), pp. 753-77.Pendleton, Andrew and Andrew M. Robinson. 2008. "Employee Stock Ownership, Participation, and Productivity." Paper presented at the IAFEP meeting: Hamilton College.Prendergast, Canice. 1999. "The Provision of Incentives in Firms." Journal of Economic Literature, Vol. 37, No. 1 (March), pp. 7-63.Thomas, H. and C. Logan. 1982. Mondragon: An Economic Analysis. Winchester, MA: Allen & Unwin.Vanek, Jaroslav 1970. The General Theory of Labor-Managed Market Economies. Ithaca, NY: Cornell University Press. 32
  33. 33. Supermarket Hypermarket All stores City COOP GESPA COOP GESPA Conventional COOP ConventionalDlnQit .0021048 -.0004499 .0042155 .0060926 .0053058 .0104688 .0019786 (.166259) (.234495) (.159047) (.169218) (.174939) (.182554) (.116191)DlnLit .0003716 -.0016009 .0023519 .0039068 .0024821 .0068902 .0022351 (.04335) (.066897) ( .09743) (.061481) (.087372) (.139007) (.102414)DMARKET it .0034202 .0035085 .002913 .003626 .003759 .0028605 .0012249 (.115210) (.107146) (.112875) (.096741) (.104789) (.115623) (.099853)YEAROPENEDi 1995.48 1999.904 1998.405 2000.626 1999.364 2000.18 2002.053 (5.46753) (4.49017) (4.74845) (2.77468) (4.94237) (2.52702) (1.84229)INVOLVEi .0024464 .0002408 .0033306 .0011858 2.23e-06 .0044161 .0000192 (.004767) (.000727) (.005705) (.003443) (.000126) (.008445) (.000332)STAKEi 33295.79 2511.332 26270.68 865.6311 1.398661 23030.07 0 (8847.05) (1010.40) (8175.98) (201.354) (23.56457) (10545.04) (0)MEMBERi .7589575 .6075966 .7289384 .5180572 0 .6442763 0 (.073878) (.135189) (.118557) (.153238) (0) (.1549173) (0)TRAININGi .0074278 .0080867 .0138549 .0102537 .0062481 .0108354 .0059075 (.01298) (.015204) (.038580) (.021452) (.053389) (.041531) (.012875)N 675 1420 4747 703 8001 967 321Table 1 Summary StatisticsSources: Individual store-level monthly data from February 2006 through May 2008 provided by Eroski.Note: Standard Deviation in parentheses.Definitions of the VariableQit=real sales of store i in time t;Lit=number of all full-time equivalent workers working in store i in time tMARKETit=monthly market index of the area served by store i in time tYEAROPENEDi = year in which store i was initially opened.INVOLVEi = proportion of scheduled hours spent on joint labor-management meetings (monthly average of store i during the time period understudy)STAKEi = average stake of employee owners (monthly average of store i during the time period under study)MEMBERi = proportion of workers who are COOP or GESPA members (monthly average of store i during the time period under study)TRAININGi = proportion of scheduled hours spent on training (monthly average of store i during the time period under study) 33
  34. 34. Table 2 Sales Growth and Ownership Types: Insider Econometric EvidenceDependent variable=DlnQit Hypermarket Supermarket Supermarket City onlyDlnL it 0.552*** 0.265*** 0.292** [6.57] [4.96] [2.03]DMARKETit 0.645*** 0.815*** 1.165*** [9.92] [19.39] [5.90]YEAROPENEDi 0.00016* 0.0004** 0.0002 [1.87] [2.36] [0.29]COOPi 0.0022** -0.0003 0.0074** [2.94] [-0.32] [2.63]GESPAi -0.0001 [-0.10]N 2070 10994 1195R-squared 0.852 0.404 0.311Sources: Individual store-level monthly data from February 2006 through May 2008 provided by Eroski.Notes:1. Absolute values of t are in parentheses (all standard errors are robust and clustered at the individual store level).2. All models include constant and monthly dummy and year dummy variables.3. For Hypermarket, all stores are either COOP or GESPA and hence the omitted reference category is GESPA. For Supermarket,there are COOP, GESPA and conventional stores and the omitted reference category is conventional stores. For Supermarket Cityonly, there are only COOP and conventional stores and hence the omitted reference category is conventional stores.* p<0.10, ** p<0.05, and *** p<0.01 34
  35. 35. Table 3 Sales Growth and HRM for Hypermarket: Additional Insider Econometric EvidenceDependent variable=DlnQit (i) (iii) (v) (ii)DlnL it 0.552*** 0.576*** 0.552*** 0.552*** [6.57] [6.53] [6.57] [6.57]DMARKETit 0.645*** 0.653*** 0.645*** 0.645*** [9.92] [9.51] [9.92] [9.92]YEAROPENEDi 0.00014 0.0002*** 0.0001 0.00007 [1.61] [2.71] [1.16] [0.87]INVOLVEi 0.558*** [2.84] -8STAKEi 6.2x10 *** [2.74]MEMBERi 0.0037 [1.01]TRAININGi 0.255 [1.15]N 2070 1889 2070 2070R-squared 0.852 0.847 0.852 0.852Sources: Individual store-level monthly data from February 2006 through May 2008 provided by Eroski.Notes:1. Absolute values of t are in parentheses (all standard errors are robust and clustered at the individual store level).2. All models include constant and monthly dummy and year dummy variables.* p<0.10, ** p<0.05, and *** p<0.01 35
  36. 36. Table 4 Sales Growth and HRM for Supermarket (City only): Additional Insider Econometric EvidenceDependent variable=DlnQit (i) (iii) (v) (ii)DlnL it 0.292** 0.292** 0.292** 0.292** [2.03] [2.03] [2.03] [2.03]DMARKETit 1.165*** 1.165*** 1.165*** 1.165*** [5.90] [5.90] [5.90] [5.90]YEAROPENEDi -0.0002 -0.0002 0.00003 -0.0002 [-0.36] [-0.27] [0.004] [-0.29]INVOLVEi 0.151 [0.48] -8STAKEi 2.26x10 [0.29]MEMBERi 0.0069 [1.51]TRAININGi -0.047 [-0.67]N 1195 1195 1195 1195R-squared 0.311 0.311 0.311 0.311Sources: Individual store-level monthly data from February 2006 through May 2008 provided by Eroski.Notes:1. Absolute values of t are in parentheses (all standard errors are robust and clustered at the individual store level).2. All models include constant and monthly dummy and year dummy variables.* p<0.10, ** p<0.05, and *** p<0.01 36

×