Françoise Carré and Chris Tilly
Françoise Carré
Chris Tilly
Center for Social Policy Institute for Research on Labor and Employment
McCormack Graduate School University of California – Los Angeles
University of Massachusetts Boston Chris_Tilly@irle.ucla.edu
Francoise.Carre@umb.edu

Much, indeed most, analysis of job quality is carried out within specific national contexts.
But globally, the largest differences in job quality are cross national, and cross national comparisons potentially offer the greatest analytical leverage in understanding the determinants of job quality. They also enable us to consider options for policies, and combinations of policies, not otherwise considered when we focus on a single national context.
Analysts have proposed varied frameworks for such cross national comparisons of job characteristics and quality (for example Bamber et al. 2004; Hall and Soskice 2001a;
Maurice, Sellier, and Silvestre 1986). However, these frameworks have typically been induced from, and applied to, manufacturing and other higher wage or unionized industries. Recently, we and others undertook six country (United States and Europe) comparisons of low-wage industries, predominantly in services, comparisons that are compiled in Jérôme Gautié and John Schmitt’s (forthcoming) edited volume,
Low-Wage
Work in the Wealthy World . In this process, we became convinced that existing frameworks must be extended in order to adequately characterize and explain differences in low-wage jobs.
In this paper, we build on, adapt, and extend existing frameworks to understand entrylevel, low-wage jobs and forces shaping job quality. We argue that entry-level, lowwage, service jobs offer a distinctive and useful vantage point on the sources of cross national differences in job quality. In particular, they shine a spotlight on government mandates and supports, and on the influence of institutions that support the reproduction of the workforce. We have observed that both macro-level comparative analyses of dominant national institutions and average labor market outcomes, and micro-level comparisons of matched plants or of jobs in the same company, while valuable, leave important analytical gaps worth exploring in order to understand job outcomes.
Meso-level analyses address these gaps; they entail comparing the same sector across countries and capturing within-sector, within-country variation as well as cross-country variation. They are useful for understanding variations in job outcomes because they shed particular light on how national-level institutions and organization-level strategies and histories interact. In this paper, we rely on, and build upon, the contributions of meso analysis to develop a framework for cross national comparisons of job outcomes that is both flexible and has broad applicability to low-wage service work--at least in wealthier countries.
A meso level analytical framework achieves many objectives of comparative analysis but nonetheless encounters three familiar challenges. First, it needs to address the complex
1

interactions between labor supply and demand on the one hand, and institutional and cultural forces on the other. A second challenge is the need to carefully contextualize institutions to avoid unwarranted presumptions of equivalence. A third challenge is presented by the issue of balancing structural influences with employer agency and strategy, as well as considering reactions of other social actors (e.g. worker organizations) to employer strategies.
We use entry-level retail jobs as the illustrative lens through which to apply existing notions as well as challenge and expand them. We use the example of retail trade jobs to illustrate the argument, drawing on our own fieldwork in the U.S. and Mexico (Carré and
Tilly 2007, Tilly 2006) as well as comparative analysis conducted jointly with others
(Carré, Tilly, van Klaveren, and Voss-Dahm 2009, Askenazy, Berry, Carré, Prunier-
Poulmaire, and Tilly 2010, Esbjerg et al 2008, Mason and Osborne 2008) and comparative work by other researchers (Baret, Lehndorff, and Sparks 2000, Jany-Catrice and Lehndorff 2005). In this earlier comparative research, we and others have wrestled with the problems highlighted in this paper, and we build very directly on these earlier insights.
Why retail entry level jobs? As a sector, retail generates high volumes of entry-level, low-wage, jobs in wealthy countries and these are jobs with a significant service component. In several nations, it has been shown to be a leader in innovative labor deployment practices that tend to erode conventional tools for safeguarding labor standards, and provide welfare state safety net benefits. It has also taken the lead in pushing for labor deregulation. Retail work per se is service work, indeed increasingly so, and so are a growing segment of low-wage jobs (and of jobs in general). The nature of the work, and its volume, provide ample cases to illustrate our argument. Furthermore, on the face of it, service producing processes may be more susceptible to national norms and customs regarding work and consumption than manufacturing processes that offer far greater opportunities to implement machine-driven work processes.
In addition to focusing on higher-wage industries, comparative analyses to date have also tended to compare relatively wealthy countries. In this paper we incorporate research on
Mexico, a middle income country (with roughly one-third to one-quarter the GNI per capita of the United States or Germany, depending on the currency conversion [World
Bank 2010]) with a high level of informal economic activity. Without claiming to arrive at a truly global comparative toolbox, we explore how comparative analysis must shift to encompass comparisons between richer and poorer countries.
In the remainder of this paper, we start by briefly discussing the challenges of evaluating job quality differences across countries. We then briefly review macro and micro comparative approaches, and argue for the unique contributions of a meso-level analysis.
We continue by elaborating why examining entry-level jobs presents particular vantage points for cross national comparisons and consider the implications of a focus on lowwage service jobs. We sketch out some of the key elements of a meso-level analysis, using retail jobs as our example. We review what we have learned from conducting such analysis and sketch out areas that require further exploration. We close with brief conclusions. Throughout, we primarily reference the United States and the five European
2

countries (Denmark, France, Germany, Netherlands, United Kingdom) we studied with
Maarten van Klaveren and Dorothea Voss-Dahm (forthcoming)
1
and with Askenazy,
Berry and Prunier-Poulmaire (2010), but also when possible consider how the analysis is altered by including Mexico.
While the great bulk of this paper tackles how to explain cross national differences in job quality, here we briefly discuss some of the issues in evaluating job quality across countries. Consistent with much of the literature on job quality (though formulations differ: see Tilly 1997, Bazen, Lucifora, and Salverda 2005), we suggest that there are five main dimensions of job quality, each of which can easily be separated into additional dimensions:
1.
Compensation (wages, fringe benefits)
2.
Basic working conditions (danger, cleanliness, stress, and so on)
3.
Work schedule and predictability of number and timing of hours and days of work
4.
Skill level and degree of worker control over the labor process
5.
Training and mobility opportunities.
Each of these pose particular challenges in translation across national boundaries. Even something as straightforward as the wage quickly raises a multitude of issues. Should absolute wage levels be compared at exchange rates (most appropriate when considering capitalist investment choices) or purchasing power parity (most appropriate when comparing worker well-being)? For what purposes are absolute wage levels the appropriate metric, and for what purposes do relative wage levels serve better? What is the impact of the wage’s embeddedness in institutional contexts ranging from systems of taxation to migrant remittances to debt peonage?
The “absolute vs. relative” question is particularly important. For comparisons of national level of well-being, there is no alternative to absolute measures. But for mesolevel comparisons of the same sector or occupation across-national boundaries, the matter gets considerably more complex. In terms of compensation, arguably the relative position of a set of jobs is most relevant. This is particularly true when comparing across nations with very different average incomes, for which the first-order difference in compensation is likely to be driven by those national averages. But as regards work schedule or mobility, absolute differences are of great interest, since they signal different ways of organizing similar functions. Perhaps the safest option is to combine absolute and relative measures of job characteristics.
Evaluations of relative measures of job quality must also take into account, at least to some extent, the full range of alternatives available to workers. While data availability makes it tempting to limit comparisons to wage workers, this is only part of the picture.
It’s also important to note the availability and importance of subsistence activities, self-
1 As relevant, we draw on national studies on Denmark (Esbjerg et al 2008) and the United Kingdom
(Mason and Osborne 2008).
3

employment, or public support. Export assembly workers in China or Vietnam, for whom the main alternative is subsistence agriculture, inevitably experience a manufacturing job differently from someone in a country with a much wider range of job options or a generous basic income policy.
Most international comparisons of work take the form of macro or micro analyses.
Macro analyses take as their unit of analysis the national system as a whole, typically as captured by a small number of indicators. Simple economic models of trade and development (Samuelson1948, Solow 1956) predict economic convergence and early sociological models of industrialization (Kerr et al. 1964) identify sources of crossnational convergence in industrial relations and labor regulations. Indeed, some academic analysts argue that convergence outcomes are taking place (Barro and Sala-i-Martin
1999). But stubborn evidence of divergence has directed most macro research energy to explaining differences. Scholars have put forward macro comparisons of the broad mode of economic regulation (Hall and Soskice 2001a), the welfare state (Esping-Andersen
1990), labor relations (Bamber, Lansbury, and Wailes 2004), and many other elements of national political economies (see the essays in Hall and Soskice 2001b for a selection).
Their typical mode of argument is to identify key institutional differences and then draw out their consequences in terms of national averages or of some set of jobs viewed as typical or archetypical. Much recent work takes the form of construction of a new typology based on some overlooked dimension of work (Dieckhoff 2008, Gash 2008,
Hult and Edlund 2008). Some of the most sophisticated macro analyses examine national systems of stratification, segmentation, and inequality (Kalleberg 1988, Kerckhoff 1995).
Macro-level analyses provide valuable insights into broad cross-national differences, but because they necessarily focus on a small number of factors and on the average or typical job, they are not well suited to exploring the interaction between varying business strategies and different national institutions.
Micro studies, on the other hand, match a small number of workplaces (often a pair) across different nations, often within the same company, in order to control for as many sources of variation as possible and focus on the impact of differing national contexts (for example Bank Muñoz 2008, Ferner and Almond 2007). Again, these studies have yielded important understandings of how and why jobs differ across national environments. However, because of their micro nature they risk focusing on idiosyncratic factors. And by construction, they typically omit within-country variation and varied employer strategies.
Meso analyses bring something new to the discussion. Meso analyses examine a single sector across countries, offering a more fine-grained look than macro analyses but capturing intra-industry variation in a way that micro studies do not (e.g. Katz 1997, Katz and Darbishire 2000, Kochan, Lansbury, and MacDuffie 1997). Such meso analyses are particularly apt at examining the interaction between business strategy and institutions.
Much of this work, though not all, looks at unionized sectors, often manufacturing. It is
4

important to extend this level of analysis to less unionized and lower paid service work.
In this regard, the contributions of Baret, Lehndorff, and Sparks (2000) and Bosch and
Lehndorff (2005a) are significant. Both ground their analyses of institutional effects in multiple domains, building on the seminal work of Maurice, Sellier, and Silvestre (1986).
Though the two formulations are distinct, they are closely related. Jean Gadrey’s (2000) chapter in Baret, Lehndorff, and Sparks specifies domains of market structure, organizational structure, industrial relations, and the domestic sphere (gender roles and social protection). Bosch and Lehndorff (2005b), in their introduction to their edited collection suggest a somewhat different, but related, set of domains: product and consumers, management strategies, labor markets and labor market institutions, and the welfare state and gender relations. Notable here, particularly in contrast with analyses of manufacturing and telecommunications, is the last sphere in each case, focusing attention on the welfare state and gender relations. In this paper, we build on these contributions, and most particularly Gadrey’s classification of spheres of action, seeking to strengthen and more fully operationalize a framework derived from this earlier research.
Examining interactions of labor supply and demand with institutions as they affect entrylevel, low-wage, jobs in particular—rather than employment systems as a whole— offers particular vantage points and priorities for analysis. First, we are more likely to prioritize the role of national institutions and their role in job outcomes. These are jobs than employers have not upgraded – in other words, it is harder to find an important role for organization-level factors and skill/technology factors in job outcomes. To experience job improvement or to make their job “work” for them and their life, workers in lowwage jobs are likely to be more reliant than average on standards set in labor regulation at multiple levels of government, or on aspects of the welfare state. Therefore, when we consider these types of jobs in comparative perspective, we emphasize the role of national institutions highlighting contrasts in their role, function, and “bite” across countries. For example, in France collective bargaining, even if a strong institution in the country as a whole, plays a weaker role than expected in retail trade because the sector has developed ways to bypass some mandates and because it is a setting where enforcement of agreements (and even labor standards) is difficult. Only some strong aspects of labor regulation like the minimum wage and hours regulation have a significant “bite” (Askenazy et al 2008).
Already, Gautié and Schmitt (forthcoming, intro) have provided a macro-economy argument for how the volume of low-wage work
2
in a country is affected by the degree of inclusiveness of national labor and social insurance institutions. A correlate to inclusiveness is the degree of availability of exit options from the national labor regulation and social insurance standards. Here we argue that key dimensions of job quality, not only the volume of low-wage work, are affected by the inclusiveness of
2 Defined as two thirds of the economy-wide hourly median.
5

national institutions. But the focus on a particular sector and jobs inherent in meso analysis warrants moving beyond the broad characterization of national institutions to explore how multiple spheres/domains interact with employer strategies.
Second, we are also more likely to give relatively more weight to institutions and social norms that constitute the reproductive sphere (as compared to labor regulation and firm strategies). The performance of low-wage work—particularly low-wage service work— is less determined by labor-market institutions (e.g. training or labor standards) than other work and more affected by other domains particularly the reproductive domain (gender roles and responsibilities in the household, welfare state institutions, customary practices). It is always true that the performance of entry-level work relies heavily on skills and behaviors acquired outside the labor market (e.g. garment shops rely on sewing skills acquired in the household). But we argue this pattern is more marked with lowwage service work because the work process is less structured through technology (less machine pacing of the work and task organization). Also, while skills and behaviors acquired outside the labor market are important in the performance of skilled service work as well, knowledge and training matter far more in these jobs.
For example, in retail, desired behavior and social skills (politeness, friendliness) are recruited for and expected to be developed outside labor market institutions. Behavioral requirements of reliability, timeliness, ability to adjust to fluctuating hours, and adaptation to short work hours are deemed to stem in part from personality factors developed outside the workplace. But they are also perceived to stem in significant part from characteristics of the worker’s household, or education commitments, or to be facilitated by national institutions that affect the reproductive sphere (e.g. access to child care, after-school care, or maternity leave). There are exceptions; German mini-jobs are a viable cost saving option for retailers because the women who hold them often are graduates of the vocational system; in this case the behavioral requirements are met through a labor market institution (apprenticeship
We discuss what we have learned about conducting cross national meso analysis of lowwage jobs using retail trade as an example. We also discuss the unresolved dilemmas we have encountered and raise questions to be addressed in future work.
A common approach of economic analysis would typically seek to explain the separate effects of labor supply, labor demand, and institutions on job outcomes. However, we argue that labor supply and labor demand are so crucially shaped by institutional context that this exercise would be misleading. Here we briefly lay out an exposition of labor
6

supply and demand in theory , then suggest how thoroughly social institutions condition these nominally economic forces.
In a simplified economic model, the demand for any particular type of labor is a derived demand: that is, it channels final demand for goods and services. Thus, it is the outcome of consumer preferences, available technologies of production, workers’ potential productivity as a function of their skills and capacities, and the market power of businesses as sellers of goods and as buyers of labor. Similarly simplified, labor supply results from demography (population size, age composition, family structure, immigration) and the preferences of potential workers (in particular, the perceived tradeoff between goods and leisure).
But as a moment’s reflection reveals, it is not meaningful to describe these factors determining labor supply and demand in some “pre-institutional” way. Consumer preferences, an element of Gadrey’s market structure sphere, are molded by norms of prestige and emulation. Choice of technology and level of worker productivity stem crucially from businesses’ choices about the organization of work (Gadrey’s organizational structure sphere), choices that involve constant consideration of institutional strictures (the sphere of industrial relations) and “exit options” from those strictures, such as union avoidance or outsourcing. Thus, although part of the explanation for U.S. retailers’ heavy use of short-hour part-time workers is the goal of matching staffing to fluctuating consumer demand, a large part of the explanation, especially in comparative context, is the lack of restrictions on store opening hours and worker schedule variations, and the low level of worker protection, which allow retailers (and others) to construct part-time work as a second-class status receiving different hourly wages and fringe benefits. Skill development institutions differ greatly across nations.
Market power reflects both business structures (corporate governance, systems of finance, nature of business associations) and government regulations (antitrust, but also a variety of other levers: for example, land use practices condition retailers’ market clout), thus building on both the market structure and the organizational structure spheres. The availability of workforce groups depends on institutional factors in the domestic sphere such as the availability of child care, requirements and norms regarding schooling, and so on. In short, employers and workers do weigh their tradeoffs, but in ways that are thoroughly structured by institutions and norms.
As noted, in the simplest sense, institutions include laws and regulations, and labor relations institutions (unions and other representative organizations, employer associations, and levels of collective bargaining). They also include institutions that regulate the reproduction of the workforce most notably education systems and welfare state institutions. Finally, social norms and customs, whether encoded in law or recognized practice, play an equivalent role. Here we discuss, direct and indirect institutional effects.
7

Direct institutional effects:
When assessing the role of institutions on low-wage jobs, we first inquire about institutions of labor regulation. There are direct institutional effects stemming from these meaning institutions play a role with direct impact on job quality per se rather than by altering the decision terms for firms, for example.
The comparison of retail jobs in the U.S. and five European countries provides ample examples of these direct institutional effects. For example, the U.S. does not mandate hourly wage and benefit parity between full-time and part-time workers whereas five
European countries do. Distinctly national institutions play a significant role as is best illustrated when we contrast France with the U.S. France has set a high relative value for the minimum wage (SMIC) whereas the U.S. has failed to raise the minimum wage over long periods of time. In 2006, the French minimum wage stood at 68% of the median wage. In contrast, the U.S. minimum wage has lost real value over the past 30 years under intensive lobbying by industries with low-wage jobs. By early 2007, the real minimum wage had declined to only 30% of the median wage.
3
The minimum wage affects the situation of retail workers in two ways: it affects both entry level wages and the reservation wage. In turn, these differences have much to do with the fact that 42% of retail workers were “low-wage” in the U.S. as compared to 18% in France for 2002
(Carré, Tilly, vanKlaveren, and Voss-Dahm forthcoming).
Strong union representation with active collective bargaining clearly has an impact as well. With European colleagues, we contrast the role of collective bargaining in Denmark where agreements enforce significant wage compression that contributes to the small amount of low-wage retail employment in that country (Esbjerg et al 2008). Collective bargaining coverage was 69% in 2004-2006 (Tijdens et al, 2007). This direct effect has ripple effects as well. For example, the terms of collective agreements are adhered to by those retailers that have not signed it. In contrast, in the U.S., retail collective bargaining coverage is very low (less than 6% in 2007; U.S. Bureau of Labor Statistics 2008).
Unionization clearly impacts compensation for unionized US retail workers but the union threat effect is minimal.
When comparing the situations of U.S. retail workers to those in five European countries we were also struck by the wide disparity of scheduling practices, in particular advance notice and the degree of control of workers over their schedule. Even while the U.S. and
European countries retailers share some practices, such as heavy reliance on part-time, worker day to day schedule predictability and control nevertheless varies substantially.
For example, in Germany and Denmark, collective bargaining agreements require significant advance notice of schedules in retail trade (26 and 16 weeks respectively).
Even while practice falls short of these requirements, retail workers in these countries have far greater advance notice than U.S. retail workers. In the U.S., in contrast, there is
3 Authors’ calculations from US Department of Labor 2008, US Bureau of Labor Statistics 2008b, and CPI-
U index.
8

no mandated notice period. U.S. retail workers typically receive notice of their schedule
3 days to 2 weeks in advance.
Mexican supermarkets offer yet another example. U.S. retailers have escalated part-time employment for three purposes: to match staffing to customer flows hour by hour, to cover weekend shifts, and to take advantage of lower wages and benefits offered to parttimers. But in Mexico, the minimum wage is set per day, making short shifts costly; the normal retail work week is six days of eight hours, making weekend coverage relatively easy, and most benefits are universally mandated by law, limiting possible savings in compensation through creation of part-time jobs. This policy results in a low use of retail part-time as compared to other countries, with virtually no part-time employment in
Tilly’s sample of retail businesses.
Indirect institutional effects
Institutions affect outcomes in low-wage jobs most often in indirect ways. What we mean by indirect institutional effects, are those affecting behavior of key actors— employers or workers rather than job outcomes per se. Indirect effects occur primarily as institutions—rules, regulations, organizational arrangements, collective bargaining agreements—also interact with social norms and cultural practices, factors that are less formalized than regulations but are impactful all the same. Through these interactions, institutional effects are mediated or reinforced (or occasionally weakened).
Importantly for job outcomes, indirect institutional effects always interface with, and are shaped by, labor supply and demand dimensions, as noted above. We provide examples of indirect institutional effects that highlight the interactions of institutions and social norms while being cognizant of the contextual factors (institutional, organizational and cultural) that affect labor supply and demand.
In short, societal norms have bearing on how labor market institutions affect jobs indirectly. In retail, norms that govern shopping hours and the frequency of shopping have clear impacts on worker schedules as well as overall staffing patterns. For example,
U.S. limited regulation of work conditions combined with social norms have driven 7-day
24-hour operation for supermarkets, a pattern that drives the high use of short-hour parttime, the pressure on full-time and part-time workers to be available as needed, unpredictable schedules, and a generalized pressure on wage levels. In five European countries, cultural norms, that have become encoded in regulations or collective bargaining, have so far yielded opening hours that are more circumscribed (restricted
Sunday openings, no instances of routine 24 hour opening) resulting in less scheduling pressure on workers, less pressure on wages but, in some countries for example France, concentrated rush shopping hours and thus great workload pressure on cashiers.
In a direct comparison of cashier jobs in France and U.S. supermarkets, we and coauthors found that the French women’s preference for full-time work—reinforced by institutions such as child care that make it feasible– and the lack of availability of
9

teenagers (who face long hours in school) limit retailers’ ability to staff with numerous workers on very short schedules. The use of short-hour is an important tool for labor deployment and cost management for U.S. grocery stores which face a ready supply of women with children (with child care constraints) and high school students (with lighter class schedules) who seek short work hours. As a result, short-hour part-timers are more prevalent in U.S. retail than in French retail.
4
Cashiers in the two countries also face different ergonomic challenges at work, with U.S. cashiers expected to stand while French cashiers sit (Askenazy et al 2010). Neither position is ergonomically correct (it is best to alternate sitting and standing); each has liabilities for workers and neither position is fully justifiable by operational considerations. Standing helps U.S. cashiers to bag groceries, a task that U.S. shoppers still expect to have done for them (consumer norm) and which is manageable for a mostly part-time and short shift workforce and affordable because of the low pay (Gadrey and
Jany-Catrice 2000). Conversely, the cashier sitting position helps French supermarkets conserve space for checkout stations. In turn, this space constraint is derived from store size regulations and concentrated shopping hours that prompt efforts to reduce checkout lines. Hence, a conjunction of regulatory features and social norms has clear consequences for job quality.
Social norms may have more direct effects on job outcomes
Social norms can affect staffing patterns and job outcomes quite directly. We expect that service work settings are particularly susceptible to these effects because service
“production” occurs at time of delivery (quite unlike good production), and is interpersonal and thus more subject to cultural particularities. For example, Wal-Mart pulled out of two countries, Germany and South Korea, for seemingly contradictory reasons. German consumers viewed “American-style” service interactions in stores as intrusive, while Korean shoppers found that the amount and quality of service was deficient. In other words, the corporate-wide service level was ill-suited to these national service preferences.
Social norms can trigger institutional changes which in turn have job impacts. One example comes from the comparison of the situation of shelf stockers in France and the
U.S.. The strong culture of consumer rights in the U.S. has led to regulatory mandates for individual labeling of products (rather than shelf labeling). This results in significantly greater workload in shelving and a strong pressure to generate large numbers of jobs at low hourly wages.
Direct and indirect institutional effects
While, for purpose of exposition, we illustrate direct and indirect effects for institutions separately, we recognize that in fact institutions tend to cluster (Esping-Andersen 1990,
4 In 2007, in the US, 34% of retail workers work less than 20 hours; as compared to 16% in France
(Askenazy et al 2010).
10

Hall and Soskice 2001). They emerge from distinct societies each with particular priorities and history. To the extent that societies govern themselves coherently, institutions with complementary purposes will tend to cluster.
Any cross national comparison of socio-economic outcomes owes a great debt to early work by Maurice, Sellier, and Silvestre (1986) as well as subsequent work in the same vein. The authors convincingly showed that great caution must be taken in attributing contents, meaning and impacts to institutions bearing the same label and, ostensibly, playing the same role in different national settings. They argued in particular that education and training institutions and their relationship to occupational definitions and stratification are deeply enmeshed with other societal effects on work organization and thus their import and impacts vary in form and extent across countries. Depending on the configuration of national institutions, skill formation in the same sector can encompass classroom education, formal apprenticeships, or on-the-job training, with very different implications.
This perspective is useful, but challenging to apply with multiple countries. This is particularly true when, in considering low wage service work in cross national comparison, we are required to consider a broad range of institutions, moving far beyond labor market institutions. As we have noted, institutions from the reproductive sphere (the family and social protection and other welfare state institutions) and social norms have greater relevance to job outcomes. These vary greatly across countries, further complicating the analysis. For example, French high schools require about 50% more hours per year than U.S. ones.
One striking example stems neither from training structures nor from the reproductive sphere, but from that core workplace institution, the union. French union membership is low, and so is U.S. union membership. But the French industrial relations environment is one in which national union federations achieve gains in labor market regulation (most notably the minimum wage) that benefit all workers in entry level jobs. Additionally, there is sectoral bargaining (though in retail, national standards matter more than bargaining agreements). In the U.S., low union membership is very closely linked to low collective bargaining coverage and, indeed, results in limited union threat effect, limited advantages to unionized workers, and low wages throughout the industry.
Naturally, employers and workers do not simply passively observe and react to the institutions that shape labor outcomes. On the one hand, large corporations or employer associations actively seek to alter, evade, or game institutions to exert more control over
11

their environment while, on the other hand, unions and worker organizations seek to resist these moves. Thus, both groups are strategists. Differing strategies are particularly important in explaining variation within countries, but particular strategies also have varied degrees of success in different institutional environments. Here we particularly focus on employers, who generally dispose of more resources in this contest.
A common channel for strategic action by employers is innovation, both in technology and in management technique. Wal-Mart provides examples of both. Because of its mastery of advanced logistical technology, Wal-Mart was able to become dominant in the
United States and Mexico via rapid organic expansion, altering market structure and evading antitrust regulation in both countries (Wrigley 2002, Tilly 2006); in fact, an initiative by Wal-Mart de México’s three largest competitors to form a purchasing consortium was delayed for months by Mexico’s Federal Competition Commission, despite the fact that Wal-Mart was at the time larger than the three competitors combined
(Castillo 2005). Similar attempts at dominating through organic growth were not as successful in smaller countries distant from existing Wal-Mart footholds and with established national chains, such as Germany and South Korea, both of which Wal-Mart left in 2005 (Tilly 2007). On the side of management technique, Wal-Mart developed a low-compensation, paternalistic model incorporating evangelical Christian notions of
“servant leadership” in white rural areas of the U.S. South, which has been highly successful in repelling attempts to unionize in the U.S., but far less so in some other countries (Moreton 2009, Tilly 2007). Of course, worker organizations also strive to develop more effective organizational models in order to reshape industrial relations, and the geographical unionism of the U.S. Service Employees International Union or the militancy of Brazil’s Central Única dos Trabalhadores have reshaped jobs as well.
Sometimes a key innovation simply involves newly expanded use of an industrial relations exit option, paving the way for adoption by other sectors. Thus, in the United
States, retailers’ tilt toward short-hour part-time employment served to evade the norms associated with full-time work, and to evade the unions focused on serving full-time workers. In Germany, likewise, retailers’ rapid expansion of short-hour “mini-jobs” has allowed them to evade normative and collective bargaining restrictions as well as achieve savings from tax exemptions. In both cases, organizational innovations pioneered in retail have subsequently diffused to other sectors. Another, more controversial example of which Wal-Mart is one of a long series of examples, is companies that set goals and evaluation criteria that essentially compel managers to violate labor laws (Rose 2005).
Most fundamentally, businesses have a market structure choice among competitive strategies: “low road” strategies based on keeping costs and prices low, “high road” strategies based on better service, quality, or innovation, and a variety of more specific configurations. In U.S. retail, Wal-Mart’s low road approach is well known. Among
Wal-Mart’s competitors, we found many attempts to develop some variety of the high road via better meats and produce, wider variety, or better service—though in most cases this was coupled with cost-cutting in ways that often undermined attempts to upgrade.
Other, less common strategies included achieving greater operational efficiency through a highly paid, experienced, and highly productive workforce, and locating in small rural
12

communities underserved by other major retailers. In France, in contrast, the high minimum wage forecloses low-wage competition, and competition based on efficiency or the exploitation of geographic monopolies is far more common (Askenazy et al.
2009).
Institutional configurations thus channel competitive strategies, but do not fully determine them.
At a certain point, strategy can move from artfully responding to institutions, to seeking to shape the institutions themselves. In the United States, associations of retailers, fast food restaurants, and other businesses fight for a lower minimum wage; unions fight for a higher one. Unions and their allies seek to use land-use restrictions and approval processes to set required labor standards for new stores (again, Wal-Mart is a major target); expanding retail chains seeks to head off such standards.
Especially in high employee turnover models like retail in the United States, employers can also shift industrial relations norms via concerted action, alienating one cohort of employees but altering the expectations of the next. Thus, in our U.S. retail fieldwork, we found that despite reported anger and distress at the point when leading consumer electronics retailers eliminated commission-based compensation, current cohorts of employees see no disadvantage to working on salary instead of commission. Similar processes are at work in the gradual reduction and elimination of premia for evening, night, and Sunday work. This U.S. experience contrasts with Western European examples, where a more long-term workforce and widely present unions with institutional memories have maintained commission systems and defended differential pay for unsocial hours (Carré, Tilly, van Klaveren, and Voss-Dahm forthcoming).
Furthermore, these economic actors can have an impact on consumer norms, once more altering the market structure. The economic crisis has led retail customers to focus much more on price, but the degree to which that will persist as economies improve will depend significantly on the marketing success of differing business models (Planet Retail 2010).
Meanwhile, the major U.S. unions have sought, with some success, to make the treatment of workers a criterion in consumers’ choice of where to shop.
We have reviewed how we might apply a cross national meso analysis framework to understanding effects of institutions, norms, and employer strategies on job outcomes. In this section, we explore issues that have arisen for us in the course of this exercise and that remain challenges for analysis to address. We also explore how including a middleincome country into a framework that was designed for comparing high income countries alters how one might use a meso analysis framework.
High road and low road . Job characteristics and job quality differ within countries as well as across them, and even within local labor markets governed by the same local and
13

national institutions. One reason for this variation is simply ability to pay. But we and other researchers have long argued that businesses also differ in ways captured by the distinction between the “high road” based on better compensation, more investment in skills as well as greater productivity; and the “low road” based on low pay, low skills, low productivity, and often high turnover (see, for example, Appelbaum, Bernhardt, and
Murnane 2003). Cross-national comparisons should allow added perspective on the degree to which this dichotomy corresponds to workplace reality. What are the circumstances that permit better outcomes, or not? Addressing within country differences, some researchers (e.g. Cappelli and Crocker-Hefter 1996) have maintained that distinct market niches are often important in underpinning these differences, and we have argued that this is an important element of the oft-cited difference in job quality between Wal-Mart and Costco (Carré, Tilly, and Holgate 2006). Are the differences large enough to be interesting, or is it more accurate to characterize job variation as a spectrum or scatter than as a distribution segmented between high road and low road?
And how do these patterns differ by sector; is the story different for low-wage service jobs than for higher-wage manufacturing ones?
Furthermore, when assessing the sources of within country differences, we draw distinctions between factors that appear to be market related (market segment as noted, but also ownership type, regional market characteristics) and those that directly stem from labor deployment strategies as part of a “high road” approach. These latter factors aim to isolate, and highlight, the degree of management autonomy in a given market context. How should we decide which of these factors to consider when drawing cross national comparisons when the tool of international comparison is firm-level case studies? To what extent are these categories comparable across nations? As sources of variation proliferate, to what extent can a relatively small number of cases effectively inform us about the respective impacts of institutional environment, firm market position, and discretionary managerial strategy?
Resilient social norms? Our discussion of the role of social norms and customers, and particularly our accounts of how they interact with institutions and contribute to the latter’s indirect effects on firm behavior and job outcomes, provide explanations of where things are, and how they got to be this way. There remain ambiguities about the predictive capability of this approach. Of course, history matters either through complex influences or simply path dependency. Nevertheless, accounts that give importance to social norms and customs—particularly those not encoded in law—must be held to special scrutiny. Social norms and customs can be eroded, amended or even overthrown at the instigation of employers and their practices. While change in social norms is expected progressively over time, sometimes it can occur surprisingly fast.
Thus we ask: what is the predictive power of analyses that hinge on the interaction of institutions and social norms? Do such analyses help reasonably predict what will happen to jobs? Of particular interest are the social norms that safeguard key aspects of job quality. For examples, the large French supermarkets have a practice—some would argue a policy— that their own workers, including part-time cashiers, ought to receive enough compensation to enable them to shop in the store. The motivation is
14

instrumental, and in keeping with a fordist political economy regime. But is also a cultural norm. How resilient are norms such as this one that safeguard job quality in the face of forces that erode job quality? How resilient are such norms in a national context under consideration? And when looking across national settings, how can we tell where protective social norms will remain strong in the foreseeable future? What else should be assessed in order to make a reasonable prediction? The literature examining job erosion has focused on economy-wide changes in areas where job quality is buttressed. So in the
U.S. example, accounts have focused on the loss of labor power (because collective bargaining is a primary way for job improvement) and diminished union threat effect.
The norm of Sunday work pay premium was first eroded in non-union stores then unionized stores. As we have noted all along, retail is one of the sectors where labor standards are subject to significant challenges (difficulty of enforcement, industry lobby for deregulation). Yet, in some countries retail has been able to maintain certain practices that protect job quality, while in others it has not. We would want to investigate how great competitive pressures need to be in order for long-standing and well established social norms to cave in. Do countries fare better where there is a good “fit” among different areas of regulations and with existing norms? And do retail worker fare least well in countries where there is the greatest lack of fit?
Industrial convergence reloaded. Debates about the likelihood of the convergence of the nature of work in industrial societies date back at least to the 18 th
century and Adam
Smith’s
Wealth of Nations . Despite the appeal of convergence arguments in simplified neoclassical economic models and the popular press (Friedman 2007), researchers have consistently found that national institutions do indeed put work on divergent paths (Kerr et al.
1964 and 1973, Maurice, Sellier, and Silvestre 1986, Hall and Soskice 2001a,b).
Thus, our emphasis on the importance of institutions is certainly not a new one.
However, our focus on consumer services, and retail in particular, as distinct from the manufacturing adds some new twists to the argument. On the one hand, it is hard to think of anything more culturally embedded than the merchandise offered by retail and consumer services: food, clothing, prepared meals, hair styling, and the like. Moreover, such services are predominantly place-based: while apparel or food can be manufactured distant from the consumer by workers unfamiliar with local tastes and customs, it continues to mostly (despite the online sales exception) be sold via direct interaction with the customer in ways that assume knowledge of those tastes and customs. This portion of the “technology” of retail sales is less likely to homogenize than manufacturing production processes. While all apparel manufacturing in the world could, in theory, concentrate in China, the same will never be true of grocery sales or restaurants. On the other hand, the global penetration of giant retailers (starting with Wal-Mart, Carrefour,
Metro, Tesco, which between them sell in over 50 countries containing the great bulk of the world’s population) and even more far-flung fast food chains such as McDonald’s and KFC (117 and 109 countries, respectively), and the rapid diffusion of technological innovations in retailing and food preparation (especially in logistics and point-of-sale), suggest powerful forces for convergence. So a question worthy of further exploration is just how the convergence-divergence dialectic differs for place-based consumer services.
15

Although we have drawn on Tilly’s research on research in Mexico, for the most part our discussion is grounded in research, ours and others’, about low-wage jobs in wealthier countries with broadly similar industry structures and predominant store formats.
Extending the analysis to middle income and poorer countries shifts the terrain. Here we suggest three promising directions for charting that shift (some of this draws on Tilly
2002).
First, market and cultural differences have a bigger impact (compared to regulatory institutions) when markets and cultures are more different. For example, lower-overhead, lower-technology segments of the market are much more important in Mexico than in the
U.S. or Europe. In the U.S., a street market is an exotic option (farmer’s market, flea market, yard sale). In Mexico and many other countries, street markets constitute a major alternative, and in some geographic areas the major alternative. U.S. and European supermarkets and hypermarkets take for granted a system of highways and trucking, the availability of car transport to customers, far-reaching print and broadcast media for advertising—all of which apply in some parts of Mexico, but not others. As an example of cultural differences, consider the frequency of shopping for food: In Mexico, the average family shops for food 11 times per week ( American Demographics 1995), compared to 4 times for the average U.S. family ( Progressive Grocer 1999) reflecting different valuation of “freshness” and varied perspective on shopping (utilitarian vs. a form of socializing). The value placed on freshness in Mexico means that for many families, only tortillas purchased hot from a shop within walking distance will do (and tortillas may be served at more than one meal per day).
Second, small individual proprietorships, which are far more dominant in lower-income countries, function quite differently from large corporate business, resulting in a much more influential role for the family as an institution. For example, as regards hiring and compensation, individual proprietors typically rely on themselves and family members
(often unpaid) for core roles. In terms of work schedule, whereas U.S. retailers use parttime work to match staffing to customer flows, self-employment in Mexico and poorer countries often drives longer hours rather than shorter ones. Upward mobility in largescale retail means moving up in management within or across companies; for a sole proprietor it means growing the business: adding more merchandise or shifting to more profitable lines, moving to a larger space, adding locations, automating.
Third, informal employment (lack of de facto coverage by employment laws) and informal enterprise (non-registration in official systems of taxation and regulation) loom far larger in poorer countries. This is correlated with small size, but the relationship is not one-to-one (one interesting counter-example is Brazil, where according to McKinsey
[2003] some technologically modern, large-scale retailers avoid paying taxes). Relatively recent independence, revolution, or major reform in many poorer countries have often brought with them ambitious new laws and constitutions. But given low income levels and limited state capacity, the laws may be more aspirational and less real than in richer countries. In particular, many poorer countries have an extremely elastic supply of
16

desperate labor (often migrants from rural areas) that feeds informal activity. Retailing and food preparation have particularly low barriers to entry due to low capital requirements and widespread possession of the basic necessary skills. Avoidance of laws can be a viable strategy for many businesses, especially smaller ones. This includes tax evasion, flouting of labor regulations, and the sale of contraband, knock-offs, and pirated electronic media. (However, it is also important to reiterate that even in rich countries, retail is a sector in which businesses often dodge or ignore labor regulations.) Evading labor regulations has a direct impact on job characteristics, but the other informal practices also have an impact, albeit less direct by facilitating the growth of market segments that are much smaller in rich countries. Indeed, some have argued, in this kind of environment, institutions such as laws or unions may have a larger impact by creating an incentive for businesses to pursue the informal path, than they do by directly regulating firm behavior. On the other hand, welfare programs such as Mexico’s
Oportunidades or Brazil’s Bolsa Familia may have a significant impact on the attractiveness and viability of informal enterprises.
In short, extending the analysis to countries at lower income levels has two broad effects.
One is to introduce wider cross national variation in market structure and cultural norms, rendering these sources of variation in job quality more relevant. The second is the greater weight of smaller and more informal businesses in less wealthy countries. This, in turn, means that alternative systems of governance, such as the family, associations of informal businesses, local patronage, or organized crime, loom larger.
We, like others, show that national institutions continue to matter to job outcomes when we consider specific jobs in particular sectors. National institutions governing labor and industrial relations but also those governing the reproduction of the workforce, matter when comparing job outcomes cross nationally. This remains true even while we recognize that job outcomes ultimately result from the interactions of employer strategies with multiple national institutions and norms. (Conversely, employer strategies matter most in accounting for within-country variations in job outcomes—albeit within a band of variation that bears the imprint of institutional effects.)
In thinking through institutional effects in cross national comparisons of job outcomes we have focused on low-wage service work, using entry-level jobs in retail trade as examples. We find that low-wage job outcomes are more susceptible to cross national differences in labor market regulations than when we consider other kinds of jobs. If generalization is possible, low-wage workers depend more on national institutions than on employer practices or organizational characteristics for improvements in wages and working conditions. We also find that because service is “produced” partly in person, characteristics of service work are more prone to be affected by cross national differences in norms, for example those regarding service style and intensity..
17

With this paper, we have also used analytical tools from a literature that has almost exclusively looked at comparisons of wealthy countries to begin to compare jobs in middle-income countries with those in wealthy countries. As comparisons of retail jobs in the U.S. and Mexico attest, differences in the organization of product markets (such as ownership structure, or the mix of formal and informal enterprises) matter much more because they are greater. Also cultural norms that manifest themselves in family structure and labor market participation, or those governing shopping patterns, also matter more in contrasts because these dimensions are far more enmeshed in economic behavior in countries with informal enterprises than they are in countries whose economic activity is wholly governed by corporate structures.
American Demographics.
1995. “Food shopping trends in Mexico.” December supplement, 17:10.
Appelbaum, Eileen, Annette Bernhardt, and Richard J. Murnane (Eds.). 2003. “Low wage America: an overview” in
Low-wage America: How employers are reshaping opportunity in the workplace, New York, NY: Russell Sage Foundation, pp. 1-29.
Askenazy, Philippe, Jean-Baptiste Berry, Françoise Carré, Sophie Prunier-Poulmaire, and
Chris Tilly. 2009. “Working in large food retailers: A France-United States
Comparison.” Presented at MIT Institute for Work and Employment Research Seminar,
March 10.
Bamber, Greg J., Russell D. Lansbury, and Nick Wailes. 2004. International and
Comparative Employment Relations: Globalisation and the Developed Market
Economies, 4th ed .
London: Sage.
Bank Muñoz, Carolina. 2008. Transnational Tortillas: Race, Gender, and Shop-Floor
Politics in Mexico and the United States . Ithaca, NY: Cornell University Press.
Baret, Christophe, Steffen Lehndorff, and Leigh Sparks, eds.. 2000. Flexible Working in
Food Retailing: A Comparison between France, Germany, the United Kingdom, and
Japan . London: Routledge.
Barro, Robert J., and Xavier Sala-i-Martin. 1999. Economic Growth . Cambridge, MA:
MIT Press.
Bazen, Stephen, Claudio Lucifora and Wiemer Salverda, eds. 2005. Job Quality and
Employer Behaviour. Basingstoke, UK: Palgrave Macmillan.
Bosch, Gerhard and Steffen Lehndorff, eds. 2005a. Working in the Service Sector: A Tale from Different Worlds . London: Routledge.
18

Bosch, Gerhard and Steffen Lehndorff. 2005b. “Introduction”, pp.1-31 in Bosch and
Lehndorff 2005a.
Cappelli, Peter, and Anne Crocker-Hefter. 1996. “Distinctive human resources are firms’ core competencies.” Organizational Dynamics , 3: 7-22.
Carré, Françoise, Chris Tilly, Maarten vanKlaveren, and Dorothea Voss-Dahm.
Forthcoming 2010. “Retail jobs in comparative perspective.” In Jérôme Gautié and John
Schmitt, eds., Low-Wage Work in The Wealthy World. New York: Russell Sage
Foundation.
Carré, Françoise, Brandynn Holgate, and Chris Tilly. 2006. “What’s happening to retail jobs? Wages, gender, and corporate strategy.” Paper presented at the Labor and
Employment Relations Association Annual Meeting, January 5-8, 2006.
Castillo Mireles, Ricardo. 2005. Across the border: Taking it to the competition,
Mexican style. Logistics Today , February.
Dieckhoff, M. 2008. “Skills and occupational attainment: A comparative study of
Germany, Denmark, and the UK.”
Work, Employment and Society 22(1):89-108.
Esbjerg, Lars, Klaus G. Grunert, Nuka Buck, and Anne-Mette Sonne Andersen. 2008.
“Working in Danish retailing: Transitional workers going elsewhere, core employees going nowhere and career seekers striving to go somewhere.” In Niels Westergaard-
Nielsen, ed., Low-Wage Work in Denmark. New York: Russell Sage Foundation, 140-
185.
Esping-Andersen, G. 1990. The Three Worlds of Welfare Capitalism . Cambridge: Polity
Press.
Ferner, Anthony and Phil Almond. 2007. “Managing people in U.S.-based multinationals: The case of Europe.”
Perspectives on Work 11(1): 4-7.
Friedman, T.L. 2007. The World Is Flat: A Brief History of the Twenty-first Century .
New York : Farrar, Straus and Giroux.
Gadrey, Jean. 2000. “Working time configurations: Theory, methods, and assumptions for an international comparison,” pp.21-30 in Baret, Lehndorff, and Sparks 2000.
Gadrey, Jean and Florence Jany-Catrice. 2000.” The Retail Sector: Why so many jobs in the U.S. and so few in France?” Service Industries Journal . 20 (4): 21-32.
Gash, V. 2008. “Preference or constraint? Part-time workers’ transitions in Denmark,
France, and the United Kingdom.”
Work, Employment and Society 22(4):655-674.
19

Hall, P.and D. Soskice. 2001a. “An introduction to varieties of capitalism.” Pp.1-68 in
P. Hall and D. Soskice, eds. 2001. Varieties of Capitalism : The Institutional Foundations of Comparative Advantage . New York: Oxford University Press.
Hall, P.and D. Soskice, eds. 2001b. Varieties of Capitalism : The Institutional
Foundations of Comparative Advantage . New York: Oxford University Press.
Hult, C.and J. Edlund. 2008. “Age and labour market commitment in Germany,
Denmark, Norway, and Sweden.”
Work, Employment and Society 22(1):109-128.
Kalleberg, Arne. 1988. “Comparative perspectives on work structures and inequality.”
Annual Review of Sociology 14: 203-25.
Katz, Harry C., ed. 1997. Telecommunications: Restructuring Work and Employment
Relations Worldwide . Ithaca, NY: Cornell University Press.s
Katz, H. C. and Darbishire, O. (2000) Converging divergences: Worldwide changes in employment systems (Ithaca N. Y.: Cornell University Press).
Kerckhoff, Alan C. 2005. “Institutional arrangements and stratification processes in industrial societies.”
Annual Review of Sociology 15: 323-47.
Kerr, Clark, John T. Dunlop, Frederick H. Harbison, and Charles A. Myers. 1964 and
1973. Industrialism and Industrial Man: The Problems of Labor and Management in
Economic Growth , 2nd ed. New York: Oxford University Press.
Kochan, Thomas A., Russell D. Lansbury, and Jean Paul MacDuffie, eds. 1997. After
Lean Production: Evolving Employment Practices in the World Auto Industry . Ithaca,
NY: Cornell University Press.
Maurice, M., F. Sellier, and J-J. Silvestre. 1986. The Social Foundations of Industrial
Power . Cambridge, MA: MIT Press.
McKinsey Global Institute. 2003. New Horizons: Multinational Company Investment in
Developing Economies . October.
http://www.mckinsey.com/knowledge/mgri/reports/newhorizons.asp
Mason, Geoff and Matthew Osborne. 2008. “Business strategies, work organization, and low pay in UK retailing.” Chapter 4 in Caroline Lloyd, Geoff Mason, and Ken Mayhew, eds., Low-Paid Work in the U.K.
. New York: Russell Sage Foundation: 131-67.
Moreton, Bethany. 2009. To Serve God and Wal-Mart: The Making of Christian Free
Enterprise.
Cambridge, MA: Harvard University Press.
Planet Retail. 2010. “Winning strategies for 2010: The new economic reality.”
Webinar, January 27. http://www.planetretail.net
20

Progressive Grocer.
1999. “Changing channels.” December, p.78.
Rose, Steven. 2005. “Behind the smiley face veneer: An insider’s observations of Wal-
Mart’s anti-labor practices and criminal behaviors.” Unpublished paper, Indiana
University-Purdue University Fort Wayne, Department of Sociology and Anthropology,
May.
Samuelson, Paul. 1948. “International trade and the equalization of factor prices.
Economic Journal 58, 163–84.
Solow, Robert. 1956. “A contribution to the theory of economic growth", Quarterly
Journal of Economics 70(1):65–94.
Tilly, Chris. 2007. “Wal-Mart and its workers: NOT the same all over the world.”
Connecticut Law Review , May.
Tilly, Chris. 2006. “Wal-Mart goes south: Sizing up the chain’s Mexican success story.”
Pp.357-368 in Stanley Brunn, ed., Wal-Mart World . New York: Routledge, 2006.
Tilly, Chris. 2002. “How retail industry restructuring affects the workforce: A U.S.-
Mexico comparison-in-process.” Unpublished paper. Department of Regional Economic and Social Development, University of Massachusetts Lowell.
Tilly, Chris. 1997. “Arresting the decline of good jobs in the U.S.A.?”
Industrial
Relations Journal , December.
World Bank. 2010. “Gross national income per capita 2008, Atlas method and PPP.” http://siteresources.worldbank.org/DATASTATISTICS/Resources/GNIPC.pdf .
Accessed February 13, 2010.
Wrigley, Neil. 2002. “Transforming the corporate landscape of U.S. food retailing:
Market power, financial re-engineering, and regulation.” Tijdschrift voor Economische en Sociale Geografie, 93(1): 62-82.
21