Briefing Paper #170
The economic impact of local living wages
by Jeff Thompson and Jeff Chapman
The modern living wage movement was born in Baltimore in 1994, when the city passed an ordinance requiring firms to pay employees a rate above the minimum wage while working on city contracts. Since then, over 120 communities have followed suit, some setting wage floors more than twice the federal minimum wage, and some requiring various benefits.
The astounding growth of the living wage movement has been a response to the predicament of Americans who work but are unable to make ends meet, as well as to the public policies contributing to the problem.
Public policies have exacerbated the problem from the federal level to the local level. Since the early 1980s, the federal government has generally neglected the minimum wage; by 2005, a minimum wage paycheck bought less than it had in 49 of the last 50 years. Local governments have contributed to the problem, following the trend of cutting costs by contracting out services to firms who frequently pay lower wages and offer fewer benefits than public employment. Too often, economic development efforts have channeled public funds in the form of tax breaks or tax incentives to businesses without regard to the quality of the jobs those businesses provide.
As a result of these policies, the two most common themes echoed by living wage proponents are (1) that wages should be high enough to allow workers to meet basic needs (i.e., “living wages”), and (2) that municipal policy should encourage or require living wages for its employees and contractors, rather than exacerbate the problems faced by low-wage workers.
Despite having common goals, living wage laws vary considerably in practice. Most cover employees working under municipal contracts. Some also cover municipal employees, employees of businesses receiving public economic development dollars, or employees of businesses located in districts that have benefited from significant public investment. Wage levels vary from one dollar above the federal minimum wage to over twice the minimum. Some exempt nonprofit organizations, while others primarily affect human service providers.
One characteristic most share is considerable scrutiny—by pushing for higher wages and challenging the way municipal governments operate, living wage policies have generated significant interest from many different parties. One of the chief concerns among all observers has been the economic effects for municipalities, workers, and firms.
Using the growing body of research that has empirically determined the actual effects of living wage policies, this study shows that:
Living wage laws have small to moderate effects on municipal budgets.
• A detailed survey of 20 cities found that the actual budgetary effect of living wage laws had been consistently overestimated by city administrators; actual costs tended to be less than one-tenth of 1% of the overall budget.
• Two separate studies of the Baltimore living wage found that city contract costs increased less than the rate of inflation.
• A study of the Los Angeles ordinance found no measurable effect on the city’s fiscal health.
• A study of living wage ordinances in three New England cities found that contract costs only rose in one city.
• Multiple studies have shown that the bidding for municipal contracts remained competitive or even improved as a result of living wage ordinances.
Living wage laws benefit working families with few or no negative effects.
• Recent studies using original surveys in both Los Angeles and Boston have shown that the workers affected were mostly adults and mostly working full time.
• Both the Boston and Los Angeles studies also showed that most living wage workers were in households struggling to meet a basic-needs budget.
• In Baltimore and Boston, empirical studies have found no evidence of diminished employment.
• In Los Angeles, surveys of workers and firms show that job losses affected just 1% of workers getting a raise.
• Two studies of San Francisco living wage policies found employment increased among airport workers and home health care workers.
• An exception to the general conclusion of research on living wages is a series of studies by David Neumark and Scott Adams that estimate relatively large wage gains and employment losses. The method of these studies has been severely criticized, and the findings discredited by many researchers.
Living wages laws have raised productivity and decreased turnover among affected firms.
• Multiple studies of Baltimore, Boston, Los Angeles, and San Francisco have shown that firms enjoy lower turnover among employees as a result of the living wage ordinance.
• A study of home-care workers in San Francisco found that turnover fell by 57% following implementation of a living wage policy.
• A study of the Los Angeles ordinance found that absenteeism declined, and the decrease in turnover offset 16% of the total cost of the living wage ordinance.
• A study of the San Francisco airport found that annual turnover among security screeners fell from 95% to 19%, as their hourly wage rose from $6.45 to $10.00 an hour.
Effects of living wages on municipalities and consumers
Costs to municipalities
One frequently raised concern is that the cost of the living wage might be passed onto the municipality through higher prices for contracts. If contract prices do increase, the municipal government will be faced with cutting services, raising taxes to pay for the higher costs, finding ways to become more productive, or some combination of the three.
A number of studies have examined changes in municipal contract costs resulting from living wage laws. In general, the evidence from enacted ordinances, as well as the more carefully prepared prospective studies, shows that the overall cost of contracts does not rise significantly.
In 1996, one year after the implementation of the first modern living wage ordinance in Baltimore, the Preamble Center for Public Policy published a study reviewing the fiscal costs of the ordinance. The Preamble study used data on city contracts and interviews with contractors and found that, in the first year under Baltimore’s living wage law, the real cost of city contracts actually decreased. Nominal contract costs rose 0.2%, but after adjusting for inflation costs declined by 2.4%. Expenses associated with implementing the law and monitoring contractors’ compliance were also shown to be minimal, “with the City allocating about 17 cents per person annually for this purpose” (Weisbrot and Sforza-Roderick 1996, 10).
Two years after Preamble’s study, the Employment Policies Institute (EmPI) published a vitriolic response, even charging the Preamble researchers with fabricating evidence in order to reach their desired conclusions.1 Despite its use of heavily charged rhetoric, nowhere does the EmPI study refute Preamble’s key finding—that the living wage ordinance had no discernible impact on contract costs. In fact, the EmPI study does not present a new interpretation of what actually happened to the overall cost of contracts, possibly because their analysis did not yield significantly different results than Preamble’s.2
EmPI’s key accusation was that Preamble “created out of whole cloth a fictitious multi-million-dollar contra
ct.” According to EmPI, the results of the Preamble study hinged entirely on the inclusion of this “whipped-up” contract. Despite EmPI’s apparent attempt to make the casual reader believe that the Preamble authors made up a contract from thin air, the disagreement between EmPI and Preamble was not whether the contract existed, but whether it was in fact an extension of an existing contract and therefore exempt from the ordinance.3 EmPI also accused Preamble of omitting contracts about which information was actually not available at the time of the earlier Preamble study.
In their reply to EmPI’s charges, the Preamble researchers showed that even accepting each of EmPI’s charges does not change the conclusion that Baltimore’s living wage ordinance did not significantly increase contract costs (Preamble 1998).
In 1999, the Economic Policy Institute (EPI) published the third study of the Baltimore experience. Analyzing contracts that could be directly compared before and after the implementation of the ordinance, the EPI research associates from Johns Hopkins University found that the nominal contract costs for the city rose just 1.2%—lower than inflation during the same period—and concluded that the “budgetary impact of the living wage [in Baltimore] has, to date, been insignificant” (Niedt et al. 1999, 6-9). Despite the overall real decline in contract costs during the period under study, there was a range of results for different contract types. Some contracts experienced moderate price decreases, while others grew considerably. The overall price for the heavily effected janitorial contracts, for example, rose 16.6% in nominal terms, with specific contracts seeing price increases ranging from less than 1% to over 50%. The overall budgetary impact of these contracts, however, was negligible as cost increases in other contract areas were more modest. The EPI study’s overall conclusion was that “the widely voiced fear that [the living wage ordinance’s] implementation would place intolerable strains on the city’s budget have not yet materialized.”
The research conducted since these early studies on Baltimore has tended to confirm the initial findings of negligible overall increases in contract costs. In 2003, Andrew Elmore surveyed administrators in 20 cities and counties that had adopted living wage ordinances that had been in force at least one year by late 2001. Each of these municipalities also had the “administrative capacity to produce cost impact estimates, formal internal evaluations, or other empirical assessments of the effects of their laws.” Elmore’s main finding was that in most municipalities “contract costs increased by less than 0.1% of the overall local budget in the years after a living wage law was adopted” (Elmore 2003, 2). Municipalities widely overestimated the costs of the living wage ordinances: the City of Berkeley, California, for example, projected the living wage would result in $479,425 in higher contract costs, but the actual increase turned out to be less than half that amount. Elmore reports that, despite the negligible overall costs of living wage ordinances, in each city there were a few contracts that did experience significant price increases. Predictably, these few contracts were labor intensive operations that employed a large number of workers concentrated at low wages, notably janitorial and security guard services.4
Elmore also found that municipalities that extended their living wage ordinances to cover human services providers, such as home health care or child care, experienced slightly higher contract price increases, ranging from 0.3%-2.8% of local human services budgets (Elmore 2003, 7). These higher (though still modest) price increases could result from a range of factors, not least of which includes large concentrations of low-wage workers and a willingness of municipalities to pick up at least some of the increased wage bill of community-based nonprofit organizations.
An exhaustive study of three New England cities by researchers from the University of Massachusetts confirmed the general finding that modest costs should be expected, but found that one particular type of bidding process that is more likely to lead to cost increases—the use of unit-cost bidding. In Boston and New Haven, Connecticut, falling or stagnant costs for other contracts balanced out increases in unit-cost contracts, resulting in 7% and 11% declines overall, respectively (Brenner and Luce 2005, 25). In Hartford, Connecticut, the ordinance covered only two contracts, both of which used unit-cost bidding, resulting in an overall increase in cost. Unit-cost bidding is used for services such as security or temporary office help, where the city is unable to predict exactly how much of a need they will have in advance. Instead of bidding on the price of providing services for the entire period, firms bid on the price of one hour of service provision. Perhaps because of the clearer connection to the costs of increased wage floors, this practice appears to make it easier for firms to pass on the costs to the city.
The authors also identified a strategy that tends to lead to lower costs—consolidating multiple services into a single contract. They give the example of Multnomah County, Oregon, where consolidating janitorial services at the Department of Corrections, courthouse, and county jail into a single contract saved the county money and may also have improved the firm’s approval rating and turnover rate.
Overall, the study adds more evidence to the finding that living wage ordinances do not put undue strain on city or county budgets. As the director of Boston’s Living Wage Division said, “We also have not seen increased costs to maintain city contracts. Vendors and the city have successfully absorbed the cost of the living wage ordinance. There has been no adverse financial impact on the city. The living wage ordinance has been good for Boston.”
Richard Sander and Sean Lokey, in a study of the Los Angeles living wage, confirmed that living wage ordinances don’t generate substantial increases in overall contract costs, concluding, “Apart from the direct costs of the living wage ordinance to the city budget (about $500,000 in 1998 and $3-4 million when the living wage ordinance is fully implemented) and the administrative costs of implementation (also around $500,000 in 1998), there is no significant positive or negative effect on the city’s fiscal health or the local economy from the living wage ordinance” (Sander and Lokey 1998, 10). They do, however, identify some contracting arrangements that can lead to higher costs. In these cases, if a city intends to absorb added costs and informs contractors of this intention, there is little reason to expect that the city won’t face increased costs on those contracts. Overall, though, they found total labor costs to have increased by about $2.5 million, a far cry from their prediction before the fact of $30 to 40 million (Williams and Sander 1997, 62; Sander and Lokey 1998, 10).
Although not primarily concerned with the cost of the ordinance to the city of Los Angeles, a study by University of California Riverside economist David Fairris, UCLA researcher David Runsten, and colleagues at the Los Angeles Alliance for a New Economy (LAANE) reaches similar conclusions on this issue. The LAANE study includes surveys of affected workers, affected firms, and a comparable control group of firms not affected by the ordinance. Conducted several years after adoption and major modification of the living wage ordinance, these surveys support the finding that some firms have been able to pass costs through to the city, and similarly conclude that specific contracting arrangements are responsible in some cases. Most surveyed firms did not answer questions about cost pass-through, but half of those answering indicated that they were able to pass at least some of their costs on to the city (Fairris et al. 2005, 93). Qualitative evidence gathered in interviews suggests that this pass-through, inc
luding some firms billing the city in excess of the cost increases they actually experienced, results from contracts that are tied to these firms’ hourly wage costs.
The general conclusion of the available evidence is that, while some firms do increase prices to municipal governments, these price increases generally have only a negligible impact on city budgets. These increases are less than what many municipal governments had expected when they were preparing cost projections in anticipation of law changes, and much smaller than what was predicted by opponents of living wage laws.
The bidding environment
A handful of studies have looked at what happens to the competitiveness of the bidding process following the implementation of a living wage ordinance and have found that the competitive bidding process itself may be an important reason behind constrained growth in contract prices following adoption of living wage ordinances.
Sander and Lokey found that contract competitiveness played an important role in their findings. Of 30 firms surveyed 18 months after implementation of the living wage ordinance, they found that for 17 firms, “the costs of the contract to the city did not change and employment levels dropped modestly, if at all. In most of these cases, it is clear that the reason costs were absorbed by employers was the presence or introduction of competitive bidding” (Sander and Lokey 1998, 8). About one-quarter of the other contracts did experience cost increases, in large part because, as Sander and Lokey explain, “these were contracts that were not competitively bid; the city has a long-term relationship with a particular firm, and the firm was asked to determine the amount of the increased cost” (Sander and Lokey 1998, 8).
In their analysis of the Baltimore ordinance, Preamble reports that “from interviews with contractors it appears that it is a common practice to try to underbid the previous year’s contract, and it may be that the competitive pressures of the bidding process were enough that contractors were forced to absorb the increased costs from the living wage” (Preamble 1998, 9). The competitive bidding environment may prevent firms from passing costs back onto the municipal government, and the relatively small size of the costs of the living wage may make it still worthwhile for the company to continue to bid on contracts. It will be easier for firms to absorb small cost increases, rather than dramatically adjust operations (moving, laying off workers, not bidding, etc.).
Other studies have found similar results. Preamble’s study of the Baltimore ordinance found that, despite opponents’ predictions to the contrary, contracts that faced the largest increases in the wage bill saw an increased number of bids (Preamble 1998, 13).5 Elmore’s survey quotes one policy maker in Ypsilanti Township, Mich., as saying that there were “more bidders than ever, at better rates” following the living wage policy. In their study of New England cities, Brenner and Luce found that bids increased in one city (Hartford), didn’t change in another (Boston), and declined in the third (New Haven). In Boston, only 6% of firms indicated that the higher wage requirements will affect their willingness to bid on future city contracts.
The LAANE employer survey asked Los Angeles contractors about their attitudes toward city contracts following the adoption of the living wage. Seventy percent of firms had not changed their attitude toward city contracting; a significant minority (19%) indicated that they were less likely to seek city contracts in the future (Fairris et al. 2005, 111). Whether such intentions are honest, or are merely political posturing, is unclear. Brenner reports that in the case of Corvallis, Oregon, despite several firms threatening to not bid on future contracts, the city finance director reports that every firm contacted has submitted a bid, “and the bids have continued to be competitive” (Brenner 2004, 22).
A common sentiment expressed by contractors in Baltimore was that the higher wage floor leveled the playing field. As one bus company manager stated, “We feel more able to compete against businesses that were drastically reducing wages in order to put in a low bid.” The LAANE employer survey found that 11% of firms consider it easier to compete for city contracts following the living wage policy (Fairris et al. 2005, 111). These firms felt that the new policy had made it possible for “scrupulous” companies paying decent wages to compete against firms whose main strategy is to drive down wages. Elmore’s survey indicates that following adoption of a living wage, some cities instituted competitive bidding for contracts that had not been put out for bid in many years. These cities report that the return to bidding led to cost savings.
Brenner and Luce (2005) determined that the large increase in the number of bidders in Hartford was the result of more security firms willing to bid because of the living wage ordinance. Previously, firms that paid their workers higher wages were unwilling to bid when the outcome of the contract was determined exclusively by who could offer the lowest wages. One security guard contractor remarked, “Most companies with any business sense would concentrate on a higher wage niche, because there is more stability involved, and it gives you better control of the business, and allows you to preserve your reputation.” Similar sentiments were expressed in New Haven, where that city’s comptroller noted that the living wage “puts all vendors on equal footing… [and] it has leveled off undercutting.” Under the ordinance, competition for contracts is determined by more than which firms can drive wages down the lowest, and is influenced by other factors, such as service quality.
Costs to consumers
One of the most comprehensive, post-passage studies of a living wage ordinance followed the implementation of the living wage at the San Francisco International Airport (SFO). The SFO policy is almost universally applied to the airport workforce, directly affecting the wages of about 5,400 workers6 (Reich 2005, 119). The living wage policy is actually part of a series of policies called the Quality Standards Program (QSP) that includes a wage floor. Unlike most other living wage ordinances, the affected firms don’t provide services for a municipal government, but instead operate in a publicly owned facility.7
Most of the study, produced by Michael Reich and colleagues at UC Berkeley, concerns the employment and other economic impacts of the QSP (which will be discussed in the next section), and addresses the issue of cost increases faced by airport consumers. Reich shows that even if the entire employee compensation cost of the QSP was passed on to consumers, the effect would be relatively minor—an increase of $1.42 per passenger, an amount unlikely to deter people from using SFO (Reich 2005, 124). This hypothetical increase is substantially less than the $5 “per segment” security tax implemented following September 11th and the $4.50 departure tax proposed by the airport in 2001 to study options for a new runway8 (Reich 2003, 49).
Effects of living wage ordinances on workers
Living wage workers
The characteristics of workers who benefit have always been of interest to researchers studying the effects of living wage ordinances. Because of the policy’s stated goals, information on the demographics and family income of the workers receiving raises is relevant when judging success.
The Los Angeles Alliance for a New Economy conducted a survey in 2002 of 320 randomly selected workers who benefited from the Los Angeles living wage ordinance. This survey proved to be a rich data source for information on the thousands of workers who received raises, showing that:
• 9
6% were age 20 and older; 58% were 35 and older
• 86% worked full time
• 71% had only a high school degree or less
• On average, workers had been in the workforce nearly 20 years
• 29% were African American
• 57% were female
The LAANE survey did not provide reliable family income data. Instead, LAANE analyzed a similar group of low-wage workers from the Current Population Survey, finding that 69% fell below a “basic needs” budget (Fairris et al. 2005, 38).9
Brenner and Luce surveyed 97 low-wage workers employed in the industries most affected by Boston’s living wage policy.10 The survey of this group of covered workers reveals a generally similar profile as Los Angeles:
• Workers were predominantly adult, full-time workers, who were disproportionately people of color11
• The average age of covered workers in Boston was 32, with 95% age 20 or older (Brenner and Luce 2005, 51-52)
• 40% of covered workers were African American, and 79% were female
• The average covered worker worked 43 hour per week (Brenner and Luce 2005, 60)
Workers benefiting from the Boston living wage policy were also disproportionately poor and low-income, especially prior to its implementation. Among those covered workers getting a wage increase under the ordinance, over half (54%) were from households with incomes too low to afford even a basic needs budget. 12
One difference between the affected workers in Boston and Los Angeles is the level of education. Among workers impacted by Boston’s living wage, 37% had only a high school degree or less, compared to 71% in Los Angeles (Fairris et al. 2005, 31). More than half of covered workers in Boston had a two- or four-year degree, and 11% had a master’s degree (Brenner and Luce, 51). The reason for this difference is that the Boston ordinance primarily covers nonprofit social service providers (a workforce with relatively low wages and relatively high educational attainment), while the workers impacted by the Los Angeles ordinance primarily work at the airport and in a variety of service contract jobs for the city.
In the study of the San Francisco airport, Reich also reports some basic demographic characteristics of affected workers. Following the implementation of the QSP, more than three-quarters of affected workers were 25 or older, and 86% were non-white13 (Reich 2005, 134).
Employment effects
A frequently expressed concern about living wage ordinances is that the increased cost might decrease employment opportunities for low-skilled workers by causing employers to hire fewer workers or even lay off employees. The employment impact of living wage ordinances is a primary focus of most recent living wage studies. In attempting to answer the question of whether or not living wage ordinances have a significant impact on employment, different researchers have used a variety of approaches, ranging from qualitative interviews with service contractors and affected workers, to detailed before-and-after analysis of impacted firms, to econometric analyses of readily available labor market data. Most of the available studies have concluded that there have been either no or only small employment losses as a result of adopting living wages.
At the time when the earliest analyses were conducted, there was not enough data to quantitatively assess the impact that living wage laws had on employment. Instead, researchers relied on qualitative surveys to develop an impression of the potential impacts on employment. In their 1996 study, researchers from Preamble interviewed 31 contractors affected by the wage increase. None of the firms, including the janitorial services most heavily impacted by the increase, reported reducing staffing levels as a result of the living wage requirement (Preamble 1996, 10). In 1999, Niedt interviewed 26 workers employed in jobs affected by the Baltimore living wage ordinance. Based on questions about conditions at their workplaces, Niedt concluded there was “no evidence that employment levels or working time had changed because of the living wage” (Niedt 1999, 27). Later studies have used quantitative data and more sophisticated techniques to answer the question about employment impacts, and have reached similar conclusions as these early studies.
In his post-passage study of the Boston living wage, Brenner found little evidence of job losses. There was no significant difference in changes in employment (total employment or full-time equivalent (FTE) employment) between contractors who were forced to raise wages because of the law and those that did not have to raise wages (Brenner 2005, 73). For example, affected firms added 22.1 FTE positions, while unaffected firms added 22.4.14 Also, the number of contract employees covered by the Boston ordinance increased more at firms that were forced to raise wages to comply than those that did not have to raise wages. Brenner’s study documents that while approximately 1,000 workers received wage gains, there was no evidence of reduced employment or hours.15
The Los Angeles living wage ordinance directly raised the wages of an estimated 7,700 workers, according to the LAANE study16 (Fairris et al. 2005, 20). This extensive study, using original surveys of firms and workers, found that job loss occurred for less than 1% of the covered workers, or 1.4% of those receiving mandatory wage increases. On the firm side, less than one in five affected firms reported making any staffing changes due to the living wage.17
The analysis by Reich et al. of the living wage policy at the San Francisco Airport concluded that there was no evidence of employment losses due to the policy. Despite a recession-induced decline in airport activity by early 2001, SFO employment in jobs covered by the QSP rose by more than 15% between 1998 and 2001—the period in which the QSP was implemented (Reich 2005, 129). As Reich et al. report, “this increase is surprising given that over the same period, airport activity declined by 9% and overall employment in the San Francisco [metropolitan area] increased by only 1%.”18
Although her research focuses primarily on employee turnover, Candace Howes’ findings from her study of the living wage ordinance for home-care workers in San Francisco also does not support claims of job loss. Over the four years of her study (late 1997-early 2002), the number of home-care workers increased by 54% (Howes 2002, 2).
A series of studies by Neumark and Adams are an exception to the general findings of studies of employment effects. They report significant decreases in employment as a result of cities adopting living wage policies. In at least five separate papers, Neumark and Adams examine the effects of living wage laws by comparing the experience of the lowest-paid workers in cities with living wage laws to those in cities without such laws.19 In each of their studies, Neumark and Adams report that the workers in living wage cities have experienced positive wage effects, but negative effects on employment relative to workers in non-living wage cities.
While Neumark’s and Adams’ research has received wide attention, it has also been criticized by a number of economists, especially work by Brenner, Wicks-Lim, and Pollin. While it is not possible to fully address all of the criticisms in this review, below is a brief summary.
To begin with, the data source used in the Neumark and Adams studies is the Current Population Survey (CPS), a national survey used by the Bureau of Labor Statistics to measure unemployment, wages, and other labor market outcomes. While an excellent data source for many purposes, it is inappropriate for the task of analyzing the impact of living wage laws. Given that in some communities the living wage law only impacts a few hundred workers, it is unlikely that any affected workers are surveyed by the CPS at all in some communiti
es. Even in Los Angeles, with one of the broadest of living wage ordinances, Brenner, Wicks-Lim, and Pollin estimate that one year of CPS data would likely include about eight affected workers20 (Brenner, Wicks-Lim, and Pollin 2002, 13). In addition, the CPS does not contain data on the workers’ employer, making it impossible to positively identify those eight workers if they do appear in the survey. Using the CPS to analyze the economic effects of living wage laws makes finding a needle in a haystack look like a relatively simple chore, which is why most researchers have eschewed it for the more costly and time-intensive process of administering new surveys targeted specifically to be able to calculate the impacts of living wages. These surveys reflect the experiences of firms and workers actually impacted by living wage ordinances, while the CPS data at best allow Neumark and Adams to analyze a broad swath of the more general, low-wage workforce.
Neumark and Adams report that their findings are driven by laws that extend the living wage requirement to firms who are recipients of business assistance (such as tax breaks). They report that laws that only cover employees working on municipal contracts (the majority of policies) do not have significant impacts on wages or employment. The finding that laws covering business assistance drive the results casts doubt on the studies because most observers believe the business assistance extensions to be weakly implemented or even redundant. Brenner et al. have argued that a large share of the cities with business assistance provisions had not actually implemented this part of the law during the time studied by Neumark and Adams; while these provisions exists on paper, firms have not actually been required to raise wages because of them.21 Economic development expert Timothy Bartik considers the effects identified by Neumark and Adams unrealistic since, “large economic development subsidies typically only go to new and expanding manufacturing companies…[which]…are a small share of the labor market and pay high enough wages that few workers would be affected by living wages” (Bartik 2004, 290). Bartik’s assessment is supported by Elmore’s survey, which found that “many business subsidy programs already emphasized attracting high-wage jobs, so living wage laws effectively formalized and reinforced existing practices” (Elmore 2003, 2).
In order to rule out the possibility that their findings were spurious, Neumark and Adams calculated the wage and employment effects for two groups of workers they call “covered” and “non-covered” workers. Since living wage beneficiaries cannot be identified directly in the CPS, they used a classification scheme that ends up including unreasonably large portions of the workforce—over 85% of the lowest-paid one-fourth of workers in cities with living wage ordinances are classified as “covered” (Neumark 2002, 60). Referring to the Los Angeles example, Fairris estimates that fewer than 10,000 workers benefited from the living wage ordinance, but Neumark’s and Adams’ classification scheme proceeds as if approximately 450,000 workers received a raise under the ordinance!22
The size of the poverty reduction effects reported by Neumark and Adams are also simply too large given that living wage ordinances affect relatively few workers (Bartik 2004, 290). Similarly, the disemployment effect reported by Neumark and Adams is unrealistic, equivalent to 91% of the total number of workers most other researchers have estimated to be affected (Fairris and Reich 2005, 10).
Brenner et al. found that Neumark’s and Adams’ key findings are extremely sensitive to the inclusion of workers from Los Angeles earning less that the state minimum wage.23 Since most firms affected by the Los Angeles ordinance are also covered by the state’s minimum wage and can generally be expected to be in compliance with it, it is doubtful that workers not covered by the minimum wage would be “potentially covered” by the living wage law.24
Because of these factors, it is unlikely that the differences in wages, employment, and poverty between the two groups of cities (living wage and non-living wage) are due to living wage ordinances. As Richard Freeman notes, “any of a host of uncontrolled factors that change the economy in an area exclusive of a living wage ordinance could explain the empirical patterns [observed by Neumark and Adams]” (Freeman 2005, 24).
All told, Neumark’s and Adams’ results are simply not believable. Their econometric analysis shows that, on average, metropolitan areas with “business assistance” provisions tended to have more negative employment outcomes and more positive wage outcomes than other cities during the time studies. For all of the reasons discussed above, however, there is little reason to believe that these results are capturing the effects of living wage ordinances. The effects measured by Neumark and Adams are too large to be reasonable, the data source they use is inadequate to capture what they are hoping to measure, and there are too many other possible factors that could be driving their findings.
In summary, the best empirical research has shown that the adoption of higher wage floors has not resulted in measurable employment loss.. Yet many prospective studies predict the opposite. While some predictions of job losses resulting from living wage ordinances have been based on perfectly defensible, if not empirically supported reasoning, others are simply re-treads from different debates that are not actually relevant to living wage ordinances. One such argument is that firms will relocate to avoid having to pay a living wage. This is a standard (and generally unproven) argument in the debate over minimum wage laws, but it is not relevant to living wage ordinances. Living wage policies, particularly the predominant contractor-only variety, are typically not place-based policies. A service contractor can elect to not submit bids for future contracts should they not wish to abide by the living wage mandate. As long as they continue to contract, however, they will be covered by the law regardless of whether they relocate or not. For the few living wage ordinances that are place-based (in that they apply to firms leasing public facilities), it is either not feasible to relocate (airlines) or the geographic region of application is so narrow that firm relocation would not necessarily imply job loss for a city even if such relocation made sense (airport concessionaires or firms leasing other types of public facilities.) In any event, estimates provided by Pollin suggest that the costs imposed on firms from living wage ordinances are too low to justify relocation as a feasible response even if it were possible to dodge the living wage ordinance requirements by doing so (Pollin 2005). In Los Angeles, 81% of firms that were forced to raise wages did not cut any jobs, in large part because “either the number of workers affected was small or the size of the required raises was minimal” (Fairris et al. 2005, 95).
The absence of predicted job losses is due in part to the small impact of living wage policy on employers, and also that some of the costs faced by employers have been offset by increased spending by municipal governments. Although such cost increases are much lower than frequently predicted, as discussed in the previous section, they have occurred to some degree and have softened the blow to contractors accordingly.
In addition, there are details of specific living wage ordinances (as opposed to the general principle behind wage floors) that might limit job losses. In their study of the Baltimore living wage ordinance, Niedt identifies that the specific nature of the major school bus contracts makes it almost impossible to reduce either worker hours or employment levels. As Niedt explains, “the bus routes have not changed and cannot be drastically sped up, nor can an aide work on more than one bus at a time” (Niedt 1999, 19). Also concerning Baltimore, the Preamble study no
tes that some of the large janitorial contracts have mandatory staffing levels that the firms cannot alter even if they want to (Preamble 1998, 12). In Los Angeles, the LAANE study shows that contractually determined staffing levels also prevented job losses at parking firms as well as airline service contractors (Fairris et al. 2005, 95).
Other studies have identified that living wage ordinances in some municipalities apply to large numbers of nonprofit/human services organizations. Although nonprofits are exempted altogether in some living wage ordinances and almost entirely in others, they are covered in some cities. Because of their nonprofit status and strict limits on uses of some funding sources, nonprofits may respond differently to living wage ordinances than for-profit enterprises. As Brenner notes in his study of the Boston ordinance, nonprofits may go to greater lengths to avoid layoffs in the face of labor cost increases from a mandated wage increase (Brenner and Luce 2005).
Implementation and enforcement
The only way for workers to benefit from living wage laws is if they are covered by laws that are implemented and enforced. If few workers are covered and/or policies are not actually implemented or enforced, there is little reason to think that workers will gain.
Regarding implementation and enforcement, there have been problems for living wage ordinances from the very beginning. Even after adopting the first living wage ordinance in Baltimore, it took many months, rallies, public hearings, complaints, and fines before some firms started to obey the law. As Stephanie Luce has documented, major post-passage struggles have been required in several cities before the law was implemented. Based on extensive interviews with city administrators, living wage advocates, and review of newspaper reporting on living wage laws, Luce considers more than half of all living wage ordinances to have been only “narrowly” implemented25 (Luce 2005, 45). As she explains:
In some places, implementation seems to simply fall through the cracks: there is no single person in charge and no one who knows much about the ordinance. There are other cities in which the staff is incompetent, ineffective, or personally opposed to the ordinances. There are also cities where the administration is outwardly opposed to the ordinance and works to stall implementation, water down, or repeal the laws. Finally, some city councilors and/or administrators continue to publicly support living wage ordinances but make it easy for employers to receive waivers or exemptions from coverage. (Luce 2005, 46)
In their study of the Los Angeles living wage ordinance, Sander and Lokey found that enforcement, compliance, and discipline were all problems. Firms did not submit required paperwork, site visits were not performed, and no action was taken against contractors violating the policy. In their 18-month review of the ordinance, Sander and Lokey considered the discipline process to be “toothless,” and one of several implementation problems limiting the effect of the ordinance (Sander and Lokey 1998, 4). Sander and Lokey did indicate, however, that by late 1998 most implementation issues were improving. More recent work by LAANE indicates that, as of 2001-02, virtually all firms surveyed were in compliance with the wage requirements, but there may be problems with compliance with other provisions.
Finally, some living wage ordinances, even if they are implemented and enforced, have such narrow coverage that they raise the wages of few workers. This is a general problem with living wage ordinances around the country. Living wage ordinances end up being narrow in scope because some sectors are excluded from coverage (nonprofits, for example). Small contracts are also usually exempted from coverage, with small being defined as anywhere from under $10,000 to under $100,000. Also, small contractors, only partly related to the size of the contract, are sometimes exempted, based on number of employees or firm revenues.
Some cities also exempt contractors based on the source of their funding. In the first year of the Los Angeles ordinance, 59% of potentially covered contracts were granted exemptions, many because the contract was funded with federal resources, which the city was allocating or “passing through” (Sander and Lokey 1998, 2). Some ordinances apply only to those employees directly working on the contract, while others set a threshold, applying only to workers putting in more than a certain portion of their work time on the contract. In some ordinances, there are provisions to exempt contractors that are identified as facing extraordinary hardship under the ordinance. The combined effect of all of these exclusions and exemptions—particularly since the total employment of service contractors is small to begin with—means that in many cases very few workers are actually covered by the living wage.
In his review of living wage ordinances, Freeman notes “living wage campaigns pay a price for targeting small groups of workers in particular localities. The price is that the ordinances and policies affect only those relatively few workers. Most ordinances and policies cover at most a few hundred workers” (Freeman 2005).
These small numbers reflect what Jared Bernstein describes as the “paradox” of the living wage movement—activists succeed in passing ordinances, in part, by agreeing to narrow the focus and lower the cost of the ordinances (Bernstein 2005, 100). Ordinances are narrowed when exemptions are granted for particular types or sizes of contracts, broad classes of industries, and certain types of workers.26
Effects of living wage ordinances on firms
Productivity and turnover
One potential benefit of living wage ordinances (which is also one explanation for the minor impact on municipal budgets and employment levels) is that higher wage floors lead to decreased turnover and greater work effort among the affected workforce, as well as spur firms to seek out and adopt other means of boosting productivity. These responses could offset at least some of the increased labor costs experienced by employers. Most of the available research on living wages suggests that these types of responses are occurring.
Increased productivity resulting from wage increases has been recognized for decades, particularly in the economics literature on “efficiency wages” and debates over the minimum wage. With higher wages, workers may feel greater satisfaction with their job and may decide to put in greater work effort.27 Increased effort could also result from fear of losing the job; now that the job is more desirable than available alternatives the “cost of job loss” is greater. A related byproduct is that workers may be less likely to leave their jobs, thus lowering the rate of employee turnover and reducing costs of recruiting and training new workers. All of these mechanisms suggest ways that increased labor costs for firms are offset.
The research on the living wage has provided new opportunities to test for evidence of these effects. The earliest living wage studies relied on qualitative interviews, and presented evidence to suggest that employees were working harder with the new wage floor and turnover had declined. In their survey of affected workers, Niedt found that most reported an improved attitude toward their job, including a greater sense of worth of the job and an intention to stay on the job longer (Niedt 1999, 2). Similarly, in their interviews with effected contractors, researchers at Preamble found evidence suggesting improved attitudes toward work as well as reduced turnover (Preamble 1996, 13). The Preamble study quotes one manager as saying “workers seem happy [and] they come to work on time because they know that at $6.10 [in 1995] per hour, somebody else wants the job if they don’t.” Furth
er anecdotal evidence of decreased turnover following living wage ordinances is reported in Elmore’s survey of cities. Sander and Lokey’s interviews with contractors following implementation of the living wage ordinance in Los Angeles also yielded evidence that some firms had responded to increased labor costs by becoming more productive.
At SFO, Reich found evidence that the living wage led to increased productivity, reduced turnover, and shorter airport lines. There were 1,550 fewer turnovers per year at SFO following implementation of the QSP (Reich 2003, 55). Security screeners, who had a notoriously high turnover rate of 94.7% before the living wage, had just an 18.7% turnover rate after the living wage, where the average wage of security screeners went from $6.45 an hour to $10.00 an hour. Annual turnover among firms experiencing “high impacts” from the QSP fell from 49% to 20%, while turnover at “low impact” firms fell from 17% to 14% (Reich 2003, 52).28
In addition, employers reported a range of other positive outcomes following the implementation of the QSP: 35% reported improvement in work performance, 47% reported better employee morale, 44% reported fewer disciplinary issues, and 45% reported improved customer service.29
David Fairris’ study of Los Angeles found that employers in that city also reaped some benefits from the living wage. While employee turnover decreased for the entire sample of firms, it appears that firms affected by the living wage experienced larger decreases than firms that were not affected. Differences in the questionnaires for the two types of firms complicate the analysis, but Fairris reported significantly lower levels of turnover at affected firms. Controlling for other factors, Fairris showed one-third less turnover among low-wage workers in firms affected by the living wage ordinance (Fairris 2005, 101). This conclusion holds when looking at the entire firm and focusing exclusively on turnover among the lowest-paid workers. Based on his regression analysis, Fairris concludes, “The lower turnover rate for prominent low-wage occupations in living wage establishments is entirely accounted for by the higher wage that prevails there.” Additionally, absenteeism declined more at low-wage firms affected by the ordinance than at low-wage firms not impacted by the ordinance. The detailed interviews with firms affected by Los Angeles’s living wage ordinance buttress the statistical results. The LAANE study quotes one employer as stating, “Higher wages mean less turnover” (Fairria et al. 2005).
One in-depth analysis of living wage ordinances’ impact on turnover was conducted by Candace Howes, an economist at Connecticut College. Howes studied the impact of a series of living wage policies, which nearly doubled the hourly wages of homecare workers in San Francisco, and provided them with health insurance. Using a unique database linking payroll records with case management files, Howes found that homecare worker turnover fell by 57% following the implementation of the living wage policies30 (Howes 2005, 140). The likelihood that a new worker would stay at least one year on the job rose by 89%, after controlling for the effects of general economic growth.
Most studies looking at the topic of turnover, absenteeism, and employee work effort have found some evidence supporting the beneficial impacts of living wages. One study reporting mixed findings is Brenner’s analysis of Boston. Although one quarter of firms reported greater employee work effort and one quarter similarly reported improved morale, there was no evidence that higher wages produced lower turnover or less absenteeism (Brenner 2005, 73-77).
While the existence of the “efficiency wage” effect seems to be well documented in the living wage literature, it is less clear how much of the increase in labor costs is offset through greater productivity and decreased turnover. The cost-savings associated with increased productivity generally are extremely difficult to calculate, and no living wage research to-date has attempted to measure them. Several studies have attempted, however, to measure the cost-savings associated with decreased turnover. These estimates are influenced by the degree of reduction induced by living wage ordinances and the cost of employee turnover at the affected firms. In his research on Los Angeles, Fairris estimates that lower turnover offset 4% of the added labor costs from the living wage ordinance31 (Fairris 2005, 102).
In their follow-up study, Fairris and his colleagues at LAANE suggest that the original estimate of the cost offset from reduced turnover is likely a lower-bound because the cost of turnover is underestimated. After taking into account other research on the cost of turnover in low-wage industries, LAANE considers 16% to be a more reasonable estimate of living wage costs that are offset by decreased turnover32 (Fairris et al. 2005, 109).
Reich addresses some of the difficulties associated with measuring turnover costs by combining survey information of affected firms with industry and academic research regarding turnover costs among low-wage workers. Published estimates from human resource practitioners and trade associations indicate that the average per-worker turnover costs (including lost productivity) for employees earning $8 per hour range from $3,500 to $8,000 depending on the industry (Reich 2003, 56). Academic research on turnover costs in the hotel industry shows that average turnover costs range from $1,300 to $7,700 depending on the position and the region of the country.
Reich’s best estimate of the savings from turnover reduction in San Francisco was $6.6 million a year, offsetting one-tenth of the originally estimated cost of the QSP (Reich 2003, 58). These narrow cost savings, however, radically underestimate the potential benefits resulting from more reliable airport security. The costs of poorly implemented airport security can be truly devastating, and indeed impossible to calculate.
There is some debate over the causes of productivity gains due to living wage ordinances, and whether those gains should be interpreted as a benefit for low-wage workers—the intended beneficiaries of living wage ordinances. Some opponents of living wages argue that firms will simply substitute away from low-skilled employees to employees with higher skills if forced to pay a higher wage. The alternative is that productivity gains will result from current employees working harder or receiving more training and being equipped with more productive tools and equipment.
Some studies have attempted to isolate the direct causes of increased productivity. Are low-skilled workers being replaced with higher-skilled workers, or are workers working harder, getting better trained, and using more productive technology? Is turnover falling because workers are feeling more respected and the job is worth keeping? Or is it that less-reliable workers are being replaced by more-reliable workers hired at the new wage?
Brenner’s study of the Boston living wage reported that no affected firms changed their hiring standards following implementation (Brenner 2005, 79). The improved morale and increased work effort identified by affected firms is not attributable to former low-wage workers being replaced with higher-skilled substitutes, but to the current employees’ response to higher wages.
At SFO, Reich concluded that while some firms raised their hiring standards as a result of being forced to pay higher wages, more firms increased training for current and new employees. Eight percent of firms reported raising the bar for hiring new employees, which in part explains a slight shift away from workers with less than a high school degree (Reich 2003, 69). However, since the QSP made completion of a high school degree a condition for being hired in certain positions, this change rather than the wage floor itself is directly responsible for the shift away from high school dropouts.
Rather than substitute their wor
kforce with higher-skilled workers, more firms raised the level of training provided; this occurred among 20% of employers at SFO. Overall, the evidence from SFO suggests that substitution away from affected workers was minimal, and it was outweighed by improved job satisfaction and work effort among those workers. One telling fact is that 45% of the firms reported decreased employee grievances following the implementation of the QSP (Reich 2003, 60). These indicators, as well as sentiments expressed by workers at the airport, support Reich’s conclusion that worker effort, rather than displacement, has driven increased productivity and decreased turnover. One union organizer volunteered that people “don’t want to lose their jobs. The mentality is different now. Before people didn’t care, [they could] always find another $6 job” (Reich et al. 2003).
The LAANE study approached this question by comparing those employees hired before the ordinance was implemented and those hired after. Despite some minor differences, LAANE’s overall conclusion is that “the worst case scenario—the displacement of the workers who are the intended beneficiaries of the living wage—has not occurred. The majority of firms have not changed their hiring standards and reported seeing no changes in the composition of their workforce. A comparison of workers hired before and after the living wage reveals that new hires are no different in terms of age at hiring, years of schooling, whether they are native English speakers, and whether they are currently attending school. The proportion of Latinos has actually increased…” (Fairris et al. 2005, 115-6). There was some evidence that new hires were slightly more likely to be male and have received formal training before being hired, with a few workers having previously earned considerably higher wages than the pre-living wage workers at affected firms.
These studies suggest that increases in productivity at firms subject to a living wage ordinance are not the result of wholesale replacement of the workforce, but something else. Some of these firms may be able to attract better qualified workers at the living wage than before, but the current workforce is experiencing less turnover, less absenteeism, and likely greater work effort that is increasing productivity enough to offset some of the higher costs associated with the living wage ordinance.
Profits and wage scales
One possible response to an increased wage floor is that profits will decline or that wage growth for higher-paid employees will be restrained. These possibilities have received less attention in the research literature than other possible employer responses to living wage ordinances. As suggested by Brenner, this area should be more thoroughly addressed by future research on living wage laws.
The possibility that firms would lower their profit margins as a response to living wage ordinances was suggested in the very first assessment of living wage ordinances. Researchers from Preamble considered this a distinct possibility given the absence of evidence of job loss along with contractors adhering to the competitive behavior of under-bidding the previous year’s contract, despite added labor costs from the ordinance.
The first research to directly assess this possibility was Brenner’s analysis of the Boston living wage ordinance. Brenner found that while few service contractors reduced employment or reported less turnover, nearly 40% indicated that they had reduced profits as a response to the ordinance (Brenner 2005, 78). If credible, this information suggests that firms might reduce profits to absorb added labor costs from a living wage ordinance. Straightforward interpretation of this evidence, however, is complicated by the fact that a large majority of firms affected by the Boston ordinance are nonprofits.
Another way firms could reallocate revenue is by directing money that would otherwise have gone to higher-paid workers to the lowest-paid workers. Although living wage ordinances have compressed the wage scales of some firms, there is no evidence of decreased wages among higher-paid workers in analyses of the SFO and Boston living wage ordinances. In Boston, the share of city service contract workers earning less than $9.75 per hour declined dramatically and significantly, but the share of workers earning more than $11.75 did not fall (Brenner 2005, 73). At SFO, the evidence shows dramatic declines in the share of workers earning less than $8 per hour or less than $10 per hour, but not higher up the wage distribution (Reich 2003, 45). In fact, the share of workers earning more than $14 per hour at SFO more than doubled between 1999 and 2001.33
While there is no evidence of wages of higher-paid workers being constrained, it does appear that living wages compress wage scales within the firm. In the survey of nonprofits in the Detroit area, several firms voiced concerns about the newly increased wages of the lowest-paid now being “too close” to the wages of more experienced or highly educated workers, making it difficult to maintain internal pay differentials that depend on very low wages at the bottom.
Concerns about maintaining wage scales may lead employers to raise the wages of workers earning slightly above the new mandated wage. This “spillover” effect has been long discussed in research on the minimum wage and is the result of voluntary action by firms seeking to maintain pre-existing wage structures. Firms that formally or informally peg wages of certain positions to the minimum wage have to raise those wages as well if they want to maintain the pre-existing wage differential. There are similar pressures on firms that provide wage increases to maintain differentials based on employee tenure. Since coverage of the minimum wage is universal, the kind of spillover raises discussed is usually this “vertical wage push” type, where workers earning above the new wage also get a raise. Coverage of living wage laws is far from universal and does not necessarily apply uniformly across firms, industries, or occupations. In fact, it is possible for a firm to have some employees in a particular occupation covered by the ordinance (those working on the municipal contract), and others in the same occupation that are not covered. These gaps give rise to the possibility of “horizontal wage push,” where firms give raises to worker in order to maintain wage parity between workers that are affected and those that are not affected.
Early discussions of spillover effects in the living wage debate were primarily speculative, based primarily on educated guesswork and application of rules-of-thumb gleaned from the limited work on this issue in the minimum wage literature. Some of the recent living wage research, however, has directly addressed this issue, providing detailed information on the size of the spillover effect.
In their research on living wages at the San Francisco International airport, Reich et al. have documented a considerable degree of spillover. Of the nearly 8,000 workers getting a raise because of the QSP, one-third (2,550 workers) received spillover raises (Reich 2005, 119). There were nearly half as many spillover raises as mandatory raises resulting from the policy. Workers receiving mandated wage increases got a 33% raise, on average, while spillover raises averaged 10%.
The LAANE study showed that 7,700 workers in Los Angeles received mandated wage increases under the ordinance, and that 1,850 received “non-mandatory indirect raises” (Fairris et al. 2005, 19). The number of workers benefiting from spillover increases is substantial, accounting for one-fifth of all workers benefiting from the ordinance, and is one-fourth the size of the group getting mandatory raises. On average, spillover raises were half the size of mandated wage increases. From its survey of firms, LAANE found that, on average, raises were granted to workers earning up to 12% higher tha
n the living wage (Fairris et al. 2005, 45).
Lessons for policy makers and researchers
To date, most living wage research on which policy makers have had to rely has been prospective—they are written before the law has been implemented. With the increasing availability of quality studies and data on the actual (as opposed to projected) effects of living wages, future prospective studies should be less speculative and instead be based on the findings of the highest quality empirical studies.
Prospective studies have typically been created to inform and influence policy decisions, and have varied widely in their methodology, predictions, and accuracy. While a comprehensive review of prospective research is not within the scope of this paper, following are two predictions that prospective studies have commonly made, but have not been borne out.
Prediction one: significant costs to the municipality
Given sufficient information on the relevant contracts and workforce, it is possible to calculate reasonable estimates of the gross costs of mandated increases in wage and benefits from a living wage policy. It is more difficult, however, to determine who will ultimately pay for these cost increases. Prospective studies frequently focus on how much a living wage would cost the municipal government.
Lacking a significant body of research until recently, prospective studies have tended to base their predictions of how much of the cost pass-through would be passed onto local governments in the form of higher contract prices on educated speculation, sometimes justified with references to economic theory.34
Some studies make the extreme assumption that local governments will absorb all of the cost increases from a living wage. Other studies, however, assume that governments will only absorb a portion of the cost increase, acknowledging that some of the costs will be offset through decreased turnover and increased productivity and that since costs from the living wage represent a very small portion of their overall cost of doing business, firms in a competitive bidding environment may ultimately pass little of the cost increase onto the municipal government.
Evidence from the retrospective studies suggests that this latter approach is probably the most realistic. In his review of the economic impacts of living wages, Brenner shows that studies predicting modest cost increases yielded estimates compatible with the effects measured by many retrospective studies (Brenner 2004, 38). Prospective studies produced by living wage opponents (e.g., Tolley 1999) have predicted massive costs that have not been reflected by the actual experience of cities. In many cases, studies have ignored factors that offset the costs, such as those described above.
Even cities budgeting for a new living wage policy have systematically overestimated the ultimate cost of the policy. Elmore’s survey of cities that have enacted living wage laws shows that all of the cities that created budget forecasts significantly overestimated the actual costs of implementation. Actual costs ended up being between 30%-52% lower than what was forecast by the municipal government (Elmore 2003, 8).
To most accurately reflect the likely cost of the policy, prospective studies need to acknowledge, at bare minimum, that municipal governments will not bear all of the cost of a living wage, and most likely will experience only relatively small budget impacts.
Prediction two: significant employment losses
The economic impact of greatest interest for most prospective studies, and policy makers as well, is jobs. Most prospective studies have discussed potential impacts on employment and some have provided estimates of job loss. Typically these studies have relied on the minimum wage literature, both the theory and the empirical research, to infer the impacts of living wages on employment.
Some anti-living wage studies cite minimum wage research to support their claims of major job losses, but as esteemed labor economist Richard Freeman has concluded, the minimum wage “debate is over whether modest minimum wage increases have `no’ employment effect, modest positive effects, or small negative effects. It is not about whether or not there are large negative effects” (Freeman 1995, 833). This emerging consensus on the employment impacts of the minimum wage, however, is of limited use in the discussion of living wages because the living wage is set so much higher—anywhere between 50%-250% higher—than the federal minimum wage, with some living wage policies also requiring health insurance and other fringe benefits.
In addition, because the coverage of the two laws is so different, it is not clear they will have the same impacts on employment; while minimum wage laws cover most or essentially all firms in a geographic region, living wage ordinances cover the relatively few firms with direct voluntary financial relationships with municipal governments, and even then provide significant exemptions based on firm size and industry, as well as employee type. Given these differences of coverage and level of benefit, findings from the minimum wage literature cannot accurately translate to a living wage policy.
Conclusion
As in the debate over minimum wages, the question of the impacts of living wages on employment is ultimately answerable empirically. Thus far, the most reliable research on living wages suggests that the impact is modest. In the largest cities with the broadest-based living wage policies, there has been little measured employment loss. Given these results, prospective studies would do best to acknowledge that offsetting factors and modest costs for employers result in only limited job losses from a living wage policy. Ignoring the importance of offsetting factors will result in extreme overestimates of costs and job losses under a living wage policy.
—February 2006
We gratefully acknowledge the support of the Annie E. Casey Foundation and the Rockefeller Foundation. Invaluable research assistance was provided by Virginia Leavell. Jared Bernstein, Mark Brenner, Carolina Briones, Michael Ettlinger, Jen Kern, Virginia Leavell, and Paul Sonn provided helpful comments and access to their own research.
Endnotes
1. The Employment Policies Institute (EmPI) is an operation housed at the lobbying firm Berman and Company and works with low-wage employers to oppose minimum wages and living wages.
2. At best, EmPI refutes Preamble’s finding that contract costs decreased during the period. However, EmPI inaccurately portrays the Preamble findings as a claim that the living wage ordinance actually caused contract costs to increase, despite the fact that Preamble states clearly in their study, “We cannot, of course, conclude that the living wage ordinance actually contributed to lowering the cost of the average contract.”
3. Preamble appears to have been given incorrect information by the city and it is likely the contract was not covered. Nevertheless, the Preamble response shows that the exclusion of the contract does not change their key finding.
4. Separate interviews of Cleveland officials (not included in Elmore’s study) conducted by researcher Dana Williams also found that contract costs had not increased.
5. They also found that contracts already paying above the living wage experienced a decrease in the number of bids.
6. An estimated 2,550 workers who were already earning at or above the living wage level received indirect, or “spillover” raises.
7. The Lo
s Angeles living wage policy also covers Los Angeles International and Ontario (Calif.) International airports, which account for 60% of the affected jobs.
8. These points were made by Reich in an earlier draft of the study, published as a UC Berkeley working paper.
9. The LAANE study used “needs-based” budgets developed by the California Budget Project and the National Economic Development and Law Center.
10. Although the survey was not based on a random sample, Brenner and Luce employed a variety of alternative sampling techniques to ensure that the surveyed workers are representative of the population of affected workers. See Brenner and Luce 2005, Appendix 6 for details.
11. Brenner and Luce 2005, tables 4.4 and 4.6.
12. Among the 76 “covered” workers with reliable before and after wage information, 32 earned below the living wage in 1998 and are considered “affected” workers.
13. Here “affected workers” are those in low-wage occupations who had been on the job between one and five years. The age of workers is their age when they started the job.
14. There was, however, a significant difference in reliance on part-time workers, with the share of part-time workers dropping considerably among affected firms.
15. The estimate of 1,000 workers getting a raise includes employees directly covered by the law, and the ripple effect on non-covered employees.
16. An estimated 1,850 workers who were already earning at or above the living wage level received indirect, or “spillover” raises. In his initial study, Fairris estimated 6,500 affected workers, but the figure was revised in later work he completed with LAANE.
17. An earlier version (Fairris 2005) reported larger, but still small, employment effects. The final version uses the same dataset, but with improved methodology.
18. Reich et al. demonstrate that this decline in airport traffic, which was also experienced by airports around the world, was due to a general decline in economic activity, and then to the events of September 11, 2001. Other Bay area airports fared better with airport traffic than SFO in 2001, primarily due to the relocation of Southwest Airlines to the Oakland airport, which left SFO after failing to secure additional terminal facilities (Reich 2005, 131-32).
19. Most of the studies also include other wage ranges as well and include the bottom quarter of workers, but the most consistent findings are for the lowest-wage 10% of workers. Also, the studies include specifications for contemporaneous effects, as well as six- and 12-month lagged effects. The 12-month lagged effects are generally the most robust findings.
20. The Brenner, Wicks-Lim, and Polling figures are based on an assumption of 7,600 affected workers that was developed before the release of either the Fairris or LAANE studies.
21. There is disagreement between Neumark and Adams and their critics as to whether only one city in Neumark’s 2002 study (as maintained by Brenner et al. 2002) had implemented the business assistance living wage provisions, or if the number is considerably larger. Whatever the exact number, it is certain that the actual impact of business assistance living wage ordinances is considerably less than an impression gained solely by looking at which cities had adopted these provisions in their city code. Neumark and Adams conduct interviews with municipal government administrators responsible for implementing the business assistance provisions of the living wage ordinance and reach a different conclusion than Brenner et al., finding that many cities are in fact implementing the provisions to some extent (Neumark and Adams 2005c, 19-20). As Bernstein points out, however, there is still a gap between what Neumark and Adams identify as “implemented” and what Brenner et al are implying in their critique: actually having to raise wages.
22. Brenner, Wicks-Lim, and Pollin indicate that Neumark’s scheme covers 97% of workers, while Neumark (2002) identifies 90% of the bottom quartile of workers as being “covered.”
23. Brenner et al. (2002) also make a technical point that by truncating his sample to focus on the lowest wage 10% of workers that Neumark could be introducing “sample selection bias.” Instead, they argue that “quantile regression,” focusing on the 10th percentile of the entire wage distribution is appropriate. Neumark and Adams respond that their truncation approach is necessary to capture the impacts of the living wage, and that, in fact, is unlikely to introduce sample selection bias. Neumark’s and Adams’ argument on this point is probably correct, but is not relevant to the main part of the critique levied by Brenner et al.
24. Neumark and Adams claim not to understand this critique by Brenner et al, and do present a defense of their position. They do not, however, refute the Brenner et al. critique.
25. For an additional 10% of adopted ordinances, implementation has been blocked either by courts or elected officials, or was overturned by voters.
26. Nationwide there were potentially 100,000 workers that had received wage increases under living wage ordinances as of 2002, although dozens more successful campaigns since that time have likely increased that number by tens of thousands (Tanner 2002, 769).
27. See, for example, Greenwald and Stiglitz 1988 and Akerlof and Yellen 1990.
28. “High impact” refers to firms where the direct increase in wage costs under the Quality Standards Program is 10% or higher.
29. The percentages of employers reporting that these areas actually suffered were in the single digits.
30. The period under study was from November 1997 to February 2002. The policy change studied by Howes was not just limited to a living wage ordinance, but also was accompanied by union organizing.
31. The 4% offset is based on sizeable reductions in turnover among affected firms, but quite low turnover costs. Fairris reports that the average reported cost of turnover was just $807 per worker. These turnover cost are considerably smaller than most other available estimates, and are actually based on his control group survey of firms not affected by the living wage law.
32. In Pollin’s survey of hotels, retail establishments, and restaurants in Santa Monica, firms reported average non-managerial turnover costs of $2,090. Similarly, in Brenner’s survey of Boston firms, it was reported that the average per-employee cost of turnover was $9,297 and the median cost was $2,500. The differences between these estimates could be influenced by the different mix of industries affected by different living wage laws (landscaping in Los Angeles versus human services in Boston) or by the possibility that mail and telephone surveys are an inadequate means by which to gather data on the cost of employee turnover. It could also be the case that the surveyed firms do not actually know the true costs of employee turnover.
33. Nominal figures not adjusted for inflation.
34. The competitiveness of the market and the price elasticity of demand for the products in question influence the ability of producers (in this case, service contractors) to pass cost increases on to their consumers (in this case municipal governments). The degree of price increases passed onto taxpayers involves a further set of assumptions regarding the responses of local governments. Some combination of higher taxes, reduced services, or greater efficiencies determine the size of any potential tax increase—essentially a parallel set of assumptions to those behind the extent of the increased costs passed onto the local government in the first place.
References
Adams, Scott and David Neumark. 2005a. A Decade of Living Wages: What Have We Learned? San Francisco, Calif: Public Policy Institute of California.
Adams, Scott and David Neumark. 2005b. The economic effects of living wage laws: a provisi
onal review. Urban Affairs Review. Vol. 40, No. 2, pp. 210-45.
Adams, Scott and David Neumark. 2005c. Living wage effects: new and improved evidence.” Economic Development Quarterly. Vol. 19, No. 1, pp. 80-102.
Adams, Scott and David Neumark. 2004. When do living wages bite? Industrial Relations. Vol. 44, No. 1, pp. 164-92.
Akerlof, George and Janet Yellen. 1990. The fair wage-effort hypothesis and unemployment. The Quarterly Journal of Economics. Vol. 105, No.2, pp. 255-83.
Baiman, Ron, Joseph Persky, and Nicholas Brunick. 2002. A Step in the Right Direction: An Analysis of the Forecasted Costs and Benefits of the Chicago Living Wage Ordinance. Chicago, Ill.: Center for Urban Economic Development, University of Illinois, Chicago.
Bartik, Timothy. 2005. Thinking about local living wage requirements. Urban Affairs Review. Vol. 40, No. 2, pp. 269-99.
Bernstein, Jared. 2005. “The Living Wage Movement: What Is It, Why Is It, and What’s Known About Its Impact?” in Richard B. Freeman, Joni Hersch, and Lawrence Mishel, eds.,Emerging Labor Market Institutions for the Twenty-First Century. Chicago, Ill.: The University of Chicago Press.
Bernstein, Jared and Jeff Chapman. 2002. Time to Repair the Wage Floor: Raising the Minimum Wage to $6.65 Will Prevent Further Erosion of Its Value. Washington, D.C.: Economic Policy Institute.
Bernstein, Jared, Chauna Brocht, and Maggie Spade-Aguilar. 2000. How Much is Enough? Basic Family Budgets for Working Families. Washington, D.C.: Economic Policy Institute.
Brenner, Mark D. 2004. The Economic Impact of Living Wage Ordinances. Working Paper No. 80. Amherst, Mass.: Political Economy Research Institute, University of Massachusetts, Amherst.
Brenner, Mark D. 2005. The economic impact of the Boston living wage ordinance. Industrial Relations. Vol. 44, No. 1, pp. 59-83.
Brenner, Mark D. and Stephanie Luce. 2005. Living Wage Laws in Practice: The Boston, New Haven and Hartford Experiences. Amherst, Mass.: Political Economy Research Institute.
Brenner, Mark, Jeannette Wicks-Lim, and Robert Pollin. 2002. Measuring the Impact of Living Wage Laws: A Critical Appraisal of David Neumark’s How Living Wage Laws Affect Low Wage Workers and Low-Income Families. Working Paper No. 43. Amherst, Mass.: Political Economy Research Institute.
Citro, Constance F. and Robert T. Michael, eds. 1995. Measuring Poverty: A New Approach. Washington, D.C.: National Academy Press.
Elmore, Andrew J. 2003. Living Wage Laws & Communities: Smarter Economic Development, Lower than Expected Costs. New York, N.Y.: Brennan Center for Justice.
Employment Policies Institute. 1998. The Baltimore Living Wage Study: Omissions, Fabrications and Flaws. Washington, D.C.: Employment Policies Institute.
Fairris, David. 2005. The impact of living wages on employers: a control group analysis of the Los Angeles ordinance. Industrial Relations. Vol. 44, No. 1, pp. 84-105.
Fairris, David and Michael Reich. 2005. The impacts of living wage policies: introduction to the special issue. Industrial Relations. Vol. 44, No. 1, pp. 1-13.
Fairris, David, David Runstein, Carolina Briones, and Jessica Goodheart. 2005. The Los Angeles Living Wage Ordinance: Effects on Workers and Employers. Los Angeles, Calif.: Los Angeles Alliance for a New Economy.
Freeman, Richard. 1995. What will a 10%…50%…100% increase in the minimum wage do? Industrial & Labor Relations Review. Vol. 48, No. 4, pp. 830-34.
Freeman, Richard. 2005. Fighting for other folks’ wages: the logic and illogic of living wage campaigns. Industrial Relations. Vol. 44, No. 1, pp. 14-31.
Greenwald, Bruce and Joseph Stiglitz. 1988. “Pareto inefficiency of market economies: search and efficiency models. American Economic Review. Vol. 78, No. 2, pp. 351-55.
Howes, Candace. 2002. The Impact of a Large Wage Increase on the Workforce Stability of IHSS Home Care Workers in San Francisco County. Working Paper. New London, Conn.: Department of Economics, Connecticut College.
Howes, Candace. 2005. Living wages and retention of homecare workers in San Francisco. Industrial Relations. Vol. 44, No. 1, pp. 139-63.
Luce, Stephanie. 2005. The role of community involvement in implementing living wage ordinances. Industrial Relations. Vol. 44, No. 1, pp. 32-58.
Neumark, David and Scott Adams. 2000. Do Living Wage Ordinances Reduce Urban Poverty? Working Paper No. 7606. Cambridge, Mass.: National Bureau of Economic Research.
Neumark, David. 2002. How Living Wage Laws Affect Low-Wage Workers and Low-Income Families. San Francisco, Calif.: Public Policy Institute of California.
Niedt, Christopher, Greg Ruiters, Dana Wise, and Erica Schoenberger. 1999. The Effects of the Living Wage in Baltimore. Working Paper No. 199. Washington, D.C.: Economic Policy Institute.
Pollin, Robert. 2005. Evaluating living wage laws in the United States: good intentions an economic reality in conflict? Economic Development Quarterly. Vol. 19, No. 1, pp. 3-24.
Pollin, Robert. 2002. Santa Monica Living Wage Study: Response to Peer Reviews and Business Critics. Research Report No. 5. Amherst, Mass.: Political Economy Research Institute.
Preamble Center for Public Policy. 1998. Baltimore’s Living Wage: Response to Critics. Washington, D.C.: Preamble Center for Public Policy.
Reich, Michael, Peter Hall, and Ken Jacobs. 2005. Living wage policies at the San Francisco airport: impacts on workers and businesses. Industrial Relations. Vol. 44, No. 1, pp. 106-38.
Reich, Michael, Peter Hall, and Ken Jacobs. 2003. Living Wages and Economic Performance: The San Francisco Airport Model. Berkeley, Calif.: Institute of Industrial Relations.
Reich, Michael, Peter Hall, and Fiona Hsu. 1999a. Living Wages and the San Francisco Economy: The Benefits and the Costs (First Release). Berkeley, Calif.: Bay Area Living Wage Research Group, Center on Pay and Inequality, Institute of Industrial Relations, University of California, Berkeley.
Reich, Michael and Peter Hall. 1999b. Living Wages at the Airport and Port of San Francisco: The Benefits and the Costs. Berkeley, Calif.: Bay Area Living Wage Research Group, Center on Pay and Inequality, Institute of Industrial Relations, University of California, Berkeley.
Reynolds, David and Jean Vortkamp. 2005. The effects of Detroit’s living wage law on nonprofit organizations. Economic Development Quarterly. Vol. 19, No. 1, pp. 45-61.
Sander, Richard and Sean Lokey. 1998. The Los Angeles Living Wage: The First Eighteen Months. Norcross, Ga.: The Fair Housing Institute.
Tanner, Jane. 2005. Living wage movement: do laws requiring higher wages cause unemplo
yment? The CQ Researcher. Vol. 12, No. 33, pp. 769-92.
Tolley, George, Peter Bernstein, and Michael Lesage. 1999. Economic Analysis of a Living Wage Ordinance. Washington, D.C.: Employment Policies Institute.
Weisbrot, Mark and Michelle Sforza-Roderick. 1996. Baltimore’s Living Wage Law: An Analysis of the Fiscal and Economic Costs of Baltimore City Ordinance 442. Washington, D.C.: Preamble Center for Public Policy.
Williams, Dana. 2004. Cleveland’s Living Wage Law: A Three-Year Review. Cleveland, Ohio: Policy Matters Ohio.
Williams, E. Douglas and Richard Sander. 1997. An Empirical Analysis of the Proposed Los Angeles Living Wage Ordinance: Final Report. http://www.law.ucla.edu/sander/L_Wage/Sander_LA-LivingWage-19970117.pdf.
