Unions and the decline of U.S. cities.

Walters, Stephen J.K.

The usual suspects in the tragic demise of many of America's core cities are well known. For decades, scholars, politicians, and pundits have condemned the racism that led whites to flee diverse urban populations after World War II, sneered at Americans' vulgar affection for cars and expansive lawns, criticized policies that encouraged us to indulge these tastes, and blamed capitalist greed and unwholesome technological change for the deindustrialization that has wrecked urban labor markets.

There is, of course, some explanatory power in these familiar stories. But they leave much unexplained, and in some cases merely describe symptoms of urban dysfunction rather than identify root causes of decline. For example, race-bias-based theories seemed quite powerful during the well-documented white flight of the 1950s through the 1970s, but less so as middle- and working-class blacks exited many core cities in recent years. Preference-based theories have a tough time explaining the enduring popularity of high-density enclaves such as Manhattan, or why cities like San Francisco and Boston depopulated and decayed for about three decades after WWII--but then revived, while others simply continued to slide downhill. If our theories need to be discarded at various times or in different places, perhaps we need new ones.

In this article, I argue that a more useful and general theory about the fate of American cities in the last half of the 20th century must begin with a discussion of the treatment of capital and the security of property rights within their borders. In particular, I will focus on the powerful influence of labor institutions in reducing the returns to capital in many American cities, thus contributing to their transformation from engines of prosperity into areas afflicted by economic stagnation, chronic poverty, and all the social problems that metastasize in such circumstances.

We tend to think of cities as dense concentrations of people living within a given land area, but it would be more fruitful to think of them as dense concentrations of capital that attract people to a certain locale. Physical capital some of it in the form of natural endowments such as a deep-water harbor, much of it the product of human investment in dwellings, factories, offices, and infrastructure--is, of course, a profoundly valuable partner in both our work and play. The greater the quantity and quality of such capital (all else the same), the higher will be our productivity and wages and the more stimulating, satisfying, and comfortable will be our leisure hours.

Such capital will, however, always be a tempting target for interest groups seeking to redistribute some of its returns from its owners to themselves. Since physical capital is immobile, it can be "taken hostage" by opportunistic actors and its value appropriated in ways that will soon be described. And since it is durable, the ill effects of such actions will generally be disguised for years or decades.

Accordingly, to develop an understanding of the nature and consequences of such behavior for the vast and varied concentrations of capital that we call cities we must take the long view but should not overlook the individuals and actions that have shaped real urban environments. What follows, then, combines standard economic and quantitative analysis with case evidence from what was America's first great high-tech industrial center and what is now its most destitute, violent, and politically dysfunctional major city: Detroit. Its rise and fall can tell us a great deal about the nature of industrial cities and about the way that labor institutions can affect their economies. The next section describes Detroit's "golden age" and the sources of its success. Succeeding sections discuss union behavior in general terms, illustrate this behavior in the context of Detroit's auto industry and the cartelization of its labor force, and enumerate the consequences of this behavior; a concluding section contains both pessimistic and optimistic speculations about the future and some policy recommendations.

A Target of Opportunity

To call Detroit a boom town during the first third of the 20th century would be to damn it by faint praise. Blessed with natural capital in the form of proximity to water transportation (eventually augmented with man-made capital in the form of rail lines) that provided low-cost access to nearby hardwood forests and mineral deposits that fueled the growth of carriage-makers, tool works, and other manufacturing enterprises, by 1900 the city was the 13th most populous in the United States--just behind New Orleans and ahead of Milwaukee. By 1930, however, Detroit was home to over 1.5 million and America's 4th largest city, its 450 percent population growth rate more than four times that of New York and Chicago and nine times that of Philadelphia over the same period)

The reason, of course, was Detroit's status as the nation's center of innovation and production in the nascent automobile industry. But this was not just a happy, accidental result of the fact that many of this industry's founding figures had grown up or begun careers nearby--including Henry Ford in Dearborn, William Durant in Flint, Ransom Olds in Lansing, and the Dodge brothers and David Buick in Detroit. Rather, this "entrepreneurial duster" (to use the phrase of Glaeser, Kerr, and Ponzetto 2009) built on a foundation of industrial, intellectual, and financial capital that was especially well-suited to working out the engineering and production problems associated with this new and rapidly evolving product. What better place to make horseless carriages than in a city where coach and tool-and-die manufacturing already were well established?

The success of the early automotive innovators in Detroit attracted more, in a dramatic illustration of Alfred Marshall's (1920) description of the economies of industrial agglomeration. Auto production in the industry's early days was not exclusive to Detroit, but it became more and more concentrated there because agglomeration economies gave Detroit firms a competitive advantage: low-cost links to suppliers of raw materials and components, access to a larger and deeper pool of labor (including managers and engineers) with specialized skills, and--perhaps most important--the technological spillovers resulting from proximity to talented minds grappling with similar problems. As Marshall (1920: 271) noted,

   When an industry has thus chosen a locality for itself, it is
   likely to stay there long: so great are the advantages which people
   following the same skilled trade get from near neighbourhood to one
   another. The mysteries of the trade become no mysteries; but are as
   it were in the air, mad children learn many of them unconsciously.
   Good work is tightly appreciated, inventions and improvements in
   machinery, in processes and the general organization of the
   business have their merits promptly discussed: if one man starts a
   new idea, it is taken up by others and combined with suggestions of
   their own; and thus it becomes the source of further new ideas. And
   presently subsidiary trades grow up in the neighbourhood, supplying
   it with implements and materials, organizing its traffic, and in
   many ways conducing to the economy of its material.

A serious disadvantage of this co-location soon became obvious, however. The factories, offices, warehouses, and transportation links necessary for the design, production, and distribution of autos and related goods were installed and augmented at an incredible rate, attracting not just laborers but those who would unionize them. In effect, agglomeration economies reduce firms' production costs and accelerate innovation, but the concentration they beget also reduces the cost of organizing and enforcing cartels of labor. To employ a military analogy, automakers had concentrated their assets and made them vulnerable to a siege by those who sought to capture them, and who could focus their forces on this task rather than divide them among many targets spread more widely.

Automakers were not unaware of this vulnerability. In 1901, for example, some workers at an Olds factory in Detroit joined a national strike for higher wages and shorter workdays. When nonunion workers kept the plant running, the strikers and about 500 sympathizers tried to occupy it, and three people were injured in the brawl that ensued. Olds soon built a new facility 90 miles away in Lansing--perhaps the first example of union-related flight of capital and jobs from Detroit. As historian James Rubenstein (1992: 234) observed, "avoiding concentrations of militant workers influenced location decisions even in the early days of the automotive industry." But agglomeration economies were too important to ignore, and, overall, the labor climate in Detroit was benign; the city was regarded as a nonunion town, and Michigan was an "open shop" state. And in 1902 the city's leading industrialists had formed the Employers' Association of Detroit, which worked to eliminate any "closed shop" agreements between member employers and unions, supplied members with substitute nonunion workers if and when a strike occurred, and marshaled legal resources to obtain injunctions against certain union practices and even arrest union leaders if these injunctions were ignored.

As a result, in the early decades of the 20th century unions usually represented less than a tenth of Detroit's labor force. And, as we have seen, the city's growth was spectacular, while its industrial base produced enormous wealth not just for entrepreneurs, managers, engineers, and traders, but for laborers as well. In 1930, there were 275 U.S. counties with at least 5,000 manufacturing workers within their borders. Those in Michigan's Wayne County (which includes Detroit and adjacent cities such as Dearborn, Hamtramck, Highland Park, and River Rouge) earned average wages higher than those in all but three other counties which contained Youngstown and Warren, Ohio, and Gary, Indiana (where the nation's largest steelmakers had facilities). Manufacturing wages in Detroit exceeded the national average by fully 33 percent, and when compared to wages in smaller factory towns elsewhere the contrasts are even more dramatic: factory workers in El Paso, Texas, earned only 60 percent as much as those in Detroit, while workers in York, Pennsylvania earned 56 percent as much and those in Greenville, South Carolina, 40 percent as much.

Detroit's absolute and relative prosperity is difficult to reconcile with pro-union rhetoric during this period (and historic treatments since), which stressed the need for countervailing power for workers in the face of employers' unfettered monopsony power. Absent collective bargaining, the story went, workers would be exploited with unjust wages and brutal working conditions. Even if this argument is accepted at face value, however, one would think that the task of raising workers' wages and improving their working conditions might start, or at least be concentrated, where wages are lowest, conditions worst, and so the need for countervailing power greatest. It did not. Instead, the efforts of America's most active labor organizations were generally most intense in those locales where abundant capital had already improved laborers' productivity and standard of living to levels far greater than experienced in areas of relative capital scarcity. The bulk of labor history for this era is written about offensives not just against the owners of the burgeoning plants of Detroit, but the mills of Youngstown (average wages 37 percent higher than the national average), Gary (34 percent higher), Chicago, Cleveland, and Pittsburgh (all 19 percent higher), and Buffalo (14 percent higher); the factory workers of El Paso, York, Greenville, and hundreds of other locales seem to have been largely ignored, at least at this time. But if union strategists' rhetoric was misleading, their logic was impeccable.

The Strategy and Tactics of Plunder

That collective bargaining allows workers to set above-competitive prices for their services is well known and much documented. While it is also true that unions might provide productivity-enhancing services both to workers and employers (see, e.g., Freeman and Medoff 1984 for a discussion of unions' "voice/response" capabilities), the evidence is not friendly to the suggestion that such effects are large or offset unionized labor's higher costs. In general, the empirical literature finds that (a) union wage premia are significant but vary considerably over time and across industries, (b) employment is lower in unionized sectors, (c) unions have a near-zero effect on labor productivity, (d) unions reduce firm profitability, capturing quasi-rents (2) associated with firms' durable tangible and intangible capital, and (e) unions reduce investment and productivity growth (see Hirsch 2007 for an excellent summary).

It is important to note that unions have the capacity to distort investment decisions by engaging in both monopolistic and opportunistic behavior, and that such behavior will have important effects on the location as well as the volume of investment. Suppose, for example, that an entrepreneur is contemplating a $10 million investment in physical capital that is specialized to a particular use; for simplicity we assume it will have zero salvage value (e.g., it is used to fabricate a unique product and has no other uses). Suppose also that costs for raw materials and other miscellaneous inputs are $8 million, that the total labor bill for the anticipated production run would be $20 million if the labor market is competitive, and that the resulting output can be sold for $40 million (all dollar values in present discounted value terms). If all these expectations come to pass, the entrepreneur would net profits of $2 million ($40m--$10m--$8m--$20m) and realize a 20 percent return on investment ($2m / $10m).

As long as the yield on alternative, equivalently risky investments is 20 percent or less, this project should get a green light. The possibility that labor might not be available at competitive prices, however, introduces complications. Note first that if the yield on alternative, equivalently risky investments is exactly 20 percent (so that this project is "at the margin" in a top-to-bottom ranking of potential investments), any increase in the wage bill would drive this project's return below the alternatives' and render it undesirable. If, on the other hand, the best available alternative investments yield, say, 10 percent (so that this project is an "inframarginal" investment yielding quasi-rents), a labor cartel could demand a wage premium of up to 5 percent (raising the total labor bill to $21 million) and this project might still be a "go," since this would simply eliminate its prospective quasi-rents while leaving returns commensurate with those of other opportunities. Of course, a 10 percent wage premium (raising labor costs to $22 million) would drive net returns to zero and kill the project.

The usual lesson drawn from such examples is that above-competitive wage rates certainly reduce investment and employment at the margin, but that if labor cartels are careful about setting wages (a big if) they can capture firms' quasi-rents on inframarginal investments without further adverse effects. It is often supposed that in aggregate, then, the welfare costs of labor cartels are not large relative to the amount of income they redistribute. But this ignores the possibility that quasi-rents may not be location-specific. (3) Investors will have a very strong incentive to investigate this matter and if, for example, the project described above can be undertaken in a locale which promises a truly competitive labor market it will tend to be placed there. This will have very grave implications for cities that are host to labor cartels.

Then there is another real-world complication: physical capital is durable, and the expectations under which it is created are not always fulfilled. This provides an additional source of possible gain for a union: returns to opportunistic behavior (see Klein, Crawford, and Alchian 1978). Suppose, in our example, the entrepreneur agrees to pay union laborers a 5 percent wage premium (i.e., $21m) and commences production, expecting to earn a 10 percent return (commensurate with that of alternatives) on the $10 million specialized investment, Shortly after the capital is installed, however, the workers strike--a "wildcat" strike that is officially unauthorized and apparently in violation of the agreement. But litigating this breach will take years and carries no certainty of victory, so the entrepreneur listens to the strikers' demands. They are shocking: unless the payment to labor is increased to $30 million--now a 50 percent premium above competitive wages--this facility will remain closed and no revenues will flow in. If the entrepreneur abandons this project, the loss is $10 million (as the non-salvageable capital investment is written off). If the strikers' demands are met, the loss is reduced to $8 million ($40m--$10m--$8m--$30m). If there are no other options, it is better (loss-minimizing) to submit to the demands of the strikers than to shutter the facility. But the entrepreneur likely would learn an important lesson from the experience: channel future investments away from this unionized sector or locale in order to protect any expected returns from such appropriation.

Of course, a real-life entrepreneur might not submit to such opportunism very easily. Other options to be explored include not only appeals to courts, waiting out the strikers in the hope that their pockets might be emptied before the entrepreneur's, or hiring nonunion ("scab") workers which might require the deployment of "security personnel" to fend off the strikers seeking to take a facility hostage. In practice, these tactics and more have been attempted in job actions, but all are costly and uncertain; they might reduce the returns to opportunism, but they will not alter the fundamental lesson that the quasi-rent value of fixed and durable capital is vulnerable to appropriation by unions, and is thus more secure when and where unions are less powerful.

The Conquest

Though firms in Detroit and other American industrial cities had kept unions more or less at bay for the first third of the 20th century, the onset of the Great Depression created an extremely favorable ideological and political environment for unionization. The erroneous but widespread belief that falling wages and prices were a cause of the Depression rather than necessary adjustments to restore growth in employment and output contributed (along with many other factors) to the passage of several pieces of federal legislation that encouraged the cartelization of labor and product markets. Of central importance was the National Labor Relations (or Wagner) Act, which eliminated many of the strategies commonly used by firms to defend the quasi-rent value of prospective or existing physical capital. For example, it took labor disputes out of the courts and vested enforcement of the Act in a politically appointed National Labor Relations Board; prohibited several "unfair labor practices" judged to be obstacles to organization; and enforced exclusive bargaining and union pay rates for all workers whether union members or not--in Board-certified bargaining units. But, as we will see, by limiting defenses against the appropriative actions of labor cartels, such legislation would actually increase firms' reliance on the remaining ones--especially redeployment of productive capital to less vulnerable locations and input substitution. Again, this would have dire consequences for "union towns."

The Wagner Act was signed into law in July 1935; in August, the newly chartered United Auto Workers held their first convention in Detroit. The UAW correctly judged that piecemeal, plant-by-plant organizing efforts could not generate the market power needed to ratchet autoworkers' wages--again, already among the highest in American industry--to the desired heights. It cast its lot with those advocating industrial unionism, which sought to organize all the workers within a given industry into a single union (in contrast to craft unionism, in which various skilled trades unions could co-exist within a firm), and set its sights on the industry's largest enterprise: General Motors.

It is hard to overstate the tactical brilliance of the UAW's campaign to monopolize GM's labor force. (4) The UAW learned that the company had only two factories producing the dies that stamped out the body components for all its cars. If it could take control of them the entire company would grind to a halt; in effect, all the firm's assets could be taken hostage by job actions in just two of its facilities. GM had placed them in Cleveland and Flint, both somewhat distant from Detroit's increasingly militant labor climate. Flint was a "company town" that seemed especially defensible; in early 1936, fewer than 200 of its 47,000 GM autoworkers had joined the UAW, city officials and police were in GM's pocket, and spies were everywhere. One organizer, upon checking into a Flint hotel, was greeted with a phone call telling him to get out unless he wanted to be "carried out in a wooden box." In late December, 1936, the union's own spies learned that GM planned to move the all-important dies out of Flint, and the UAW quickly initiated what came to be known as the Great Flint Sit-Down. Though UAW membership was still less than a tenth of GM's total Flint labor force, by sitting down at their machines, occupying the buildings that housed them, and keeping all others out, such strikers could idle an entire facility--and, in contrast to more conventional tactics such as picket lines, do so with less risk that they would be attacked or arrested by police or have their jobs given to strikebreakers so that production could continue. GM tried all the usual counter-measures during this 44-day "siege from within," from injunctions issued by friendly local judges to assaults by police armed not just with tear gas but machine guns; all failed. And without the crucial supplies from these plants, production slowed or stopped everywhere else; eventually, GM's output fell from 53,000 cars per week before the sit-down to 1,500, and 140,000 of its 150,000 workers were idle.

On February 11, 1937, GM capitulated and recognized the UAW as the exclusive bargaining agent for its unionized workers. A sit down strike at the union's next target, Chrysler, won a similar agreement the next month. Thus legitimized, within a year UAW membership grew from 30,000 to 500,000--fully half in Detroit. Henry Ford would be a tougher nut to crack, vowing that the UAW would organize Ford "over my dead body" and attempting to use violence and intimidation to fend off organizers; an April 1941 sit-down strike at the River Rouge plant finally led to his surrender. (5)

Labor historian Sidney Fine (1969: 341) has, therefore, judged the 1936-37 GM sit-down strike "the most significant American labor conflict in the twentieth century." It not only gave the UAW the means to capture a large portion of the returns to the auto industry's capital in succeeding decades, but it demonstrated to laborers and employers the tactics that could lead to successful cartelization of labor supply in other industries. For example, iron and steel workers' unions had been moribund for decades, but in March 1937, U.S. Steel (of Pittsburgh and Gary), its management fearful of the stone sort of upheaval that had cost GM so dearly, signed a contract with the union that would become the United Steel Workers (though its full name better conveys its reach: the United Steel, Paper and Forestry, Rubber, Manufacturing, Energy, Allied Industrial and Service Workers International Union). "Little Steel"--which included Republic (of Cleveland and Chicago), Bethlehem (of Pennsylvania and Baltimore, and also including its subsidiary Lackawanna of Buffalo), Youngstown (Ohio) Sheet and Tube, National (of Weirton, West Virgnia, and Detroit), and Inland (of Chicago)--attempted the Ford approach of violence and intimidation, including the Memorial Day Massacre of 1937 (in which Chicago police opened fire on strikers and sympathizers approaching the Republic mill, killing ten). Like Ford, however, by 1941 Little Steel capitulated.

In sum, at the onset of World War II most of America's great industrial firms which, thanks to agglomeration economies were concentrated in cities throughout the East and upper Midwest--now faced labor cartels. These cartels needed some time to consolidate their power, so increases in employers' wage costs would be significant but gradual. Further, WWII and its aftermath, during which time America's industrial rivals' productive capacity suffered heavy damage that would be restored only slowly, insulated the unions and firms to some degree and for some time from the most severe competitive consequences of monopolistic and opportunistic prices for labor. (6) But the employers started to adapt immediately in ways that standard economic theory would predict--and that would ultimately help create what became known as America's Rust Belt. Union actions, clearly, were not the only reason that industrial cities would decapitalize, depopulate, and become poorer in the second half of the 20th century, but they merit inclusion on the list.

The Occupation

As we have seen, industrial firms recognized the vulnerability of their physical capital to appropriation and engaged in some defensive deployment strategies even before unions achieved full monopoly status; as union power grew, the pace accelerated. Just as, for example, the UAW had operated with near-military discipline and precision in defeating opposing forces in Flint and elsewhere, they would now see those forces retreating to safer environs followed by a stream of refugees.

In the decade following WWII, General Motors spent $3.4 billion, Ford $2.5 billion, and Chrysler $700 million on new facilities, almost all in rural areas "as a means of reducing wages and inhibiting union militancy in manufacturing cities like Detroit" (Sugrue 1996: 128). Detroiters decried these "runaway shops," but many simply followed the capital and the jobs it supported. From 1947 to 1958, manufacturing employment in Detroit fell 40 percent, a net loss of 134,000 jobs. Accordingly, between the 1950 and 1960 censuses, the city's population fell by 180,000, or 10 percent. In effect, the same multiplier effect that had made Detroit one of America's fastest-growing cities in the first half of the 20th century was now operating very powerfully in reverse--and well in advance of the racial tensions of the 1960s or the completion of the interstate highway system popularly assigned much of the blame for flight and sprawl.

And Detroit, of course, just illustrates more dramatically and rapidly the trends unfolding in other industrial cities. Unfortunately, city- or metro-area data on the extent of unionization are not available for the decades of the 1950s and 1960s, but data from the early 1970s makes clear that the strength of unions was negatively and significantly related to subsequent population changes in core cities. Figure I displays a scatter plot and trend line relating the proportion of a metro area's production workers that were unionized over 1973-75 and the core city's population growth between the 1970 and 2000 censuses. Cities with above-median unionization depopulated by an average of 7 percent over that period; those with below-median unionization grew an average of 32 percent.

Faced with above-competitive prices for labor, the threat of opportunism, and work rules that limited their flexibility, managers of industrial firms also engaged in input substitution, reconfiguring their production processes in important ways. Using land intensively in urban settings previously had enabled firms to realize the benefits of industrial agglomeration. Now they applied a different economic calculus, reducing their reliance on the more-expensive labor input and substituting capital and land, the relative cost of which had fallen. In describing the resulting trend toward sprawling, more heavily automated plants in low-density areas, observers have generally supposed that these productive technologies were new and superior and that adopting them was inevitable (given "capitalist greed"), if unfortunate. But the availability of cheap land outside cities was not new or unknown to capitalists, and neither was the ability of capital to substitute for labor. The trend, in short, was not an "exogenous change," but rather an adaptation to a new array of input prices. Had the relative cost of labor, land, and capital not changed, it is entirely possible that more manufacturing firms would have decided not to abandon the prospective benefits of urban location and agglomeration. We'll never know--nor will we know whether any technological spillovers of the kind described by Marshall might have, over the decades, made U.S. industrial firms more innovative and globally competitive. Defenders of unions like the UAW generally blame the declining fortunes of heavily unionized industries on bad managers selling poor products; they rarely contemplate whether such results become more likely once the "mysteries of the trade" are no longer "in the air," as it were. (7)

[FIGURE 1 OMITTED]

What we do know is that the consequences of capital flight and reduced labor demand were dire for the residents who remained in America's (formerly) industrial cities. Figure 2 shows a scatter plot and trend line linking a metro area's level of unionization in 1973-75 to the subsequent (1979-2007) change in the core city's real median income. Again, the correlation is negative and significant. Core cities with above-median unionization rates got poorer, their median real incomes falling an average of 7.6 percent over the relevant period; those with below-median rates of unionization averaged 4.5 percent growth in their real median incomes.

Of course, such summary statistics do not convey the enormity of the problems that result from or are compounded by diminished economic opportunities in cities. The sociologist William Julius Wilson (1997) has written forcefully of the economic, social, and cultural consequences of the "spatial mismatch" between labor demand and supply that was a by-product of urban deindustrialization. Abundant industrial capital had long made American cities an economic launching pad for successive generations of immigrants, but its flight left the most recent migrants to cities--especially African-Americans participating in the "Great Migration" from the rural South to northern cities over 1916-70--with far more limited economic options. For many, Wilson has argued, this has meant persistent joblessness; their detachment from the labor force and limited exposure to the working- and middle-class populations that followed the capital out of cities have contributed to the creation of an urban underclass. For this population, capital- and job-flight kicked off an unwholesome cycle that has adversely affected a host of social variables, from family formation to educational attainment to propensity to engage in crime; as Murray (1990) has pointed out, this cycle is by no means racially exclusive. The foregoing suggests that, in part at least, it began with the cartelization of urban labor markets.

[FIGURE 2 OMITTED]

What Does the Future Hold?

It is easy to be pessimistic about the future of the American city. Much of Detroit is in ruins; its median household income, once 29 percent above the national figure, is now 44 percent below it; its poverty and crime rates are over three times the nation's. It would be nice to be able to say that Detroit's experience is an aberration, mad that other formerly industrial core cities are healthy. In fact, however, most are just "less unhealthy." Since 1950, St. Louis, Pittsburgh, Buffalo, and Cleveland all have suffered population declines greater than or equal to Detroit's 50 percent; Newark, Cincinnati, Baltimore, and Philadelphia have lost roughly a third of their populations. All told, about 5.5 million people exited America's largest cities in the second half of the 20th century, and many of those who remained experienced declining economic and social well-being on many dimensions.

What is more, there are troubling clouds on the horizon even for those cities that many consider to have transited successfully toward service-based economies from goods-producing ones. Unions have adapted, too. While private-sector membership has declined (as manufacturing firms have engaged ha defensive deployment and increased automation), public sector unions have shown enormous growth. Though these organizations often face limits on their capacity to engage in opportunism (e.g., many states or municipalities enforce no-strike laws for at least some governmental functions), they have been ingenious in their acquisition and exercise of monopoly power. In particular, they have employed the bloc-voting power and dues-paying capacity of their members to exert significant influence on municipal office-holders. This has enabled them to capture sizeable quasi-rents associated with the public capital (buildings, infrastructure, etc.) concentrated in cities. Often, the near-term costs of above-competitive wages can be disguised by raiding cities' capital budgets and deferring maintenance, and some of the monopoly returns can be taken in the form of long-delayed payments (e.g., pensions)--all of which, of course, has the advantage for current officeholders of coming due on the watch of others. Nevertheless, the burden of such behavior for taxpayers is growing. As the quality of public capital erodes and taxes rise in particular cities, we will inevitably see some of their residents "voting with their feet" (as industrial employers long have been doing), with ill consequences for these cities' futures.

On the other hand, there are at least a few case examples of cities that provide reasons for optimism--and, in fact, hint at some genuine strategies for urban revitalization. Consider Boston and San Francisco. Between 1947 and 1972, Detroit lost 47 percent of its manufacturing jobs, but Boston lost 42 percent and San Francisco 28 percent. Between 1950 and 1980, Detroit lost 35 percent of its population, Boston 30 percent, and San Francisco 12 percent. In other words, differences in the fortunes of these cities in the three decades after WWII were more a matter of degree than of kind. Then, Boston and San Francisco turned things around. While Detroit's population fell another 21 percent over 1980-2000, Boston's grew 5 percent and San Francisco's 14 percent. Between 1979 and 2005, Boston's median household income grew 38 faster than the nation's, and San Francisco's 51 percent faster. What happened? Did racism materialize in Boston and San Francisco in 1950 and evaporate in 1980? Did a taste for suburban lawns do likewise?

The most common explanation for the renewal of these "superstar cities" is that they participated in the high-technology boom of the 1990s. But, clearly, the inflection point for the revival of these cities predates that boom; as well, high-tech employment has flourished elsewhere, and not all of the formerly industrial cities nearby have joined in. (8) What did coincide with the turning points in Boston mad San Francisco were two statewide property tax revolts that suddenly, significantly, and favorably transformed the climates for capital investment in these two cities. California's Proposition 13 (1978) and Massachusetts' Proposition 2 1/2 (1980)--both passed over the objections of the political leaders of those states' largest cities did not just greatly reduce property owners' annual tax liability on existing and, more importantly, new capital investment (in both cities by roughly two-thirds), they also secured their property against appropriation with formulae limiting future tax hikes. The results were immediate and substantial: inflows of capital, repopulation, and enhanced quality of urban life on many margins.

The lesson is that increased capital-friendliness is a necessary condition for a successful, enduring, and organic urban redevelopment strategy. But policymakers should not just focus on tax policy in this regard. Even Boston and San Francisco have their critics: in particular, such cities have tended to ignore how inefficient, inequitable labor market institutions can chill capital investment that might be of greatest value to city residents who are not well-equipped for jobs in, say, finance or biotechnology, and not beneficiaries of the "Potemkin villages of art museums, performance centers, tourist attractions, luxury hotels, and condos enthusiastically promoted to locals and visitors as evidence of urban renewal" (Kotkin 2006: 25). In effect, too much redevelopment policy is focused on attracting middle-class residents, and not enough on creating an urban middle class.

The good news is that there are some policies that have demonstrably improved the environment for investment in the kind of physical capital that can fuel growth and enhance employment opportunities in cities. For example, Holmes (1998) has shown that right-to-work laws (which simply ban "union shops" in which all employees are required to join the union) have a positive and significant effect on manufacturing activity and employment. All else the same, manufacturing employment increases by a third when one crosses the border from a non-right-to-work county to one with such a law (which might be correlated with other pro-investment policies, of course), and growth in right-to-work areas is much higher. Other initiatives (e.g., charter schools and voucher programs) that temper the monopoly power of public employees' unions and increase the quality and/or reduce the cost of key public services also have the capacity to improve the fortunes of struggling cities. In most eases, of course, entrenched interest groups within city borders will fight tooth-and-nail to preserve the status quo. In many cities, the flight of capital and capitalists has proceeded for so long that scarcely any political competition or sentiment for pro-capital policies remains. As Boston and San Francisco demonstrate, however, sometimes statewide referenda can rescue a jurisdiction from its own failed policies. Those who care about the future of cities and their residents should heed the lesson and get to work on similar initiatives; there is no time to waste.

References

Atleson, J. B. (1998) Labor and the Wartime State: Labor Relations and the Law during World War H. Champaign: University of Illinois Press.

Bureau of the Census (various years) County and City Data Book. Available at www2.1ib.virginia.edu/ccdb.

Fine, S. (1969) Sit-Down: The General Motors Strike of 1936-1937. Ann Arbor: University of Michigan Press.

Freeman, R. B., and Medoff, J. L. (1979)"New Estimates of Private Sector Unionism in the United States." Industrial and Labor Relations Review 32 (2): 143-74.

-- (1984) What Do Unions Do? New York: Basic Books.

Glaeser, E. L.; Kerr, W, R.; and Ponzetto, G. A. M. (2009) "Clusters of Entrepreneurship." NBER Working Paper, No. 15377.

Hirsch, B. T. (2007) "What Do Unions Do for Economic Performance?" In J. T. Bennett and B. E. Kaufman (eds.) What Do Unions Do? A Twenty-Year Perspective. New Brunswick, N.J.: Transaction Publishers.

Holmes, T. J. (1998) "The Effect of State Policies on the Location of Manufacturing: Evidence from State Borders." Journal of Political Economy 106 (4): 667-705.

Klein, B.; Crawford, R. G.; and Alchian, A. A. (1978) "Vertical Integration, Appropriable Rents, and the Competitive Contracting Process." Journal of Law and Economics 21 (2): 297-326.

Kotkin, J. (2006) "Urban Legend." Democracy: A Journal of Ideas, No. 2 (Fall). Available at www.democracyjournal.org/article .php?ID=6483.

Marshall, A. (1920) Principles of Economics. 8th ed. London: MacMillan.

Murray, C. (1990)"The Underclass Revisited." American Enterprise Institute, Papers and Studies. Available at www.aei.org/paper/ 14891.

Rubenstein, J. M. (1992) The Changing U.S. Auto Industry. New York: Routledge.

Sugrue, T. J. (1996) The Origins of the Urban Crisis: Race and Inequality in Postwar Detroit. Princeton, N.J.: Princeton University Press.

Wilson, W. J. (1997) When Work Disappears: The World of the New Urban Poor. New York: Vintage Books.

(1) It is also worth noting--so that race-based theories of urban form can be kept in perspective--that Detroit's black population increased 20-fold, to 120,000, from 1910 to 1930 and would increase another 150 percent (to 304,000) by 1950. That the city was attractive to both whites and blacks during this period was understandable: by 1949, the median family income of Detroiters was higher than that of any other city in America except Chicago (whose residents enjoyed a 1949 median family income exactly one dollar higher), and 29 percent above the national figure. In other words, while Detroit's economy functioned well its large and rapidly growing minority population was not a destabilizing force but both a reason for and symptom of its success; it was only after its economy began to erode and its population began to fall that commentators began to draw the conclusion that racism (no doubt present, but likely showing little variation over the decades) was a key driver in this process.

(2) The quasi-rent value of any asset is the excess of its value in its current use over its salvage value--that is, its value over that in its next best use. The potentially appropriate portion of any asset's quasi-rent value is that amount, if any, in excess of its value to the next highest-valuing user (see Klein, Crawford, and Alchian 1978).

(3) It is arguable that the agglomeration economies that made, say, Detroit attractive to automotive entrepreneurs also gave rise to site-specific quasi-rents that unions could extract with only marginal effects. This would imply that the onset of unionization in Detroit (and other, similar industrial cities) might result in a slowing of investment and employment growth, but not a wholesale redeployment of industrial capital. The evidence suggests otherwise, however, and will be discussed further later.

(4) See Fine (1969) for a thorough history.

(5) Still, Ford reportedly planned to break up his company rather than sign a contract with the UAW, but his wife threatened to leave him if he did not cooperate with the union so that the family business could survive and their son and grandsons continue to run it.

(6) The main tool of opportunism was the wildcat strike. During WWII, for example, when lives hinged on abundant and steady output from auto plants that had been retooled to produce war materiel, the UAW made a no-strike pledge. Nevertheless, in 1943 there were 153 wildcat strikes in its plants, and in 1944 there were 224 involving over half the workers in the industry. Across all sectors, according to Bureau of Labor Statistics data, in 1943 alone over 13 million mandays of production were lost to strikes--equivalent to idling over 53,000 full-time workers for the year (Atleson 1998: 145-47).

(7) One source of inefficiency associated with the threat of opportunism merits mention: U.S. automakers' reluctance to employ "just-in-time" production techniques, in which components arrive at assembly plants shortly before needed, thus reducing storage and carrying costs. Their Japanese competitors, with more compliant unions, exploited such economies to great advantage, but U.S. firms could not because interruptions in supply due to wildcat strikes or other labor frictions would cause costly ripple effects throughout other facilities (a la Flint); they adapted by holding much larger inventories of parts. Such extra costs generally are not counted in evaluations of union-related sources of competitive disadvantage for U.S. firms, which usually focus on wage and benefit disparities, but they are important nonetheless.

(8) Surveys suggest that the three largest high-tech employment centers in the United States are the San Francisco Bay area (including the Silicon Valley), greater New York, and the Washington-Baltimore metro area. It would be pleasant to report that the old industrial cities like Newark and Baltimore are showing San Francisco-like signs of revival, but they are not.

Stephen J. K. Waiters is Professor of Economics at Loyola University Maryland.