CIAO DATE: 8/00

Globalization and Its Impact Across Sectors: Developmental Effects and Public Sector Implication in the United States

Gerald A. Hendrickson

International Studies Association
41st Annual Convention
Los Angeles, CA
March 14-18, 2000

Introduction

Globalization is not a topic that immediately attracts the attention of most public administrators. However, understanding the implications of globalization for labor markets across economic sectors can affect the way public sector managers approach personnel planning. This paper attempts to enhance this understanding in three ways. It will carefully examine some of the issues around globalization, subject the ideas from the dominant perspectives to empirical testing and propose alternative hypotheses. In the end, the mission of this paper is to provide some insight into the kind of labor environment that public sector managers will encounter as their plans come to fruition in the future.

The need for this kind of analysis comes from the increased emphasis put on planning in the public sector in the United States in a time when budgetary conservatism is prevalent. The “re-inventing government” movement encourages many public sector entities in the United States to more effectively plan for the future and anticipate the needs of future generations (Osborne and Gaebler, 1993). However, it does not provide funding for the bulk of local governmental entities that try to carry out this mission. Local funds for this purpose are scarce as well due to the specter of ever-present budget cuts that affect essential service delivery and widespread cynicism about the role and capability of the government. As a result, public managers who seek to heed the call of anticipation often engage in classic “satisficing” behavior (Simon, 1955). They try to “bound” the future themselves instead of hiring the expensive, analytical staff required to perform the sophisticated analysis they need.

A typical end result of this attempt to plan is a reliance on trend analyses. Unfortunately, there are multiple problems associated with the use of these analytic products. First, they are often ill suited at describing and identifying the kinds of trends and issues public managers would gain insight from examining because they come from commercial sources who write to an audience dominated by business travelers. In many cases, taking these specific circumstances into account may lead to the complete dismissal of the claims of the trend analysis (Schroeder, 1980). Second, these analyses are also more prone to reflect mainstream fads. As trends emerge and devolve, and the publications about them multiply, any intellectual content often becomes watered down and the trend itself loses much of the meaning it originally had. Third, trend analyses also have a propensity to either over-predict or under-predict the actual effect of a given trend on a given manager’s environment due to reliance on highly aggregated data.

This paper represents a clear improvement from prior analyses of this sort. It pays attention to the intellectual debate at the heart of the globalization discourse and subjects the proposed theories of that debate to thorough empirical testing. This approach draws attention to Marxist world systems and neoliberal political economy as the two dominant schools of thought in this issue area and delineates the dominant way globalization is believed to impact core labor markets in each perspective. It finds both of these mechanisms to be wanting, and brings in development and consumption theory to compensate for these lapses. In addition, this paper uses sectoral data to test this set of mechanisms and reveal some of the underlying structural changes at work in aggregated core labor markets.

This approach to the study of the impact of globalization on core labor markets reveals a great deal that had previously gone unnoticed. The deeper look offered in this paper provides evidence that globalization and traditional economic development processes are taking place simultaneously in the United States. This simultaneity suggests some of the effects attributed to globalization by contemporary scholars can be seen in sectoral changes coming from traditional developmental processes. In addition, this result also implies that scope-limited industries attenuate globalization’s impact on the domestic economy in ways that have often been ignored.

This structural change in the economy is likely to affect the composition and functioning of the public sector in the United States in a fundamental way. One of the primary ways the public sector will feel the impact of this trend is in the types of personnel likely to be filling public sector positions in the future. Typically, the public sector will become the choice of work of employees from lagging industries, many of whom do not have the skills to participate effectively in many of the complex tasks that characterize the public sector operating environment. As a result, the public sector will feel a continued pressure to provide training opportunities to shore up the sagging skill sets of its new employees. At the same time, the sector will continue to be bombarded by high turnover in the ranks of employees who have opportunities in the private sector.

Defining Globalization

Globalization must first be defined before any of these latter goals can be achieved. This task is not as easy as it seems. There are a number of issues and topics that come up when globalization is discussed. Many alleged experts talk about the globalization of American values, information technology or terrorist activity. The literature is filled with references to new world orders and the ability of the Internet to put the developed and developing worlds on common footing with each other. In essence, it appears as if the tag of globalization is applied to almost every ongoing trend, and as a result, the term itself now has little meaning.

It is apparent from this breadth of application that this bloated definition of globalization needs to be clarified so it retains any real meaning. The path to clarification starts with a look at the significance of the descriptor by itself. This most basic of steps alone often enhances the theoretical clarity of the topic under consideration - a quality the globalization literature clearly requires in order for any rigorous analysis to be worthwhile.

The term globalization alone implies a trend of expansion from an earlier system based on smaller units of analysis like nation-states or regional blocs to a system existing worldwide. This factor limits the span of issues that are feasible to consider in this fashion. Globalization can not occur in systems that are already global in scope. In such cases, a given analyst may actually be attempting to describe a condition of increasing integration between constituent elements within the system. This difference is an important element in the confusion that is so much apart of the issue area itself.

The second step toward a more precise definition for globalization involves looking at its traditional usage. In terms of its intellectual history, the mainstream literature concerned with globalization tends to have a Marxist heritage with minor Weberian undercurrents. As one would expect, this scholarly foundation compels theorists in this area to perceive this change in the “dynamic density” between nations largely in terms of production relations. Thus, the literature surrounding this topic tends to focus on two primary channels for globalization to occur – the globalization of capital and the internationalization of labor. In the end, however, only one channel actually presents a valid “globalization” trend.

The so-called “globalization” of capital falls into the category of being a trend of integration. Since the emergence of the modern international political and economic systems in the late 1500s, capital assets have been notoriously difficult to restrict to national borders (Wallerstein, 1976). The only difference between the beginning of this dual system and the contemporary time period is the velocity of capital’s application and its distribution of costs and benefits. Put more simply, the vehicle for capital has moved from physical to electronic transport and the volume of capital transactions has increased accordingly. Thus, international capital is merely going through a process of enhanced integration, not a “new internationalization of capital”(Wallerstein, 1976).

Labor is another case altogether. In the early period of the dual system, domestic firms used domestic labor efforts with generic capital assets to produce goods. As the period has progressed, however, domestic labor inputs have been increasingly combined with international labor inputs to produce goods for firms with origins outside the domestic environment: foreign firms and multinational corporate entities who appear to be “nationless.” The scope of international labor involved in this type of production is continually growing as more and more countries throw aside state development schemes in favor of the market-based alternative. It is apparent that a new internationalization of labor is taking place as local labor efforts now are used to promote globally consumed goods.

The conception of globalization as the internationalization of labor (also known as the internationalization of production or the cross-country substitution of labor) forms the base definition of globalization used in the confines of this paper. This paper asserts that globalization is defined as the constrained process by which firms spread across national boundaries in their pursuit of cost-minimization and profit maximization. It should be noted this definition choice for globalization does not preclude the utility of other definition used in other more narrow contexts. It is simply the broadest definition of globalization with the greatest intellectual foundation.

Review of Literature

Globalization is a topic with a complex discourse at dual levels in multiple issue arenas. Thus, an examination of the literature focusing on this topic reads like the intellectual equivalent of a travelogue. There are large amounts of literature in both the policy and theoretical areas. The more pragmatic literature that focuses on globalization deals primarily with distributional issues associated with the impact of the trend on local economic and political systems. The mass of theoretical literature concerns itself with a number of questions about the trend’s existence, its real impact and potential limiting factors.

It should come as little surprise that the policy literature is tainted with the biases of political struggle. The perspective of a given policy actor is often highly correlated with their position. This type of perspective formation by self-interest maximization often leads policymakers to spike discussions with rhetoric borne of the perennial North-South debate or of economic nationalism and geopolitical isolationism to curry favor either internationally or domestically. Clearly constructed sound analysis is often ignored while those who hold other perspectives and positions sound alarms, hold celebrations or dismiss the other’s concerns altogether.

In the policy arena, globalization is often construed as either evidence of a conspiracy or another opportunity to transform the world. In less-developed countries, the debate over globalization focuses more on the normative aspects of this globalization process. It is seen as a foregone conclusion for their future with only revolutionary or experimental alternatives. In the industrialized world, the debate indicates that there is more uncertainty about the processes’ effect on national attributes like wealth, power and prestige. However, unlike the discourse in the developing world, there is a sense that the trend can come to a halt with adverse domestic reactions on the part of any of the great powers.

The developing worldview is nicely summarized by Malaysian Prime Minister Dato Seri Dr Mahathir Bin Mohamad in a 1998 speech. In his speech, Mahathir suggests that globalization is inherently unequal – that the result of this “one level playing field for all nations” is the evaporation of chances for less developed areas to escape their current position in the international order. It is his fear, as it is many others, that the more highly developed firms of the United States, Japan and Western Europe will outcompete indigenous firms and rob areas of the chance to grow and develop. In essence, it is seen as the key that locks-in the current unequal international division of labor

The policy literature in the developed world is not quite so neatly summarized. The governments of many core nation-states are split on the issue. It is seen as a threat to national security as well as an enhancement. Competitiveness effects also appear to be a political wash. The civil societies those governments represent are sharply divided as well. WTO protestors often consider the scope expansion of rules concerning worker and environmental conditions to erode more stringent protections in the core. Fears about cultural erosion due to migration have been linked to improvements in the political standing of right wing reactionaries and their parties, such as Jorge Haider’s Freedom Party in Austria and Le Pen’s National Front in France.

Like the work on globalization in the policy realm, the more academic work on the topic often suffers from a lack of objectivity. The body of theory that academia brings to the debate carries with it the baggage of entrenched conflict between opposing schools of thought in the international relations and development communities. The result of this devotion to base level assumptions is that few scholars if any appreciate globalization’s true impact. In many issues, the conflict boils down to a debate between Marxist world systems scholars and neo-liberal political economists. Other schools of thought tend to deny the existence of the trend itself.

Both of these particular schools tend to see the trend as an indication of fundamental change in the international system. For the Marxist world systems scholar, it is the instrument by which the capitalist world economy comes to its ultimate end. For the neo-liberal political economist, globalization is the mechanism by which substate actors receive power to contest and limit the authority of the state. Immanuel Wallerstein and Dani Rodrik are the primary apologists in this debate, representing the world systems scholars and cautious neoliberal viewpoint respectively.

Wallerstein suggests the internationalization of labor is taking place via a fundamentally different process than the process of capital internationalization, which occurred sometime prior to the formation of the existing dual international system order (Wallerstein, 1997). Part of the difference between the two emerges from their fundamental differences. Capital is fungible and relatively easy to conceal as one crosses the boundaries between nations. Labor is easier to restrict because it is nested within a set of social relations that can give indications about migration inside the domestic environment and is more difficult to conceal in crossing international borders.

According to Wallerstein, this lack of labor mobility relative to capital drives the internationalization process. Owners of capital assets in the core often experience lower returns than they would in the periphery or semi-periphery because labor standards have been bid up through the political process over time. Wishing to maximize profit, these capital holders than move their assets to international locations that minimize their labor and transportation costs outside of the original country. As the conditions in the new area improve for workers through the political system, the profit margins decline for the profit takers - precipitating yet another movement to an area with lower costs. In essence, foreign capital is more likely to move to domestic labor rather than vice versa.

The long run implications of this movement are obvious. Core capital will move to the periphery. It is clear the largest gains for capital holders are to be found in the periphery states. These states tend to have highly rural populations who are geographically dispersed and poorly educated. In addition, the majority of exchanges tend to be outside the official goods market. This combination initially ensures the capital owner access to a work force that has significant barriers to organization and a lack of knowledge about the fair market value of their labor - all factors that suggest a low cost source of labor.

The clearest implication of this process is that the internationalization of labor clearly diminishes the demand for labor in the area of the core and the semi-periphery as capital moves to take advantage of the lowest cost labor in peripheral areas (see Figure One). This decline in demand reduces wage levels (from w0 to w1 in Figure One) and employment (shown by the movement from L0 to L1). The presence of greater substitute employment in the core areas (Wallerstein, 1976) suggests that the impact of the trend will be greater in these areas than if similar changes occur in semi-peripheral states (see Figures Two and

Figure One: Impact of Declining Labor Demand on Employment and Wages in the Core and Semi-Periphery

Figure Two: Effect of Periphery Labor Substitution on Core Labor Markets

Figure Three: Effect of Periphery Labor Substitution on Semi-Periphery Labor Markets

Three). As firms expand their production chains into the periphery, the wages offered across markets declines (from w0 to w1) and domestic workers in those industries are put out of work (unemployment is indicated by the distance between L0 and Ld). The presence of greater substitutes in the core countries generates the flatter labor demand curve in Figure Two and is responsible for the difference between the core and semi-periphery in their response to the openness shock. Oddly enough, the semi-periphery suffers a smaller unemployment effect than that witnessed in the core. In the end, however, if this theory holds true, both areas should exhibit lower levels of employment during the internationalization period.

Rodrik does not dispute the central hypothesis that emerges from Wallerstein’s analysis (i.e. the extension of product chains to the periphery creates lower levels of labor demand in the core countries). He suggests that a second change is taking place, inferring that more substitutes exist for given types of labor in a world economy of global product chains than in one solely reliant on either regional or nation-state scale productivity networks. This increase in substitutes makes the domestic labor market in the core increasingly sensitive to changes in the wage rates as a whole. (Rodrik, 1997)

Rodrik does dispute the causal factor behind Wallerstein’ inward shift in the labor demand curve. He proposes a general inward shift in labor demand caused specifically by a large decline in the demand for low skilled workers. It is this specific subset of the labor force that is affected by the forces of globalizing production chains. Thus, the migration of capital is diminished as a source of the shock to employment as a result. The increase in labor intensive exports from peripheral countries that occurs with the adoption of a more open economic system is viewed as the more significant factor of the two in altering factor markets in general, and labor markets specifically.

Rodrik’s notion of increased wage elasticity in labor markets follows from observations made about rising wage inequality between workers educated at the high school level and those with higher levels of education attainment and a possible connection with trade. First, wage inequality and the durable goods trade deficit are observed to be cointegrated (Borjas and Ramey, 1994). This evidence suggests that as imports increase, wage inequality increases in durable goods producing industries. Second, worker immigration may be perpetuating the wage gap as well (Borjas, Freeman and Katz, 1996). Indeed, the inflow of workers at low wage rates may be generating a kink in U.S. labor supply that makes it much more elastic at low wage levels.

Rodrik and Wallerstein’s debate hinges on whether the change in the labor market over time is the result of a demand shift or in a change in the wage elasticity of workers in the labor market. These two mechanisms imply different kinds of economic relationships between inputs in the production process. The decrease in labor demand position suggests that in the presence of globalization there is no change in the number of available substitutes for a given kind of labor and that the willingness to pay for that labor decreases at every point. The increase in elasticity argument suggests that globalization increases the number of substitutes in the market such that low-wage workers will be employed to the point of exhaustion. It is of course, more likely that both changes have occurred, but both theorists have chosen to avoid their exact estimation.

There are some clear difficulties with both of these perspectives. Wallerstein paints a picture of a semi-periphery abandoned by the expansion of production chains into the periphery. However, his discussion of labor markets in the core actually suggests the greatest losses will occur in the developed countries because he asserts these workers face a market with more substitutes for their labor. Rodrik’s perspective does not suffer from inherent contradiction so much as it fails to go far enough. Rodrik suggests the globalized world has the elements to create a countertrend toward isolation, but neglects to consider the location and size of these conditions in the domestic economy.

Theory Made Applicable by Studying Sectors

Both of these problems can be linked to these two apologists’ level of analysis. The examination of data generated by third orders of aggregation creates an observation environment that often indicates volatility with smoothness. Concentrating on a smaller unit of analysis can provide insight into structural changes that come with openness by expanding the set of applicable theory. In the case of globalization, concentrating on sectors allows for the application of development theory and intermediate types of economic analysis.

To this point, the debate over globalization has paid little, if any attention to the role that national economic structures have in determining the impact of product chain expansion. In both cases above, the state is perceived as too weak or too bound to the old international system to matter. As a result, the sectoral changes that have occurred in advanced industrial societies seem to have disappeared from the perception of these analysts. It is as if the basic precepts of development theory have disappeared.

Observation of low-income and high-income countries and their GDP composition reveals at least one primary trait. As nation-states grow and develop, they turn away from subsistence to concentrate on the acquisition of wealth. This fact is seen in the high agricultural components to the poorest countries ₁ and the relatively low importance of agricultural production in the advanced industrial countries. ₂ This fact extends to primary manufacturing and the export of raw materials as well. Export oriented industry (whether it be the extraction of mineral or the assemblage of relatively low technology items) focused on the acquisition of foreign exchange to prop up contrived exchange rates for urban consumption will be substituted against as well (Bates, 1981).

The economy will most likely substitute growth in the service sector for decline in other sectors as it increases in income. Service sector growth occurs when the nation-state has enough resources to keep itself from the worry of subsistence and can concentrate on the welfare of its citizens. As it grows, the positions that existed in other more subsistent level industries will diminish in number as productivity increases in those sectors, and service positions will proliferate.

Not all of these positions can be assumed to disappear with technology gain. It is logical to assert that in an open economy, the industry that is diminishing in an advanced state is likely to migrate to a less advanced state given the proper conditions. Technology changes, even those in the organizational sense, do not come without cost. It is likely that not every firm in the factor market will be of sound enough fiscal position to stay in the presence of lower cost entrants or existing firms that innovate.

A cursory examination reveals that there are a significant number of economic mechanisms that cause the movement of industries. Migration can and does occur through the natural functioning of the market. As service industries become more prominent in the market for labor, it is likely that service firms will drive up the demand for labor – making it more expensive for existing firms in the extraction or manufacturing pursuits to hold on to their workers. That addition to these firms’ variable costs will push their marginal cost schedules higher and diminish the economic profits that each sees to the point that losses may drive many firms to other areas of production. It is unlikely that these firms will simply transform themselves into a producer of some other commodity in the short run – the acquisition of human capital and information is too high. They may concentrate on another aspect of their business, but most likely, they will follow the economic gradient.

There is also another reason to doubt the impact of such industrial movement away from a particular community as well. It appears as though the supply of workers in diminishing sectors decreases in elasticity as industries leave. The probable mechanism for this increase in elasticity is the movement of workers from one sector to another to avoid the problem of layoffs and prolonged unemployment. Thus, some populations are entrenched due to their skill sets and are likely to go to work anyway despite the lower wages offered by the employer in the presence of globalization.

In addition, there is significant debate as to how this process of development takes place given what appears to many scholars as a fairly rigid division of labor in the international system. Many scholars in the development community suggest breakout development comes from adopting the appropriate developmental paradigms (Stallings, 1995). Others suggest the path to development begins with involvement in goods that have a “multi-dimensional conspiracy” in favor of development (i.e. specializing in a product like machine woven cloth in the case of Great Britain during the 1800s or automobiles in the case of the United States in the 1900s) (Hirschman, 1977). The timing of involvement in those sectors is almost as important as the decision to pursue a given good, because involvement on the upswing of this “product cycle” creates wholly different effects than lagging involvement (Vernon, 1966).

This last viewpoint leads to an interesting observation. If some sectors create “multidimensional conspiracies” for development and constitute so-called, “leading sectors” than other sectors must be lagging behind and should generate little in the way of expanded economic opportunities. This bifurcation should lead to some form of difference in input markets and the way that these markets respond to shocks in the form of diminishing barriers to input substitution and the global expansion of production chains because they are “derived” demands. Thus, this leading/lagging sector distinction could be a primary source of variation in the way sectors respond to globalization pressure and could provide refutation for the idea that all sectors increase in wage elasticity with the pressure to globalize.

If one considers these distinctions, it is clear that development theory provides clear alternatives to Wallerstein and Rodrik’s mechanisms for labor internationalization. Sectors that typically have large barriers to entry and exit in the form of capital investment are likely to remain relatively concentrated in areas close to their markets, due to the cost of transporting products from some international destination. As a result, local subsidiaries are more likely to emerge in new areas rather than the vertical integration one often expects in many markets. Leading and lagging sectors are likely to have different responses to globalization. The lagging sector will typically exhibit much greater wage elasticity than that in evidence in the leading sector because of its longer exposure time and greater technological diffusion than its more restricted counterpart.

If either of these hypotheses is correct, it can be said that at least part of the capital migration hypothesis is faulty. It is clear that this migration does not have to occur through the issuance of entitlements by some governing body alone. Indeed the development of government programs of social insurance may very well occur long after the initial migration has occurred – violating the basic law of causality – chronology of events.

Another dividing line with potential impact on how labor markets are affected by globalization is found in the notion of product scope. The idea of product scope is a simple but essential concept. It follows from the notion that the demand for labor in a given industry is sensitive to many of the same pressures that affect the demand for products. If this is true, it provides a clear counterargument for the notion that only unskilled workers feel the effects of openness. If one believes in the notion of product scope, it is reasonable to expect that many positions in both high skill and low skill areas are subject to export in some sectors while positions in other sectors are not.

A clear example of this lack of sensitivity to trade can be observed in the service industry. Service industry positions at either end of the skill spectrum are not subject to export. Physicians and attorneys represent two sets of service positions that can not be filled by physicians and lawyers overseas – there is a clear regional scope to their service and market. Likewise, the employee at McDonald’s only serves a customer in a region – the product of their labor can not be exported. Even some intermediate skill positions in the service industry are insensitive to openness. Financial services, such as real estate brokerage, are a bit more complex in scope, but essentially sell immediate regional exchanges and information that may be unavailable or unimportant in other markets.

In general, manufacturing jobs can be exported. Consumer durables can be made almost anywhere and shipped to any other place on the planet given a few caveats (i.e. the presence of electrification for consumer electronics and appliances, etc.). The only set of goods in this product definition that might be affected by globalization is the area of food and food product manufactures. Technological improvement in the areas of preservation and refrigeration work against this regional dependency. In a sense, the good made through manufacturing exists with little regard for the legal, cultural, and time constraints that diminish the utility of other goods before they can escape regional or national boundaries. That quality alone lends itself to the exportability of production positions overseas more than any other factor.

As a result of the observations in both of these areas, this paper argues that globalization has distinctly different impacts across four types of sectors. Lagging sectors with global scope will typically have the greatest sensitivity to changes in wages because their positions are the most exportable due to their products and they represent industries in decline, who specifically require cost savings. Scope limited, leading sectors will be the least wage sensitive due to the converse of that logic. Scope limited, lagging sectors and global leading sectors will exhibit intermediate wage elasticities, with the global sector becoming more wage elastic at the higher rate. In the empirical analysis, this set of propositions will be tested in concert with Wallerstein and Rodrik’s hypotheses.

Methodology

As stated before this analysis of globalization uses a different method in examining the issue than many other analysts who have previously written on this topic. Traditional globalization theory often uses aggregate level data to validate its conclusions about the social, economic and political structures resulting from the changes in the ongoing pattern of integration between nation-states. In doing so, it loses a great deal of information about the compositional changes that occur within a country’s economic structure in the presence of these forces. When this data is used in analysis, it weakens the theoretical position of these theorists. This data causes analysts to ignore other trends by restricting the variability of the explanatory variables. Often the only action observable in this data is the simple ebb and flow of integration of nation-states that has occurred since the 19th century (Krasner, 1976).

The push for sectoral analysis to take the place of aggregated analysis in this case is motivated by the estimation bias that emerges from OLS or GLS parameter estimation using aggregated data. In essence, the use of aggregated data when sectoral or individual level data is required leads to a variant of the errors-in-variables problem which is characterized by parameter estimates biased below their actual values and are correlated with the model’s error term. This bias can be shown in the following simple demonstration where the dependent variable (Y) is only a function of a constant and the error term:

To avoid this errors-in-variables problem more thoroughly, this research also takes a sector by sector approach to characterizing the wage elasticities that emerge from each sector. An example of each sector in each of the categories mentioned in the preceding section is used to represent each type as is seen in the following schematic:

Figure Four: Typology of Tested Sector Selections

The apparel sector was mentioned earlier as being an example of a sector that had generated a multi-dimensional conspiracy for development in the case of Great Britain early in the Industrial Revolution. Computer manufacturing is arguably the manufacturing sector on the leading edge. Garment services as a sector refers to dry cleaning and laundromat services, typically services that can not be traded due to the transitivity of their effect. Financial instruments like insurance are limited in the sense that they often only apply in a given institutional or national context.

The data in this treatment comes from two primary sources: the Bureau of Labor Statistics within the U.S. Department of Labor and the National Bureau of Economic Research. As a result of the reliability of these sources and the continued attention these institutions give to potential problems of over- or under-estimation of population statistics like inflation, problems of measurement error, though potentially significant, are assumed to be negligible for this data. The data is time-serial in nature and spans the 1947-1998 timespan. All of the numeric information used in attaining the econometric results in this paper can be found at http://stats.bls.gov/datahome.htm via the use of the selective access database or at http://www.nber.org/nberprod.htm.

This study treats the quantity of labor in the marketplace as the total number of worker-hours per time period in a given industry. The worker-hour metric is used because workers in the sectors of interest are either paid via an hourly wage standard or they have greater hour flexibility. Typically, each year represents an observation, but in sectors where data is sparse due to short periods of statistical reporting, six-month periods are used to enhance measurement precision.

Comparing the sectors in worker hours shows an interesting trend. Finance hours outpace apparel hours by a factor of five in the last reported time period. While it is true that finance hours are always greater than hours spent in apparel manufacturing during this time period, it is important to note that the gap between the two is increasing. One can see potentially two effects from this data. Sectoral reallocation may be driving the finance industry forward while removing labor from the manufacturing of apparel. Globalization pressure may also be pushing clothing manufacturers to reallocate their productive assets to other countries. Both service sectors appear to be demonstrating a shallow growth trend.

It is easy to see that finance and computer wages have always been higher than those in the apparel and garment services industries. However, closer examination of the graph reveals some interesting traits of these four markets. The real wage in 1998 is roughly similar to the real wage rate that was offered in 1947 for apparel workers. Two factors potentially contribute to this stagnation: lower apparel prices in the market place – which drive down the value of marginal product – and a decline in labor skill needed to produce garments due to the presence of productivity enhancing capital. The close tracking of the garments services sector to the apparel industry indicates that it experiences many of these same pressures.

The pricing data used in this study comes from the Department Store Inventory Price index for apparel, the Producer Price Commodity Index for computers and the Consumer Price Index with the current base for finance and garment services. The apparel, computer and garment services prices revealed in this data are themselves indices that measures price changes of all the products within their respective production bundles. Coming up with the price of financial services is a difficult one due to the vast variety of potential financial instruments and the paucity of data. In place of such an index, this study uses the price of motor vehicle insurance as a proxy. Perhaps the most interesting trend to note about the data in is that the price of insurance is increasing at nearly an exponential rate while the price of apparel has only doubled over the past 50 years. One could see this as a difference in the transferability of goods – after all clothes are durable items that can be manufactured anywhere, but insurance is specific to a legal system.

Testing the hypotheses proposed by the three perspectives compels specification of a number of empirical models to be used later in estimation. At first blush, the use of multiple models seems to be only a reaction to significant Chow tests, but to examine the theoretical propositions offered by these two authors, multiple specifications are clearly necessary. To examine Wallerstein’s hypothesis that labor demand is shifting backward, one must compare the constant terms in estimated equations over two different time periods. Testing Rodrik’s elasticity change hypothesis requires a similar procedure. Examining the conclusions of development theory alone requires the estimation of at least four different regression models due to the variation in elasticity change that must be shown across the two economic sectors.

The specification of multiple empirical models is also necessary because of concerns about autocorrelation and unresolved questions about the endogeniety of wages and hours in a model of labor supply and demand. Part of this difficulty is due to the data that must be involved in the estimation procedure. Time series data is required for the examination of these propositions and observations in each time period are often correlated with each other. The debate over endogeniety is far more troubling, because of its effects on the consistency and efficiency of parameter estimations to test the above theories. There is precedent for assuming exogeneity of wages and hours in labor demand in Thomas Mroz’s study of female labor supply (Mroz, 1987). However, it also seems contrary to economic logic for the two to be exogenous.

Despite their different components, each of these empirical models share the same origin in the following maximization problem. This problem attempts to represent a profit maximization objective on the part of firms in a given industry. The firm is assumed to have a Cobb-Douglas production technology as well as costs that are a function of inputs and their per unit costs.

Labor demand can be derived from the function that comes from finding the first order conditions (2) of the firm’s profit maximization problem with respect to L(1).

This resulting expression asserts that wages are some function of good prices, and marginal product of labor.

Where labor supply is used in this treatment, its specification emerges from previous empirical studies. Typically, the labor supply in a given industry is thought to be a function of wages, non-work income and a class of demographic variables such as age, experience and years of education. To avoid under-identification in the simultaneous equation system, the specification of labor supply consists of two variables, a lagged wage variable and the labor force in the particular economic sector.

The Cobb-Douglas functional form is used in the empirical specification because of its attractiveness in estimating elasticities as well as its reflection of the relationship between wages, good price, marginal product and labor as multiplicatively interactive. Use of this functional form provides the deterministic model (3), where B1 is the real wage elasticity of labor demand, B2 is the price elasticity of labor demand, and B3 is the responsiveness to marginal productivity change:

Unfortunately, the availability of time series data constrains the quality of variables and forces a re-specification by proxy. Marginal product data is unavailable for manufacturing and five item total factor productivity (Bartelsman and Gray, 1996) is used in its place. Thus, what should have been (4):

became (5):

The empirical specification of labor supply in this model did not suffer from these same data problems. This fortunate turn of events results from its specification as a function of the current real wages, lagged real wages and the size of the labor force as in 7).

Research Conclusions

The empirical analysis began with specification of each sector in the log-log model form shown above. Labor demand equations were estimated for the pre-1973 era, the post-1973 era and the entire span of available data in the apparel and finance sectors. Only one equation was estimated for the computer and garment service sectors due to the relatively short span of official data collection for each industry.

Table One: OLS Results Using the Log-Log Specification

Type\Sector		Apparel	Computers	Finance	Garment Service
Pre-1973	F	17	-	951	-
	R^2	.73	-	.99	-
	B1	1.24	-	.69	-
	T	4	-	6.9	-
Post-1973	F	104	-	261	-
	R^2	.94	-	.96	-
	B1	.87	-	-.49	-
	T	1.97	-	3.5	-
Pooled	F	92.198	3.473	3501.51	35.5
	R^2	.969	.881	.997	.893
	B1	1.075	4.558	.545	.757
	T	8.006	3.217	13.73	1.748

Chow tests were only carried out on the two sectors with the longest data histories- apparel and finance. The two Chow tests carried out on this dataset indicate that the relationship between wages, productivity, price and work-hours have changed significantly in the past fifty years represented by the dataset. A structural change is especially apparent in the case of the apparel industry (F=72.8), which appears to have expanded in labor demand from 1947-1973 (a result that creates havoc with the estimation of wage elasticity in that period) and stabilized since that time. The finance sector does not show such a dramatic change (F=22.3).

OLS is clearly only a starting place for analysis of this topic. In this context, OLS is likely to be non-stationary and suffers from problems of endogeniety. Of the two problems, the latter is probably the worst problem as it causes OLS to generate parameter estimates that are both inconsistent and inefficient. Non-stationarity only causes the generation of inefficient parameter results. The OLS estimators for wage elasticity in Table One exhibit have both of these problems so they exhibit degrees of both of these properties. All of the time series variables used in this empirical analysis exhibit statistically significant unit roots.

The pervasive endogeniety problem is a bit more difficult to explain. An intuitive economic interpretation of these results would suggest that this positive slope is simply a reflection of the notion that labor is a normal good in the production process (i.e. as firm incomes increase, more workers are hired) and that the labor demand curve is just tracing out the labor supply curve in each of these sectors. Unfortunately, this explanation is probably not feasible, because of dual action on the labor supply side. Over the past fifty years, there have been dramatic increases in the labor force participation rate as middle and upper class women entered the workforce in increasing numbers – an event which should shift the supply curve outward, attenuating much of the wage change generated by increasing demand.

To clear up the non-stationarity problem, first differences were taken of all the time series under analysis and OLS was again applied after scaling the variables to control for heteroscedasticity. The differencing allowed efficient estimation of the wage parameter in both the apparel and garment service industries. Precise estimation of the wage parameter in the case of the garment services industry required the use of semi-annual data. The wage parameters are shown in the table below along with their GLS estimators to indicate the sensitivity of the results with respect to problems of heteroscedasticity and autocorrelation.

Table Two: OLS and GLS estimators using the first-difference models

	Apparel		Garment Service
Models	OLS	GLS	OLS	GLS
F	4.851	194.97	1.318	.96
R^2	.41	.89	.29	.04
B1	-4.2	-5.1	-5.9	-4.9
T	-3.096	-8.876	-1.536	-1.25

The two GLS estimators use different correction methods to ameliorate the autocorrelation problem that exists in each of the models. The apparel GLS estimator corrects for autocorrelation via the Prais-Whinston algorithm while the garment service GLS estimator uses a Cochrane-Orcutt search method at 500 iterations.

To find the stream of point elasticities for these two sectors, the wage parameter is multiplied by the ratio of wages to worker hours. This calculation is possible because the wage parameter of these difference models is essentially the same as the first derivative of the worker-hour function with respect to wages. It should be noted, however, that the resulting elasticity results will only supply point results.

The finance and computer sectors resisted attempts to estimate reasonable wage elasticities at every level. OLS, GLS and 2SLS all failed to produce negative values for a wage parameter that could be used to calculate a corresponding elasticity. All of these techniques are thwarted by essentially the same phenomena. Each of the sectors is experiencing a significant growth in demand and thus, is under-identified in each of these econometric specifications. This growth in demand is apparent when one examines either graph of work-hours plotted against hourly wages. There is simply not enough variation with which to calculate an accompanying wage elasticity in either case.

The most interesting observation that can be gained from this exploration is that the apparel and garment services sectors are undergoing changes in wage elasticity. As expected apparel is growing more wage elastic more quickly. However, the garment services sector is

exhibiting a decline in wage elasticity – potentially reflecting the effect of stagnant skill sets in workers who hold positions in this area.

Despite the attainment of precise parameter estimates in most of the estimated relationships, this examination of labor demand is still incomplete. In many of the models estimated with 2SLS, as with OLS and GLS, the wage elasticity was not estimated with the right sign. This fact implies that the underlying empirical model did not wash the endogeniety from the model. It also implies that control variables need to be added to account for other determinants of labor supply.

While this examination of labor demand has not provided conclusive evidence about which of these theories is best supported with the available data, it has shown two results that are theoretically useful. The useful results are that in at least two sectors, the wage elasticity of labor demand has increased over time. All of the results, while not providing a conclusive refutation of the perspective offered by Wallerstein, infirm his notions about declining labor demand. It supports the two alternative theses by demonstrating over time that there is an increasing wage elasticity of labor markets in the global lagging sectors over the past 50 years and a decreasing wage elasticity in the scope limited lagging sector over the period of record-keeping. In doing so, it also demonstrates that Rodrik’s notion of overall elasticity increase must be taken with caution.

Conclusions

These seemingly esoteric results have important implications for public administrators because they indicate how parts of the managerial context will evolve in the near future. Public administrators across positions are likely to encounter some effect from these changes, but public sector managers who deal specifically with personnel or finance issues are likely to feel these impacts of globalization most directly. Public personnel administrators face a special challenge from this trend, because they will be forced to restructure their image of public sector employment. In comparison, the public sector manager in finance will only have to deal with a potentially eroding tax base and the political difficulty of enacting changes in revenue extraction in the presence of hostile interest groups – challenges that many of these managers already face on a regular basis.

The lack of an apparent decline in U.S. labor demand across these economic sectors is a positive sign for the American public sector in general. A typical decline in labor demand would indicate macroeconomic trends that are unfavorable over time in the form of an increasing natural rate of unemployment and lower rates of economic growth. Such macroeconomic indicators indicate a recessionary trend in the overall economy - a time that is often associated with larger demand for services from the government in the form of benefit transfers. This larger demand for services often comes at a time of lower tax revenues, a dual relationship that forces many governments into untenable deficit spending positions that often endanger their long-term financial condition.

The increasing wage elasticity across economic sectors creates some difficulties for public personnel managers. For some time now, personnel administrators have faced high turnover rates in support staff positions in areas like information technology and finance, because the public sector is often incapable of matching the wages and benefit packages that private sector employers can offer to these trained and skilled individuals. The increasing wage elasticity simply implies that these workers will continue to be difficult to hold on to for any appreciable length of time. The change that these administrators will face under globalization is an increasing turnover rate among workers who were originally employees in what many people would describe as lagging sectors. These workers now have more options available to them at a given wage rate than they had previously. Continued economic expansion will only worsen this problem due to tight labor markets.

This increased turnover across the set of public sector positions should force the personnel administrator to alter the way employees are viewed. Instead of long term, life-long employees, public employment will be transitory for the majority of workers. To attract individuals to public sector positions, the personnel administrator and the local governments they represent will have to focus on providing more non-monetary benefits. Pervasive ongoing training programs should broaden employee skill sets and allow a few long term employees to take the place of multiple people in different, more specialized positions while a steady flow of employees enter and exit the public sector to provide a given service level in the public sector. The public sector should also take advantage of workers retraining and could provide tuition waivers instead of an additional benefit package or some percentage of wages.

Increasing wage elasticity also poses a difficulty for administrators in public finance. The enhanced sensitivity of markets to wages also indicates a heightened sensitivity to market distortions. In this environment, the income taxes used in some localities as a source of revenue increase the tax burden on employees across sectors. This burden immediately decreases area employment. More employees choose leisure over labor due to its greater expense and tend to rely more on gray market transactions, pensions and non-labor income to cover the costs of living. Increased reliance on these modes of income lowers tax revenue immediately. Over time, the tax base itself will erode, as workers with opportunities in other areas will move to other areas without income taxes.

To create an unemployment neutral system of revenue generation, the governments in these areas will have to alter their mix of taxes and fees. Additional income taxes should not be used to prop up sagging tax revenues in the short run because levying additional taxes only aggravates the problem. Income tax rates should decline, and disappear if possible to make the tax system truly unemployment neutral. Taxes on consumption items should increase to compensate for the loss of revenue to make the change revenue neutral as well.

The change to the tax system will clearly vary by the demographic make-up of the region that has levied the tax in the past. Areas with large immobile populations (i.e. groups aged 55+) could seek to move more of the tax burden onto those who are unlikely to move for wage reasons to protect their more mobile counterparts. Greater reliance on property taxes could achieve such an end, as could a number of other revenue generating schemes. Tax expenditures could give these mobile populations incentive to stay as well by lowering property tax rates on entry level houses.

Recommendations

Future research in this area should focus on improving the econometric analysis presented within this paper and validating the implications for public personnel and financial administration that emerge from this set of research conclusions. The inability of OLS, GLS and 2SLS to produce negative wage elasticities in the case of the computer industry indicates that new statistical formulation and analysis are needed for further insight. Ex post analysis of the match between predicted and real trends in the public sector must also take place to assure that a feedback loop is maintained between theory and reality so that understanding of this trend and its impact can improve over time.

The econometric work of future analysis needs to focus on resolving two issues – one of which has not been dealt with before in the context of this paper. The first issue needing resolution – at least in the case of the apparel industry - is the endogeniety problem that is apparent from the positive signs on the estimated wage elasticities. Resolving this issue involves at least three factors: attention to the loss in efficiency from either GLS or 2SLS, better measurement of marginal productivity in the industry and improved specification of the labor supply function that is needed for estimation via 2SLS.

The ex post analysis should take existing models of turnover and revenue estimation and look for structural changes in the validity of these models as time passes. If the research conclusions within this study hold, governments should observe more rapid turnover in employees who are less skilled. Additionally, revenues from income taxation should also decline as local labor markets become more sensitive to market distortions – causing higher steady state levels of unemployment.

References

Cabellero, R. and Jaffe, A. (1993) How high are the giants’ shoulders: an empirical assessment of knowledge spillovers and creative destruction in a model of economic growth. National Bureau of Economic Research Macroeconomic Annual 1993 p. 15-73.

Krasner, S. (1976). State power and the structure of international trade. In J. Frieden and D. Lake (Eds.) International Political Economy: Perspective on Global Power and Wealth (p. 19-36). New York: St. Martin’s Press.

Marquardt, M. J. (1998). The global advantage: how world class organizations improve performance through globalization Houston, TX: Gulf Publishing Company.

Martin, H.P, Schumann, H., and Camiller, P. (1997). The global trap: globalization and the assault on prosperity and democracy London: Zed Books.

Mroz, T. (1987). The sensitivity of an empirical model of married women’s hours of work to economic and statistical assumptions.

Econometrica, Vol.55, No.4.

Rodrik, D. (1997).

Has globalization gone too far? Washington, DC: Institute for International Economics.

Wallerstein, I. (1974). The rise and future demise of the world capitalist system. Comparative Studies in Society and History v.16, #4 p.387-415.

---

Endnotes

Note 1: Rwanda, Malawi and Botswana average 70-80% of their GDP coming from agricultural pursuits  Back.

Note 2: The US, UK, France and Japan all derive less than four percent of their gross domestic product from agriculture.  Back.