Do unions impact efficiency?: evidence from the U.S. manufacturing sector.I. INTRODUCTION
The relationship between labor unions and firm performance has garnered a vast amount of attention from scholars, unions, and businesses, as well as from policymakers as extensive studies have explored both the theoretical foundations and the empirical evidence regarding the impact of labor unions on productivity, as well as on other aspects of business, such as sales, profitability, investment, and employment growth (Doucouliagos and Laroche 2003). Perhaps one of the most relevant public policy questions is how unions affect productivity. Although studies have empirically investigated union productivity effects for a variety of industries and public sectors, particularly in the U.S. and European countries, empirical evidence on this question, however, is mixed.
Brown and Medoff (1978) and Freeman and Medoff (1984), in their early and influential work, argued that labor unions can raise productivity by inducing managers to adopt more efficient production methods and policies, and by providing better communication channels between workers and management. Supporting this impression, Dworkin and Ahlburg (1985) argued that opening communication channels between management and workers can result in integrative, rather than distributive dis·trib·u·tive
a. Of, relating to, or involving distribution.
b. Serving to distribute.
2. , bargaining. Allen (1984) used the example of the construction industry, and cited union hiring halls, apprenticeship programs, and managerial shock effects as the reasons for union productivity improvements. In recent years, many employers have adopted new ways of organizing work that emphasize employee involvement. Empirical evidence suggests that various human resources The fancy word for "people." The human resources department within an organization, years ago known as the "personnel department," manages the administrative aspects of the employees. management practices have a positive impact on firms' productivity. Black and Lynch (2001) found that unionized plants with these new approaches to organizing work had higher productivity than other similar non-union plants. Recent studies indicate that improved productivity of firms with labor unions may reflect better human resources management practices (Bloom and Van Reenen 2007; Doucouliagos and Laroche 2006; Ichniowski and Shaw 2003; Machin and Wood 2005).
However, Lewis (1963) pointed out that one of the most well-established impacts of labor unions on firms is the ability to raise wages above competitive levels. Such a monopoly power may exert adverse impacts on firm productivity by distorting the labor market labor market A place where labor is exchanged for wages; an LM is defined by geography, education and technical expertise, occupation, licensure or certification requirements, and job experience . Other unfavorable effects of unions are on R&D spending and investments, since by acting as a tax on the return on investment, unions limit innovative and investment activities (Connolly, Hirsch, and Hirschey 1986; Hirsch and Link 1987; Link and Siegel 2002; Lommerud, Meland, and Straume 2006; Menezes-Filho and Van Reenen 2003). These effects can have a detrimental impact on the dynamic productivity of firms. Thus, unions can both enhance and detract from detract from
verb 1. lessen, reduce, diminish, lower, take away from, derogate, devaluate << OPPOSITE enhance
verb 2. firms' productivity performance. Because of the ambiguity over the net effect of unions, an assessment of the union's effects on economic performance hinges on empirical evidence (Doucouliagos and Laroche 2003). (1)
Although most literature focuses on the performance or productivity impacts of unions, a strand of literature also looks at the channels through which unions affect measured performance/productivity. Measured productivity can be decomposed de·com·pose
v. de·com·posed, de·com·pos·ing, de·com·pos·es
1. To separate into components or basic elements.
2. To cause to rot.
1. into two important components: the adopted production technologies and the efficiency of utilizing the technology. The former represents a shift in the production function and the latter shows the distance between actual production level and frontier production level for given inputs. An increase in the measured productivity may result if a firm adopts a better production technology or improves "technical efficiency" by moving production toward the frontier level of a given technology.
The arguments for or against unions regarding productivity can trace their causes to the impacts on one or both of the productivity components. For instance, the communication effect of Brown and Medoff (1978) and Freeman and Medoff (1984) results in improved efficiency in using installed technology, whereas the outcome of changing production methods, also mentioned by the same authors, amounts to choosing new production technologies. Freeman and Medoff (1984) also suggested that "union shocked" management is able to extract more output from a given amount of inputs than management, which may be because of improved information flow among workers or between workers and management (Carmichael and MacLeod 1993). This channel of effect also amounts to improved technical efficiency.
For theoretical arguments, Kuhn (1985) provided a model that explains unions' effect on firm productivity, utilizing an agency cost model to show that unions may improve firm productivity by reducing its monitoring costs. Based on the agency cost theory of Jensen and Meckling (1976), Kuhn (1985) assumed that firm managers own a of the firm, whereas outside shareholders own (1- [alpha]) of the firm. Because a < 1, the effort level of managers would be less than the optimal level required to maximize the total value of the firm. The agency problem is partially mitigated if managers' efforts can be monitored by outside shareholders. However, because shareholders may be numerous and not well organized, the cost of monitoring could be high. However, organizing the union makes monitoring much easier with lower cost. The union stewards and representatives are always in the plant and can easily observe managers' performance, enhancing managers' efforts and resulting in improved technology efficiency. Better monitoring would most likely lead to a more efficient use of current technology, that is, improving technical efficiency.
A strand of empirical research Noun 1. empirical research - an empirical search for knowledge
inquiry, research, enquiry - a search for knowledge; "their pottery deserves more research than it has received" employs the data envelopment analysis This article or section may be confusing or unclear for some readers.
Please or discuss this issue on the talk page. (DEA) approach, which is a mathematical programing method. Byrnes et al. (1988) used the DEA approach to estimate the effect of unions on technical efficiency using data from the U.S. mine establishments in the 1970s. They did not find evidence that unions harmed technical efficiency. Doucouliagos and Laroche (2006) also used the DEA approach to investigate the effects of human resource practices and unionization on technical efficiency of French industries. They found that while the human resource practice improves efficiency, existing unions moderated the effect. A different strand of literature uses the econometrics econometrics, technique of economic analysis that expresses economic theory in terms of mathematical relationships and then tests it empirically through statistical research. approach of stochastic By guesswork; by chance; using or containing random values.
stochastic - probabilistic frontier models. Cavalluzzo and Baldwin (1993) compared the productivity differential between union and non-union contractors in the United States United States, officially United States of America, republic (2005 est. pop. 295,734,000), 3,539,227 sq mi (9,166,598 sq km), North America. The United States is the world's third largest country in population and the fourth largest country in area. and found a substantial union productivity premium. Dickerson et al. (1997) used the model proposed by Cornwell et al. (1990) to examine the effects of union strikes on productivity and efficiency in a sample of UK manufacturing industries manufacturing industries npl → industrias fpl manufactureras
manufacturing industries npl → industries fpl de transformation
in the 1970s. They found that strikes only had a negligible net impact on output in this period, and the presence of unions mitigated the effect.
This article estimates the impact of labor unions on the U.S. manufacturing sector using state-level panel data on 48 states from 1983 to 1996, focusing on technical efficiency effect of the union. The main motivation of this article stems from observing that unionization rates in the U.S. manufacturing sector at both the national and state levels have been continuously declining for the last few decades. Understanding how much these changes over time have affected efficiency or productivity of the U.S. manufacturing sector would be useful for scholars, unions, and businesses, as well as policymakers. Given this objective, using the state-level manufacturing sector as a unit of analysis would be appropriate as data on dependent and independent variables In mathematics, an independent variable is any of the arguments, i.e. "inputs", to a function. These are contrasted with the dependent variable, which is the value, i.e. the "output", of the function. are not only available, but also provide a rich panel data set that consists of 48 states and over 14 years (672 observations). (2)
The current study adopts the econometric e·con·o·met·rics
n. (used with a sing. verb)
Application of mathematical and statistical techniques to economics in the study of problems, the analysis of data, and the development and testing of theories and models. approach (the stochastic frontier model) in this research. Stochastic production frontiers are preferred for a number of reasons. Although the DEA approach has the advantage of not requiring the specification of any particular functional form of the underlying production process, using deterministic 1. (probability) deterministic - Describes a system whose time evolution can be predicted exactly.
2. (algorithm) deterministic - Describes an algorithm in which the correct next step depends only on the current state. frontiers is based on the unrealistic assumption that total error is because of inefficiency and that there are no random errors that occur in the production process. (3) Another benefit of stochastic frontiers is that they allow inefficiency to exist in all states. In contrast, DEA identifies the frontier by assigning some of the states as operating under best practice (100% efficient) and defining the frontier in relation to these firms. With a stochastic production frontier, every state can be inefficient, although some will be more inefficient than others. Compared to the DEA method, the econometrics approach of the stochastic frontier model is more structural in specifying the relationships between variables and easily accommodates random noises in the data.
This article adds to the union-efficiency literature in several ways. First, this article employs a Tran slog stochastic production specification which is more flexible than the Cobb-Douglas specification adopted by most existing studies (Cavalluzzo and Baldwin 1993; Dickerson et al. 1997). Second, the cross-sectional nature of the data typically limits studies in the union-efficiency literature at the firm level, and the labor union labor union: see union, labor. measure is commonly defined with a dummy variable This article is not about "dummy variables" as that term is usually understood in mathematics. See free variables and bound variables.
In regression analysis, a dummy variable with a value of 1 if a union is present and 0 otherwise, because the preferred union density figures are hardly available. Not only does state-level manufacturing data enable us to cover an extended period, from 1983 to 1996, but the availability of union density also allows us to measure the effect of the continuous decline in union density in U.S. manufacturing on technical efficiency and productivity. Third, by employing the "true-fixed effect" panel stochastic frontier model of Greene (2005), econometric analysis presented in the article captures the effects of unionization rate on technical efficiency on U.S. states' manufacturing, while unmeasured attributes that are fixed over time are controlled for estimating the effect of labor unionization. To our knowledge, this is the first study that applies the true-fixed effect model to estimate union effects.
The remainder of this article is organized as follows. Section 2 discusses the empirical model and data, Section 3 presents the results, and Section 4 concludes.
II. EMPIRICAL MODEL AND DATA
The stochastic frontier model for cross-state panel data is specified as follows:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII ASCII or American Standard Code for Information Interchange, a set of codes used to represent letters, numbers, a few symbols, and control characters. Originally designed for teletype operations, it has found wide application in computers. ]
where y (state manufacturing output) is subscripted with i (state) and t (year), [[mu].sub.i] are state-specific fixed effects and m = 1, ... M indexes inputs that affect the frontier (maximal max·i·mal
1. Of, relating to, or consisting of a maximum.
2. Being the greatest or highest possible. ) level of output. The input variables (xs) consist of state public capital stock (g), state private manufacturing capital stock (k), manufacturing labor (l.), and unionization rate (union). (4) Non-neutral technical change is assumed in Equation (1) by introducing t (time trend), [t.sup.2] (time trend squared), and cross-products of t and inputs to capture technical change effect. (5) [u.sub.it] is an error term which is independent and identically distributed [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] and independent of [u.sub.it] which are non-negative random variables that account for technical inefficiencies in production. [u.sub.it] are assumed to be independently distributed as truncations at zero of the [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]distribution where [S.sub.it]= [Z.sub.it] [delta], where zit zit
A pimple. is a vector of variables that affects the deviation of output from the frontier causing technical inefficiency and [delta] is a vector of parameters to be estimated.
The inefficiency is determined by U.S. states' private manufacturing unionization rates, unemployment rates, wage index, U.S. unemployment rates, and time trend. The primary variable of interest is U.S. states' private manufacturing unionization rate. To disentangle the union effect from other effects, several variables are included in the inefficiency function. U.S. states' unemployment rate and U.S.'s unemployment rate are included to account for cyclical cyclical
Of or relating to a variable, such as housing starts, car sales, or the price of a certain stock, that is subject to regular or irregular up-and-down movements. changes in unutilized capital and hoarded labor that are likely to hinder a state from reaching the production frontier (Munnell 1990). We also controlled for worker quality by including the wage index in the equation. The quality of union and non-union workers may be different because unions encourage their members to stay longer in the firm and thus their skill levels may be higher because of average longer job tenure. Given the fact that workers are generally paid according to according to
1. As stated or indicated by; on the authority of: according to historians.
2. In keeping with: according to instructions.
3. their performance, the wage index obtained by dividing the gross earnings of all employees on the payroll of operating manufacturing establishments by the number of employees is used as a proxy for labor quality. (6) Finally, this study includes time trend, commonly employed in the frontier production literature, to capture other time-related factors. Following Battese and Corra (1977), we replace [[[sigma].sup.2].sub.v] and [[[sigma].sup.2].sub.u] and [[sigma].sup.2] = [[[sigma].sup.2].sub.v] + [[[sigma].sup.2].sub.u] and [gamma] = [[[sigma].sup.2].sub.u]/([[[sigma].sup.2].sub.v]+[[[sigma].sup.2].sub.u]). The maximum-likelihood estimator (MLE MLE Maximum Likelihood Estimation
MLE Managed Learning Environment
MLE Maximum Likelihood Estimate
MLE Medical Laboratory Evaluation (Medical Laboratory Proficiency Testing Program, Washington, DC) ) (7) estimates the model parameters ([theta Theta
A measure of the rate of decline in the value of an option due to the passage of time. Theta can also be referred to as the time decay on the value of an option. If everything is held constant, then the option will lose value as time moves closer to the maturity of the option. ], [delta], [[sigma].sup.2], and [gamma]).
The above-specified model allows the production function (frontier) to have state-specific fixed effects ([[mu].sub.i]). That is, instead of assuming a single U.S. frontier function for all the states, we allow heterogeneity het·er·o·ge·ne·i·ty
The quality or state of being heterogeneous.
the state of being heterogeneous. via state-specific intercepts in the frontier production function (the "true-fixed effect" panel stochastic frontier model by Greene 2005). This effectively eliminates biases in the parameter estimates that arise from omitting unobserved time-invariant variables correlated with included explanatory variables. As emphasized in theoretical and empirical studies Empirical studies in social sciences are when the research ends are based on evidence and not just theory. This is done to comply with the scientific method that asserts the objective discovery of knowledge based on verifiable facts of evidence. by Kumbhakar (1991) and Greene (2005) respectively, failure to include individual-specific effects in a panel stochastic model is likely to bias the estimate of the one-sided error, [u.sub.it], which is one of the most important estimation elements. In a recent study, Kumbhakar and Wang (2005) provided evidence of heterogeneity in the production functions, for both the translog and Cobb-Douglas specifications, across countries.
The data covered 48 contiguous states observed for the period 1983-1996. (8) The primary independent variable of interest was the U.S. private manufacturing unionization rates, defined as the percent of employed workers covered by a collective bargaining agreement The contractual agreement between an employer and a Labor Union that governs wages, hours, and working conditions for employees and which can be enforced against both the employer and the union for failure to comply with its terms. obtained from unionstats.com. (9) The state-level data on manufacturing output (million dollars), labor quantities, wage index, state-level manufacturing private capital stock, and public capital stock were obtained from Cohen cohen
(Hebrew: “priest”) Jewish priest descended from Zadok (a descendant of Aaron), priest at the First Temple of Jerusalem. The biblical priesthood was hereditary and male. and Morrison Paul (2004). (10) Manufacturing output was deflated de·flate
v. de·flat·ed, de·flat·ing, de·flates
a. To release contained air or gas from.
b. To collapse by releasing contained air or gas.
2. by manufacturing the Gross State Product deflators for each state. The perpetual inventory Perpetual Inventory
An accounting method of maintaining up-to-date property records that accurately reflect the level of goods on hand.
The current balance of inventory is sustained daily by the addition of inventory to the account when goods are received and the method was applied to the data on state-level new capital expenditures (Annual Survey of Manufactures) as well as to the state-level public infrastructure investment data to obtain private manufacturing capital stock and public capital stock estimates. (11) U.S. and U.S states' unemployment rates were obtained through the U.S. Bureau of Labor Statistics Bureau of Labor Statistics (BLS)
A research agency of the U.S. Department of Labor; it compiles statistics on hours of work, average hourly earnings, employment and unemployment, consumer prices and many other variables. . Table 1 provides summary statistics.
TABLE 1 Summary Statistics Variables Mean Standard Min Max Deviation Panel (a) Manufacturing 62,998 63.858 2.229 368,329 outputs Private 19.894 20.758 645 105.892 manufacturing capital stock Public capital stock 15.495 13.410 2.108 78.105 Manufacturing 13,488 14,862 186 79,339 labors Private 21.34 10.02 2.70 50.00 manufacturing unionization rate (%) States' unemployment rate (%) 6.25 2.00 2.30 17.40 U.S. unemployment rate (%) 6.59 1.14 5.30 9.60 Wage index 0.81 0.12 0.54 1 Year Observation State-Averaged Private Manufacturing Unionization Rate (%) Panel (b) 1983 48 27.19 1984 48 25.88 1985 48 24.99 1986 48 23.86 1987 48 22.22 1988 48 21.75 1989 48 20.91 1990 48 20.60 1991 48 20.56 1992 48 19.49 1993 48 18.81 1994 48 18.30 1995 48 17.18 1996 48 16.96 Sources: (a) Cohen and Morrison Paul (2004), U.S. Bureau of Labor Statistics, Statistical Abstract of the United States (various issues) and www.unionstats.com and (b) www.unionstats.com.
We estimated two versions of Equation (1), and Tables 2 and 3 report the results. Model 1 in Table 2 assumes no state-specific fixed effects (i.e., [[mu].sub.i) = 0, [for all] i) and Model 2 in Table 3 allows for the full set of fixed effects. Before discussing the individual coefficients, we note that the absence of the inefficiency effects of the generalized likelihood-ratio tests of the null hypothesis null hypothesis,
n theoretical assumption that a given therapy will have results not statistically different from another treatment.
n are strongly rejected in both models. (12) Because each individual coefficient of the inputs (and interaction terms of inputs) of a translog model does not convey interpretable information, we calculated output elasticities for the inputs and the results are reported in Table 4.
TABLE 2 The Translog Model without State-specific Fixed Effects (Model 1) Frontier Function Coefficient Standard Inefficiency Error Function [[theta].sub.0] -7.269 *** 0.921 [[delta].sub.0] [[theta].sub.g] 1.603 *** 0.355 [[delta].sub.w] [[theta].sub.k] -1.157 *** 0.351 [[delta].sub.t] [[theta].sub.l] 2.053 *** 0.332 [[delta].sub.un] [[theta].sub.kk] 0.071 0.126 [[delta].sub.un2] [[theta].sub.ll] -0.014 0.039 [[delta].sub.union] [[theta].sub.gg] (-0.051) 0.110 [[theta].sub.kl] 0.040 0.108 [[theta].sub.kg] 0.227 *** 0.080 [[theta].sub.gl] -0.347 *** 0.063 [[theta].sub.t] 0.097 *** 0.028 [[theta].sub.tt] 0.004 0.001 [[theta].sub.kt] -0.028 0.008 [[theta].sub.gt] -0.011 0.005 [[theta].sub.lt] 0.029 0.007 [[theta].sub.union] 0.947 *** 0.213 [[theta].sub.unionunion] 0.140 *** 0.052 [[theta].sub. unionk] -0.446 *** 0.067 [[theta].sub. uniong] 0.033 0.036 [[theta].sub. unionl] 0.383 *** 0.061 [[theta].sub. uniont] -0.006 0.005 Frontier Function Coefficient Coefficient Marginal Efficient [[theta].sub.0] 4.283 4.283 *** [[theta].sub.g] -8.996 -8.996 (-1.101) *** [[theta].sub.k] 0.254 0.254 (0.031) *** [[theta].sub.l] 0.190 0.190 (0.023) *** [[theta].sub.kk] -0.039 -0.039 (-0.005) [[theta].sub.ll] -0.029 -0.029 (-0.004) *** [[theta].sub.gg] [[theta].sub.kl] [[theta].sub.kg] [[theta].sub.gl] [[theta].sub.t] [[theta].sub.tt] [[theta].sub.kt] [[theta].sub.gt] [[theta].sub.lt] [[theta].sub.union] [[theta].sub.unionunion] [[theta].sub. unionk] [[theta].sub. uniong] [[theta].sub. unionl] [[theta].sub. uniont] Frontier Function Standard Error [[theta].sub.0] 1.198 [[theta].sub.g] 2.459 [[theta].sub.k] 0.070 [[theta].sub.l] 0.044 [[theta].sub.kk] 0.028 [[theta].sub.ll] 0.007 [[theta].sub.gg] [[theta].sub.kl] [[theta].sub.kg] [[theta].sub.gl] [[theta].sub.t] [[theta].sub.tt] [[theta].sub.kt] [[theta].sub.gt] [[theta].sub.lt] [[theta].sub.union] [[theta].sub.unionunion] [[theta].sub. unionk] [[theta].sub. uniong] [[theta].sub. unionl] [[theta].sub. uniont] Other Parameters coefficient Standard Error [[sigma].sup.2] 0.094* 0.015 [gamma] 0.0823* 0.027 Model Statistics Log-likelihood 300.58 LR test of one-sided error 134.265 Notes: Subscripts of the symbols indicate the variable or variables, where g is the public capital, k the private capital, the labor. the time trend, union the union rate, iv the wage index, un the state unemployment rate, and un2 the U.S. unemployment rate. * 10% significance level, ** 5% significance level, and *** 1% significance level. TABLE 3 The Translog Model with State-specific Fixed Effects (Model 2) Frontier Function Coefficient SE Inefficiency Function [[theta].sub.0] 20.128 *** 3.667 [[delta].sub.0] [[theta].sub.g] -5.501 *** 0.804 [[delta].sub.0] [[theta].sub.k] 2.378 *** 0.755 [[delta].sub.0] [[theta].sub.l] 0.097 0.633 [[delta].sub.0] [[theta].sub.kk] -0.004 0.217 [[delta].sub.w] [[theta].sub.ll] 0.013 0.126 [[delta].sub.t] [[theta].sub.gg] 0.807 *** 0.152 [[delta].sub.un] [[theta].sub.kl] 0.036 0.163 [[delta].sub.un2] [[theta].sub.kg] -0.216 * 0.126 [[delta].sub.union] [[theta].sub.gl] -0.018 0.108 [[theta].sub.t] 0.139 *** 0.029 [[theta].sub.tt] 0.004 *** 0.001 [[theta].sub.kt] -0.011 0.005 [[theta].sub.gt] -0.025 *** 0.005 [[theta].sub.lt] 0.025 *** 0.004 [[theta].sub.union] 0.048 0.028 [[theta].sub. unionunion] 0.143 *** 0.046 [[theta].sub. unionk] 0.146 ** 0.064 [[theta].sub. uniong] 0.064 0.05 [[theta].sub.l] 0.125 ** 0.054 [[theta].sub.t] -0.01 *** 0.003 Frontier Function Coefficient (Marginal SE Efficient) [[theta].sub.0] 2.694 *** 0.301 [[theta].sub.g] -7.257 (-0.533) 0.456 [[theta].sub.k] 0.212 (0.016) *** 0.017 [[theta].sub.l] 0.162 (0.012) *** 0.011 [[theta].sub.kk] 0.022 (0.002) 0.019 [[theta].sub.ll] -0.024 (-0.002) *** 0.004 [[theta].sub.gg] [[theta].sub.kl] [[theta].sub.kg] [[theta].sub.gl] [[theta].sub.t] [[theta].sub.tt] [[theta].sub.kt] [[theta].sub.gt] [[theta].sub.lt] [[theta].sub.union] [[theta].sub. unionunion] [[theta].sub. unionk] [[theta].sub. uniong] [[theta].sub.l] [[theta].sub.t] Other Parameters cofficient Standart Error [[sigma].sup.2] 0.062 *** 0.009 [gamma] 0.937 *** 0.007 Model Statistics Log-likelihood 838.01 LR test of one-sided error 309.882 Notes: Slate-specific fixed effects are not reported. Subscripts of the symbols indicate the variable or variables, where g is the public capital, k the private capital, / the labor, / the time trend, union the union rate, w the wage index, un the state unemployment rate, and un2 the U.S. unemployment rate. * 10% significance level, ** 5% significance level, and *** 1% significance level. TABLE 4 Output Elasticities and Returns to Scale Model 1 Model 2 Coefficient SE Coefficient SE Output Output elasticities elasticities G 0.193 *** 0.024 G -0.142 0.099 K 0.464 *** 0.034 K 0.128 ** 0.052 L 0.402 *** 0.033 L 0.938 *** 0.050 union 0.032 0.029 union -0.099 *** 0.024 T 0.010 *** 0.003 t 0.018 *** 0.003 Returns to 1.058 *** 0.013 Returns to 0.924 *** 0.113 scale scale (g, k l) (g, k, l) Notes: Coefficients are evaluated at the mean, and the associated standard errors are computed using the Delta method. Subscripts of the symbols indicate the variable or variables, where g is (he public capital, k (he private capital. / the labor. / the time trend, and union the union rate. * 10% significance level, ** 5% significance level, and *** 1% significance level.
Overall, the estimates of the frontier production functions are different between models with and without state-specific fixed effects. Take, for example, the union effect, which is the main focus of our study. Without the fixed effects, Model 1 shows an insignificant output elasticity In economics, output elasticity is the percentage change of output (GDP or revenue for a single firm) divided by the percentage change of an input.
It is calculated as marginal product of an input to its average product. It is a local measure, defined at a point. , but the elasticity of union becomes significantly negative when the fixed-effect specification is used in Model 2. The negative impact of the union on the production frontier indicates that the union may hinder the adoption or progress of production technology, which may occur if the union limits the firms' ability to invest in R&D and other types of innovative activities (Connolly, Hirsch, and Hirschey 1986; Hirsch and Link 1987; Link and Siegel 2002; Menezes-Filho and Van Reenen 2003). Our union model exerts its influence on productivity through both the technology channel (shift of the frontier function) and the efficiency channel (distance between the actual production and frontier production). This article discusses the effects from the second channel in a later section.
The output elasticities of other production inputs also show noticeable differences between Models 1 and 2. The elasticities of both public and private capitals decrease when the fixed effects are included, whereas the labor elasticity increases from 0.402 (Model 1) to 0.938 (Model 2) with fixed effects. Technology progress (effect of t) also increases horn 1% to about 2% per year after controlling for fixed effects, and the later figure is consistent with the literature's findings of U.S. industries (Kumb-hakar and Wang 2006). The result sheds light on the impacts of individual heterogeneity on estimates.
Because Model 1 nests within Model 2, we conducted the likelihood-ratio (LR) test on the joint significance of state-specific fixed effects. The LR statistics are 1074.86. Given the degree of freedom equal to 47, the test overwhelmingly rejects the hypothesis that fixed effects are jointly insignificant. Therefore, the data provide evidence of heterogeneity captured via the state-specific fixed dummies in the production functions of the U.S. manufacturing sector. The results reinforce the finding of Kumbhakar and Wang (2005).
In terms of technical inefficiency determinants, Tables 2 and 3 report the estimated effects. Because the exogenous Exogenous
Describes facts outside the control of the firm. Converse of endogenous. determinants of inefficiency enter the inefficiency function in a nonlinear A system in which the output is not a uniform relationship to the input.
nonlinear - (Scientific computation) A property of a system whose output is not proportional to its input. way, we reported not only the coefficients of determinants, but also their marginal effects defined as [alpha] E ([alpha] / [Z.sub.it)). The marginal effects were calculated using the formula given in Wang (2002). Qualitatively, the marginal effects of the various exogenous determinants were similar between the two models. In particular, findings revealed that labor unions reduce technical inefficiency (increase technical efficiency). The marginal effects are -0.004 in Model 1 and -0.002 in Model 2, and both are statistically significant. A figure of -0.004 (-0.002) implies that the output can be improved by 0.4% (0.2%) if the existing technology and resources are used more efficiently. The finding that unions improve technical efficiency is consistent with the communication effect of Brown and Medoff (1978), the union shocked management hypothesis of Freedman and Medoff (1984) and Carmichael and MacLeod (1993), and the agency cost story of Kuhn (1985).
Findings for other variables show that although the U.S. unemployment rate does not have significant effects on technical inefficiency, the state-level unemployment rate is detrimental to manufacturing sectors' technical efficiency (increases inefficiency). A plausible explanation is that the unemployment rate positively correlates with firms' excess capacity (because of fixed capitals and labor hoarding). Therefore, production potentials are not fully realized when the unemployment rate is up, resulting in increased technical inefficiency. Finally, as expected, wage increases, which is a proxy for labor quality, have significantly positive effects in improving technical efficiency (reducing inefficiency).
Table 5 reports the efficiency index defined as E (exp exp
2. exponential (-[u.sub.it]) | [u.sub.it] - [u.sub.it]), which measures the percentage by which the potential production is achieved. The results show that the sample mean of the index is about 91% (Model 1) to 94% (Model 2), indicating fairly high efficiency levels. The efficiency index of Model 1 is expectedly lower than that of Model 2, because Model 1 may incorrectly capture some of the time-invariant and state-specific effects as inefficiency.
TABLE 5 Efficiency Indexes Model Mean 25% 50% 75% 1 0.908 0.892 0.930 0.951 2 0.940 0.928 0.958 0.973 Note: The efficiency index is E (exp (-[u.sub.it] | [u.sub.it] - [u.sub.it]) based on Battese and Coelli (1988).
As discussed earlier, unions may affect productivity through both the technology frontier and the technical inefficiency channel. Here we calculate the union's overall effect on productivity using the results of Model 2. The effect is calculated by
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
Because we used the log of unionization rate (In [union.sub.it)) in the translog specification of the frontier equation (i.e., the f (.) function) and the rate level in the inefficiency function (i.e., the u (.) function), the first term on the right-hand-side of the equation may be transformed using the following approximation equation:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
As shown, union elasticity on the technology frontier is--0.099 and union marginal effect on technical inefficiency is--0.002. Given that the sample mean of the unionization rate is 21.34, the overall effect measured at the mean is equal to -0.0026 (=-0.099/21.34 + 0.002). The technology channel dominates the inefficiency channel. Interestingly, the conclusion would have reversed if we based our analysis on Model 1, which ignores time-invariant and state-specific fixed effects.
The unionization rate declined in the sample period from 1983 to 1996. Given the findings that unionization increases technical efficiency but reduces overall productivity, it would be interesting to see the impact on production efficiency and productivity had the unionization rate remained at the higher level of 1983. To this end, we conducted two sets of counterfactual simulations. First, we used the estimated parameters of Model 2 to recalculate re·cal·cu·late
tr.v. re·cal·cu·lat·ed, re·cal·cu·lat·ing, re·cal·cu·lates
To calculate again, especially in order to eliminate errors or to incorporate additional factors or data. the efficiency index for the year 1996 assuming the unionization rate in each state was at the 1983 level. The result yields a mean efficiency index of 0.975. If the actual data is used, the mean efficiency index in 1996 is 0.951. Therefore, declined unionization rate impairs the technical efficiency by about 2.4 percentage points.
Second, we used the results of Model 2 to compute the (conditional) expected level of log-output with and without the unionization rate fixed at the 1983 level.
The expected level of output is defined as follows:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
where E ([v.sub.it] | In x, z) is zero and the formula of E ([u.sub.it] | In x, z) is given in Wang (2002). Results show that the expected log-output in 1996 is 10.704 using the actual data, and the figure would reduce to 10.630 had the unionization rate been fixed at the 1983 level. This indicates an industry gain of 7.4% of output owing to owing to
Because of; on account of: I couldn't attend, owing to illness.
owing to prep → debido a, por causa de the declining unionization rate.
This section shows that the unionization helps improve the technical efficiency of firms but hampers their technology adoption. The results provide important policy implications. As the long-term trend of the declining unionization rate continues in many of the U.S. states, efforts should focus on implementing efficiency-improving measures where there are less union stewards and representatives in the plants to monitor managements and help communications between workers and managers. One of the possible measures is to improve the quality of workers, which is shown to be positively correlated with technical efficiency, through better human resource management practices such as training, incentive compensation plans, employee monitoring, profit sharing profit sharing, arrangement by which employees receive, in addition to their wages, a share of the net profits of a business. The purpose is to give them an incentive to increase their output through enhanced morale, less wasteful use of materials, better care of , and information sharing See data conferencing. . However, for a few states which show signs of increasing unionization rates in the recent years (e.g., California and Colorado), the policy should be designed to mitigate the union's negative impacts on firms' technology adoptions. Lowering the cost of R&D and providing more incentives for adopting new technologies are among the possible actions.
This article investigates channels through which unions affect measured performance/productivity. Utilizing state-level panel data on 48 states from 1983 to 1996, we estimate the impact of labor unions on the U.S. manufacturing sector, focusing on the technical efficiency effect of the union. Employing the "true-fixed effect" panel stochastic frontier model, the results indicate that while labor unions reduce firms' technical progress, they improve the technical efficiency by which firms utilize existing technology. The fall of the unionization rate in the 1983-1996 period thus impaired firms' technical efficiency by 2.4 percentage points. However, because the effect on technical progress dominates the effect on technical efficiency, the overall result of the declining unionization increases the manufacturing output by 7.4% in the sample period.
Although the use of state-level manufacturing sector as the unit of analysis enables us to investigate the impact of labor unions on technical efficiency of the U.S. manufacturing sector, the data does not allow for the investigation of union effects at the less aggregate levels such as industry level. As noted by Kuhn (1998), the union effects may be different among industries within the manufacturing sector because of different management approaches, production methods, and other industry-specific characteristics. In addition, most decisions affecting technical efficiency are made by the firm and, therefore, are likely affected more by union activities at the firm level than by those at the aggregate level. Thus future studies at the industry and firm levels would be nice complements to the current study.
Allen, S. G. "Unionized Construction Workers Are More Productive." The Quarterly Journal of Economics The Quarterly Journal of Economics, or QJE, is an economics journal published by the Massachusetts Institute of Technology and edited at Harvard University's Department of Economics. Its current editors are Robert J. Barro, Edward L. Glaeser and Lawrence F. Katz. , 99, 1984, 251-74.
Battese, G. E., and T. J. Coelli. "Prediction of Firm-Level Technical Efficiencies with a Generalized Frontier Production Function and Panel Data." Journal of Econometrics, 38, 1988, 387-99.
--. "A Stochastic Frontier Production Incorporating a Model for Technical Inefficiency Effects." Working Papers working papers
Legal documents certifying the right to employment of a minor or alien.
Noun 1. working papers in Applied Statistics, No. 65, Department of Economics, University of New England The University of New England can refer to:
Noun 1. NSW - the agency that provides units to conduct unconventional and counter-guerilla warfare
Naval Special Warfare , Australia, 1993.
Battese, G. E., and G. S. Corra. "Estimation of a Production Frontier with Application to the Pastoral Zone of Eastern Australia." Australian Journal of Agricultural Economics Agricultural economics originally applied the principles of economics to the production of crops and livestock - a discipline known as agronomics. Agronomics was a branch of economics that specifically dealt with land usage. , 21, 1977, 169-79.
Black, S., and L. Lynch. "How to Compete: The Impact of Workplace Practices and Information Technology on Productivity." The Review of Economics and Statistics, 83, 2001, 434-45.
Bloom, N., and J. Van Reenen. "Measuring and Explaining Management Practices Across Firms and Countries." Quarterly Journal of Economics, 122(4), 2007, 1351-408.
Brown, C., and J. L. Medoff. "Trade Unions in the Production Process." Journal of Political Economy, 86(3), 1978, 355-78.
Byrnes, P., R. Fare, S. Grosskopf, and C. A. Lovell. "The Effect of Unions on Productivity: U.S. Surface Mining of Coal" Management Science, 34, 1988, 1037-53.
Carmichael, L., and B. MacLeod. "Multiskilling, Technical Change and the Japanese Firm." Economic Journal, 103, 1993, 142-60.
Cavalluzzo, L., and D. Baldwin. "Unionization and Productive Efficiency," in The Measurement of Productive Efficiency, edited by Fried Harold, C. A. Knox Lovell, and S. Schmidt. New York New York, state, United States
New York, Middle Atlantic state of the United States. It is bordered by Vermont, Massachusetts, Connecticut, and the Atlantic Ocean (E), New Jersey and Pennsylvania (S), Lakes Erie and Ontario and the Canadian province of : Oxford University Press, 1993, 210-20.
Cohen, J. P., and C. Morrison Paul. "Public Infrastructure Investment, Inter-state Spatial Spillovers, and Manufacturing Costs." The Review of Economics and Statistics, 86, 2004, 551-60.
Connolly, R. A., B. T. Hirsch, and M. Hirschey. "Union Rent Seeking In economics, rent seeking occurs when an individual, organization, or firm seeks to make money by manipulating the economic and/or legal environment rather than by making a profit through trade and production of wealth. , Intangible Capital and Market Value of the Firm." The Review of Economics and Statistics, 68, 1986, 567-77.
Dickerson, A. P., P. A. Geroski, and K. G. Knight. "Productivity. Efficiency and Strike Activity." International Review of Applied Economics, 11, 1997, 119-34.
Doucouliagos, C., and P. Laroche. "What Do Unions Do to Productivity? A Meta-Analysis." Industrial Relations industrial relations
Relations between the management of an industrial enterprise and its employees.
the relations between management and workers , 42, 2003, 650-91.
--. "Human Resource Practices, Unionization and the Organizational Efficiency of French Industry," in Advances in Industrial and Labor Relations, edited by E. Levin and B. Kaufman. Amsterdam, The Netherlands: JAI JAI Java Advanced Imaging
JAI Justice et Affaires Interiéures (French: Justice and Home Affairs)
JAI Journal of ASTM International
JAI Just An Idea
JAI Jazz Alliance International
JAI Joint Africa Institute Book Series, Elsevier Science, 2006, 67-102.
Dworkin, J. B., and D. A. Ahlburg. "Unions and Productivity: A Review of the Research." Advances in Industrial and Labor Relations (A Research Annual Series), 2, 1985, 51-68.
Freeman, R. B., and J. L. Medoff. What Do Unions Do? New York: Basic Books, 1984.
Greene, W. "Fixed and Random Effects Random effects can refer to:
Henderson, D. J., and S. C. Kumbhakar. "Public and Private Capital Productivity Puzzle: A Nonparametric Approach." Southern Economic Journal, 73, 2006, 219-32.
Henderson, D. J., and R. R. Russell. "Human Capital and Convergence: A Production Frontier Approach." International Economic Review, 46, 2005, 1167-205.
Heshmati, A., S. C. Kumbhakar, and L. Hjalmarsson. "Efficiency of the Swedish Pork Industry: A Farm Level Study Using Rotating Panel Data, 1976-1988." European Journal of Operational Research, 80, 1995, 519-33.
Hirsch, B. T., and A. N. Link. "Labor Union Effects on Innovative Activity." Journal of Labor Research The Journal of Labor Research is a journal which publishes articles regarding labor relations in the United States.
The journal publishes articles which cover a wide variety of topics in labor relations, including the nature of work, labor-management relations, , 8, 1987, 323-32.
Hirsch, B. T., and D. A. Macpherson. "Union Membership and Coverage Database from the Current Population Survey: Note." Industrial and Labor Relations Review Industrial and Labor Relations Review is a publication of the Cornell University School of Industrial and Labor Relations. It is an interdisciplinary journal publishing original research on all aspects of labor relations. , 56, 2003, 349-54.
Hoxby, C. M. "How Teachers' Unions Affect Education Production." Quarterly Journal of Economics, 111, 1996, 671-718.
Ichniowski, C., and K. Shaw. "Beyond Incentive Pay: Insiders' Estimates of the Value of Complementary Human Resource Management Practices." Journal of Economic Perspectives, 17(1), 2003, 155-80.
Jensen, M. C., and W. H. Meckling. "Theory of the Firm: Managerial Behavior, Agency Costs Agency Costs
The costs resulting from an agent performing services for a principal.
Agency costs are generally the commissions earned by agents.
See also: Agency Problem, Agent, Principal
Agency costs and Ownership Structure." Journal of Financial. Economics, 3, 1976, 305-60.
Kuhn, P. "Union Productivity Effects and Economic Efficiency." Journal of Labor Research, 6, 1985, 229-48.
--. "Unions and the Economy: What We Know; What We Should Know." The Canadian Journal of Economics, 31, 1998, 1033-056.
Kumbhakar, S. C. "Estimation of Technical Inefficiency in Panel Data Models with Firm- and Time-Specific Effects." Economics Letters Economics Letters is a scholarly peer-reviewed journal of economics that publishes concise communications (letters) that provide a means of rapid and efficient dissemination of new results, models and methods in all fields of economic research. Published by Elsevier. , 36, 1991, 43-8.
Kumbhakar, S. C., and H. J. Wang. "Estimation of Growth Convergence Using a Stochastic Production Frontier Approach." Economics Letters, 88, 2005, 300-05.
--."Estimation of Technical and Allocative Inefficiency: A Primal pri·mal
1. Being first in time; original.
2. Of first or central importance; primary.
pri·mali·ty n. System Approach." Journal of Econometrics, 134, 2006, 419-40.
Lewis, Harold G. Unionism and Relative Wages in the United States: An Empirical Inquiry. Chicago: University of Chicago Press The University of Chicago Press is the largest university press in the United States. It is operated by the University of Chicago and publishes a wide variety of academic titles, including The Chicago Manual of Style, dozens of academic journals, including , 1963.
Link, A. M., and D. S. Siegel. "Unions and Technology Adoption: A Qualitative Analysis Qualitative Analysis
Securities analysis that uses subjective judgment based on nonquantifiable information, such as management expertise, industry cycles, strength of research and development, and labor relations. of the Use of Real-Time Control Systems in U.S. Coal Firms." Journal of Labor Research, 23, 2002, 615-30.
Lommerud, K., F. Meland, and O. R. Straume. "Globalization globalization
Process by which the experience of everyday life, marked by the diffusion of commodities and ideas, is becoming standardized around the world. Factors that have contributed to globalization include increasingly sophisticated communications and transportation and Union Opposition to Technological Change." Journal of International Economics, 68, 2006, 1-23.
Machin, S., and S. Wood. "Human Resource Management as a Substitute for Trade Unions in British Workplaces." Industrial and Labor Relations Review, 58(2), 2005, 201-18.
Menezes-Filho, N., and J. Van Reenen. "Unions and Innovation: A Survey of the Theory and Empirical Evidence." CEPR CEPR Centre for Economic Policy Research (London, UK)
CEPR Center for Economic and Policy Research (Washington, DC)
CEPR Centre Européen de Prévention des Risques Discussion Paper No. 3792, 2003.
Munnell, A. H. "How Does Public Infrastructure Affect Regional Economic Performance?" New England New England, name applied to the region comprising six states of the NE United States—Maine, New Hampshire, Vermont, Massachusetts, Rhode Island, and Connecticut. The region is thought to have been so named by Capt. Economic Review, September/October 1990, 11-33.
Wang, H. J. "Heteroscedasticity and Non-Monotonic Efficiency Effects of a Stochastic Frontier Model." Journal of Productivity Analysis, 18, 2002, 241-53.
DEA: Data Envelopment Analysis
LR: Likelihood Ratio
(1.) As noted by Kuhn (1998), a fair summary of the literature suggests that not only are both positive and negative effects conceivable, but both may well exist in different industries and situations. Supporting this impression, Hoxby (1996) used district-level data to find a negative effect of unions on productivity of U.S. public elementary schools. Black and Lynch (2001) found that unions' productivity effects depend on other aspects of the human resources environment, which tends to be negative when unions maintain more traditional labor management relations, but positive when unions adopt new or "transformed" industrial relations practices that promote joint decision making coupled with incentive based compensation. A recent meta-analysis by Doucouliagos and Laroche (2003) on 73 studies shows negative association between unions and productivity for the United Kingdom, whereas a positive association for the United States in general and for U.S. manufacturing.
(2.) While there is substantial intra-industry variation in union coverage in the United States, at least at the two-digit SIC-level, 75 of 82 census industry codes (three-digit CIC CIC
circulating immune complexes.
CIC Circulating immune complexes. See Immune complexes. ) in the U.S. manufacturing sector have experienced declining unionization rate, on average 13%, from 1983 to 1996 (www.unionstats.com). This study also found a similar decline in state-level unionization rate for the manufacturing sector as a whole (Table 1).
(3.) It should be noted that here we employed a translog specification on the stochastic frontier model which is considered more flexible than the Cobb-Douglas specification.
(4.) In Equation (1), we allow unionization to affect the frontier by including the union variable into the production function as suggested by empirical evidence on unions and technology change. For example, Link and Siegel (2002) investigated the relationship between unions and implementing new technologies by utilizing a qualitative analysis (field-based methods) in the U.S. coal mining industry. The findings suggest that unions constitute an obstacle in adopting and using new technologies. Menezes-Filho and Van Reenen (2003) focused on surveying recent microeconometric results in the areas of R&D, innovation, technological diffusion, and productivity growth. Empirical evidence shows consistently strong and negative impacts of unions on R&D in North American North American
named after North America.
North American blastomycosis
see North American blastomycosis.
North American cattle tick
see boophilusannulatus. studies.
(5.) Non-neutral technical change has been widely used in studying stochastic frontier models (Heshmati, Kumb-hakar, and Hjalmarsson 1995). In addition, recent empirical evidence on aggregation production (Henderson and Russell 2005) suggests that the production function shifted in a non-neutral way.
(6.) Cohen and Morrison Paul (2004) used a similar strategy to determine the price of labor at the state-level manufacturing sector.
(7.) The likelihood function can be found in the appendix of Battese and Coelli (1993).
(8.) The data excluded the states of Alaska and Hawaii and the District of Columbia District of Columbia, federal district (2000 pop. 572,059, a 5.7% decrease in population since the 1990 census), 69 sq mi (179 sq km), on the east bank of the Potomac River, coextensive with the city of Washington, D.C. (the capital of the United States). . The data coverage was jointly determined by the availability of all variables used in the model. While data on some variables were available to date, the others were limited to preserve data consistency Data consistency summarizes the validity, accuracy, usability and integrity of related data between applications and across the IT enterprise. This ensures that each user observes a consistent view of the data, including visible changes made by the user's own transactions and . Because the U.S. Bureau of Economic Analysis does not construct state-level private capital in the manufacturing sector, we relied on the dataset constructed by Cohen and Morrison Paul (2004), considered the most recent and refined state-level data estimate for private capital stock in manufacturing. However, the availability of data was limited to 1996 because of the change of SIC industry definitions (before 1997) to NAICS NAICS North American Industry Classification System industry definitions (after 1997). This discontinuity dis·con·ti·nu·i·ty
n. pl. dis·con·ti·nu·i·ties
1. Lack of continuity, logical sequence, or cohesion.
2. A break or gap.
3. Geology A surface at which seismic wave velocities change. results from many sources, including differences in source data and different estimation methodologies that may affect estimates for state-level private capital stock. Although the data employed in this article might appear out-dated, the empirical evidence found herein proves to be relevant and useful as several studies on union efficiency/productivity in the United States employed similar period data. Therefore, this paper's results can serve as a comparison (Doucouliagos and Laroche 2003, Table 1).
(9.) Workers are counted as covered by a collective bargaining agreement if they are union members or if they are not members, but are covered by a union contract. For greater details, see Hirsch and Macpherson (2003). The database also provides union membership, which is the percent of employed workers who are union members. Using union membership, instead, does not change the findings of the paper.
(10.) Researchers have often used state-level manufacturing private capital stock data when dealing with U.S. capital stock (Henderson and Kumbhakar 2006).
(11.) See Cohen and Morrison Paul (2004) for more details. The author is grateful to Daniel J. Henderson for making the data available.
(12.) LR test of one-side error with the null hypothesis of [gamma] = [[delta].sub.o] = [[delta].sub.us] = [[delta].sub.i] = [[delta].sub.un2] = [[delta].sub.union] = 0.
Chintrakarn: Lecturer, Business Administration Division. Mahidol University Mahidol University (Thai มหาวิทยาลัยมหิดล, pronounced: ma-hi-don) is a public university in Thailand. International College (MUIC), Salaya, Nakornpathom, Thailand. Phone 66 (0) 2-441-5090#1238. Fax 66 (0) 2-441-5095, E-mail email@example.com
Chen: Assistant Professor, Department of Economics, Tamkang University Tamkang University (Traditional Chinese: 淡江大學; Simplified Chinese: 淡江大学 , Tamsui 251, Taiwan. Phone 886-2-26215656 #2862, Fax 886-2-26209654, E-mail firstname.lastname@example.org