Introduction

From the evolutionary approach of organizations presented in the work by ^{Hannan and Freeman (1977)}, the impact of environmental factors on the differentiation of organizations as a survival process has been researched. However, based on ^{Powell and Di Maggio (1999)}, a series of studies derived from the premise which the environment causes organizations that are grouped in certain sectors to experience isomorphism, that is, a process of homogeneization process of organizational structures.

The literature addresses several factors that cause isomorphism, such as accounting legislation reforms (^{Doadrio, Alvarado and Carrera, 2015}), the interference of regulatory agencies (^{Moura, Domingos, Cabral and Santos 2014}; ^{Filho and Sousa, 2013}), the implementation of environmental practices (^{Gellers, 2012}; ^{Morais, Olveira and Souza, 2014}; ^{Borges, Dutra and Scherer, 2014}), culture (^{Gallén and Peralta de Grado, 2016}; ^{Fuentes, 2013}), the general legal system, governance and globalization (^{Gallén and Peralta de Grado, 2016}), the participation of the Dow Jones Sustainability Index (^{Jacomossi, Casagrande and Reis, 2015}), foreign financing (^{Mir and Bala, 2015}; ^{Salas, 2010}), and the clients, workers and unions (^{Filho and Sousa, 2013}). Evidently, the role of the workforce in organizational homogenization has been barely studied.

Regarding the workforce, migration is a variable of particular importance because it acts as an environmental mechanism that influences organizations by altering the available workforce, and thus determines its evolution and structure to a certain degree (^{Manic, 2016}; ^{Castelhano, Lawell, Sumner and Taylor, 2016}; ^{Mora Rivera and Arellano González, 2016}; ^{Santermer and Giangaspero, 2011}; ^{Nava-Tablada, 2012}; ^{Chen, Frank, Thomas and Jianguo, 2012}). In the case of Mexico, there are few studies that address isomorphism (^{Sánchez Ramos, 2007}; ^{Ruiz Ortega, 2015}; ^{Lara Gómez and Pérez Sosa, 2015}; ^{Rivera and González, 2016}; ^{Cabrera Rubio, 2016}; ^{Rosas, Sánchez and Martínez, 2012}); yet although ^{Gonzáles Zepeda and Escala Rabadán (2014)} analyze the case of migrants from Michoacán, they do not deem migration as a factor for isomorphism. In this sense, considering the historical and recent importance of national and international immigration in Mexico, and that organizational theory highlights that the environment acts as a symbiotic force in the development of organizations, interest arises in researching the role of immigration of skilled labor in the isomorphism of the corporate structure of the main Mexican cities.

In this manner, this study continues in the second section addressing theoretical aspects regarding the two dimensions of analysis, i.e. the organizational environment and migration. On the one hand, organizational theory points out the argument that the environment and the organization affect each other (there is endogeneity); also (along with some migration theories discussed in section 2.2), that the evolution of both can be shown on an isomorphism caused by the highly skilled workforce that looks for greater competitiveness. Subsequently, the third section presents the methodology used to empirically analyze the effect of skilled labor migration on the isomorphism of the economic sectors of seventy-eight Mexican cities controlling the endogeneity of migration. The fourth section shows the results, and the fifth section presents some conclusions.

Theoretical aspects

Methodology

To estimate the effect of an environmental factor on organizational isomorphism, regardless of the effect of other environmental factors, an statistical analysis of the interest variables is carried out, so as to subsequently adapt a multivariate econometric analysis of the interest variables. For this, the variables proposed in the analysis are based on theory or on empirical evidence of previous studies. Regarding the dependent variable related to isomorphism, ^{Powell and Di Maggio (1999)} mention that it is reflected on the organizational characteristics related to the market (competitive isomorphism) or to the rules, policies or legislation (institutional isomorphism); therefore, the proposal is to carry out the estimation of two dependent variables that correspond to these two types of isomorphism. Similarly, ^{Powell and Di Maggio (1999)} mention that the unit of analysis in isomorphism research can not only comprise the companies that compete with each other, but even covering the totality of the important actors. Thus, the analyzed organizational field in this study is the total number of companies established in the seventy-eight main Mexican cities.

According to ^{Hannan and Freeman (1977)} cited by ^{Powell and Di Maggio (1999)}, competitive isomorphism can be generally measured by the changes in market niches; this ranges from modifications in the business structure in a specific field to modifications in groups of aggregated economic sectors. Thus, it is possible to investigate if a decrease in the number of economic sectors in a city is influenced by competitive isomorphism mechanisms that seek, for example, to make use of the economies of scale in an industry through their concentration in a locality. On the other hand, as previously mentioned, institutional isomorphism can be measured by variables related to change in organizational policies; thus, the idea is to measure this type of isomorphism through the ratio of companies that adopt a quality or sustainability policy manifested through the acquisition of an ISO 9000 or ISO 14000 certification. Regarding the explicative variable, the aim is to identify whether a greater immigration of skilled people (potential workers) has an effect in a competitive or institutional isomorphism in the established companies of a specific city, measured by a convergence in the attainment of certifications or a fewer number of economic sectors.

According to econometric theory, it is important to include the appropriate control variables to obtain a reliable estimator of the independent variable of interest. Based on theory and the empirical evidence referred to in the first section, these variables include forces related to the degree of dependence of the organization to external factors. ^{Powell and Di Maggio (1999)} argue that the dependence and the imitation of other organizations are coercive and mimetic mechanisms of isomorphism; therefore, the number of large companies in the CNN-expansion ranking of the 500 more important companies of Mexico (empgrand) were included, since they are considered leaders. Furthermore, given that they refer to isomorphism as a process in which organizations change according to the changes of others in the same field, it is important to control the past values of the dependent variable (the certification or the diversification). In the same manner, ^{Powell and Di Maggio (1999)}, ^{Mir and Bala (2015}), and ^{Salas (2010)} suggest that a greater dependency of external financing generates greater isomorphism, therefore, the amount of credits granted to the companies in pesos (créditos) and net foreign investment (ied) are included. A productivity variable is also included, as it is linked to isomorphism mechanisms such as the interference of regulatory agencies, workers and syndicates (^{Filho and Sousa, 2013}), culture (^{Fuentes, 2013}), and the implementation of environmental practices (^{Gellers, 2012}). Finally, a variable of the population with higher education (pobeducsup) is also included given that ^{Powell and Di Maggio (1999)} argue that professionalization is a normative mechanism of isomorphism.

Econometric model

Finally, endogeneity is econometrically addressed through instrumental variables, that is, variables that are not correlated to the dependent variable, but are correlated to the explicative endogenous variable (^{Wooldridge, 2009}). When data are available in time, as is the case of longitudinal panel, the strategy is to use lags of the variables as instruments of themselves (^{Cameron and Trivedi, 2010}). Thus, to estimate the effect of skilled labor migration on the two variables of corporate isomorphism controlling the endogeneity of the migration variable, the following econometric model is proposed (1):

2

where the isomorphism indicator (vd) is expressed by the number of economic sectors (diversec) based on the North American Industry Classification System (NAICS) or on the number of companies with ISO 9000 and ISO 14000 (eciso) certification for each thousand companies of entity i in time t. For the migration indicator for skilled labor (mtc) the number of migrants above 25 years of age with a higher education in thousands (migc) or the ratio of these migrants with relation to the total population (pmigc) of entity i in time t is utilized as variable. C is the vector of control variables explained in the abov section, ε is the idiosyncratic error, α and β are parameters, and γ is a parameter vector.

Results

Figure 1 shows the number of classes of economic sectors in the seventy-eight cities considered in the analysis. From this graph, fit can be first observed that the main cities of the Mexico (Mexico city, Monterrey and Guadalajara) have a broad diversification of economic classes, approximately 900⁴, due to their size. On the other hand, the cities with a lower average diversity were Rio Verde-Ciudad Fernández and Guanajuato (a little less than 360 economic classes). Another interesting information is that in most cities (56 out of the 78 analyzed) the number of economic sectors decreased between 2012 and 2013, indicating a smaller diversity. Disaggregating the analysis by levels, this decrease was mainly present in those cities with a little more than 600 economic sectors.

Figure 1 Diversification of economic sectors by city 2012-2013. 

The fact that the cities with greater diversity are those that have a bigger population, in addition to the fact that most showed a decrease in the number of sectors since 2013, could arise the problem of presenting a spurious relation to arguing that the migration of skilled people is positively correlated to the size of the locality where they migrate; however, by observing Figure 2 ⁵, it is clear that there is not a constant positive relation between the number of sectors and the ratio of skilled migrants, but it rather increases to a decreasing rate.

Figure 2 Economic sectors and skilled migrants in Mexican cities (2014) Skilled migrants as a ratio of the total population above 25 years of age. 

In the same manner, it is evident that the average and the standard deviation of the number of sectors decreased during the analyzed period (quadrant I of Figure 3), which according to ^{Powell and Di Maggio (1999)} suggests the presence of isomorphism. Besides, the number of skilled migrants and their ratio in accordance with the total population in the same age range (25 years or more) show a sustained increase in the observed period, with a slight sustained decrease in the ratio of skilled migrants between 2013 and 2014. From this figure, the inverse relation between both variables in time is evident, i.e., while the number of migrants and their ratio increased, the number of economic sectors decreased. However, although this inverse relation seems to back up the hypothesis that skilled labor migration causes isomorphism, it is necessary to carry out the adequate analysis of causality through the econometric model.

Figure 3 Evolution of the economic sectors and skilled migrants in Mexican cities (2008-2014). Averages and standard deviations correspond to the total number of cities. 

Departing from the relation found in Figure 2, model (1) was estimated by first analyzing whether there was evidence of a positive linear relation, and then allowing a quadratic behavior. In models 1 and 2 of Table 1, migc is used as an independent variable, and pmigc is used for the rest; the functional form of models 1, 3 and 5 allows a linear effect of migration, while models 2, 4 and 6 assume a quadratic relation; models 3 and 4 prove whether or not the significance of pmigc is robust despite the high correlation (Pearson correlation coefficient of .99) of pobeducsup with the above of 25 years of age population variable utilized in the construction of pmigc. Finally, it is important mentioning that tests were done to confirm the compliance of the consistency assumptions for Arellano-Bond estimators as well as valid overidentification by Sargan, with all models being 95% adequate in both tests. Moreover, given that the standard errors normally reported show bias (^{Cameron and Trivedi, 2010}), robust standard errors are utilized, which also resolve potential heteroscedasticity in the idiosyncratic error.

Table 1 shows that the créditos coefficient for companies presents a different sign depending on whether the number or ratio of skilled migrants is included in the model; when the number of migrants is included, the créditos coefficient has a positive sign, contrary to the theory, but it is only significant in model 2 (although not all decimals appear due to space issues, around two hundred billion pesos in credit are needed for companies to create one more economic sector). In the case of models that estimate the ratio of migrants, the sign of the créditos coefficient for companies is negative according to the theory, but is only significant at 95% in models 5 and 6; furthermore, the effect is once again minuscule. These discrepancies question the role of the credits in the economic diversification of the cities, and encourage the realization of a more in-depth study on the characteristics of these credits as the granting mechanisms, the acquierers characteristics, and the real destination of resources. For example, it is possible that a considerable portion of these credits are destined to small companies that offer traditional products or services such as food preparation (restaurants) or beauty salons, which belong to economic sectors that present few or no changes over time, and which require little innovation in their processes and, therefore, do not have a significant interference in the diversity of other economic sectors.

Table 1 shows that the foreign direct investment (FDI) coefficient has the expected sign and is significant in the first four models, indicating that an increase of approximately one hundred million dollars of FDI leads to a reduction of 1 or 2 economic sectors on average. This result provides evidence that the economic sectors in the Mexican cities have a greater dependency (as organizational field) on FDI than locally granted credits, possibly because they are mainly destined to research projects and development, which translates into greater isomorphism. Finally, there is no doubt that the last control variable included (large companies) influences isomorphism, as it is significant in all models; however, it is inconsistent by having the expected sign only in models 5 and 6 of Table 1, when it is not controlled by population with a higher education. Evidently, there is a high correlation between the variables of large companies and the population with higher education (Pearson correlation of .98), which makes it difficult to isolate the partial effect of both variables.

Regarding the estimators of the interest variable of models 1 and 2 of Table 1, it can be observed that the number of skilled migrants variable presents a positive sign and there is no statistical evidence that the relation is quadratic. This coefficient explains that if the number of skilled migrants in a city increases by approximately two thousand individuals, then an economic sector is expected to rise; in other words, there is no evidence that a greater number of migrants will lead to isomorphism, as argued after Figure 3. However, the arguments presented in figures 2 and 3 on the effect of the ratio of skilled migrants on the number of economic sectors are supported by models 5 and 6. The coefficient of the migrant ratio in model 5 indicates that an increase of 1 percentage point leads to a decrease of 1.6 economic sectors on average (there is isomorphism according to what was expected). Besides, the specification of model 6 is the only one with all the significant coefficients and the expected sign; in this case, it indicates that as the ratio of migrants with higher education increases, the number of economic sector classes in a city increases to a decreasing rate, meaning that there is evidence of a threshold in that ratio (approximately 10.34%) from which a higher ratio of skilled migrants in a city decreases the number of economic sectors, i.e. there is isomorphism. Therefore, it seems to be a reconciliation with the findings of models 1 and 2 in which a positive effect on a greater number of migrants is found: in order for isomorphism to be present, it is not enough to have a greater immigration of skilled workers, as the population of the city will naturally increase; it is necessary that these skilled immigrant workers cross a threshold so that they represent the majority in relation to the rest of the skilled people. These results make sense under the argument that the greater a group of workers with a certain ideological profile, the easier it is for that ideology to be instilled in the other workers, which is basically the justification behind normative isomorphism presented by ^{Powell and Di Maggio (1999)}.

Continuing with the analysis of institutional isomorphism, Figure 4 shows something similar to Figure 1 in the sense that the most economically relevant cities or entities are those with a greater number of companies certified with ISO. Moreover, it can be observed that even though the number of these companies increased annually in all entities, it is quite evident the relatively low number of companies that obtain this certification in the main Mexican cities- where the average is of only approximately 71 per thousand companies in the city-. Also, the relation between the number of certified companies and migrants with higher education is shown in Figure 5, which seems to support the hypothesis of the study as there is a clear positive relation between both variables in each year.

Figure 4 Companies with ISO 9000 and ISO 14000 certification by entity. Number of annually certified companies per thousand (2008-2014). 

Figure 5 Migrants with higher education and companies certified with ISO 9000 and ISO 14000 State data for years 2008-2014. 

Thus, the estimations of model (1) are shown in Table 2 ⁶, considering the logarithm of the number of ISO 9000 and ISO 14000 certified companies as dependent variable. In models 1 and 2 the ratio of skilled migrants was estimated. The results of Table 2 indicate that the models were estimated with data for the 32 states of Mexico for four years (due to the loss of information when using the instruments); furthermore, all the models were statistically significant overall. Almost all variables were individually significant, although some present the opposite sign to what was expected, since according to the theoretical discussion on isomorphism it is expected that an increase in these would generate a greater number of certified companies.

Thus, Table 2 first shows that the dynamic of the dependent variable in time is significant and positive, that is, an increase in the number of ISO 9000 and ISO 14000 certified companies in a certain year will have a positive effect of almost the same magnitude in the number of certified companies the following year. This result was evident in Figure 4, where the sustained increase in the number of certified companies for all states in the country can be clearly observed.

Therefore, in this case, it can be confirmed that institutional isomorphism follows the feedback of the same homogenization process, i.e. there is a “spread” through which the presence of certified companies causes other companies to seek and obtain a certification.

The models of Table 2 show that an increase in the population with higher education has a negative effect on the participation of certified companies per every thousand. If the increase in population with higher education occurs only through residents (non-migrants), then an additional 50 thousand skilled migrants would be needed for the number of certified companies per every thousand to decrease by 1%, and considering that this average is lower than one hundred companies (Figure 4), then many more residents would be needed to reduce a certified company. Now, if the inclusion of migrants with a higher education is allowed, an increase in the skilled population of around 20 thousand people would be needed for such reduction to occur. The negative effect of the pobeducsup variable towards the certified companies variable can be explained when changes in this segment of the population are linked to a disproportionate change in the relation of certified companies to total companies. In other words, it is possible that a greater population with higher education increases both the number of total companies as well as the number of certified companies, but the speed with which the number of total companies increases is greater (due to the time that it takes to obtain an ISO certification and/or because some companies cannot cover the costs), which reduces the value of the eciso variable despite the fact that there are actually more certified companies.

A similar effect shown in Table 2 is the increase in foreign direct investment and the number of large companies; a 1% increase in the FDI or in a large company causes a decrease in the number of companies certified with ISO 9000 and ISO 14000 of approximately 0.12% or 0.01%, respectively. Once again, the explanation is based on the relation that the FDI or the establishment of large companies could have with the relation of certified companies to total companies: it is possible that the FDI flow or the establishment of a large company could have a lower impact on the certification of existing companies.

Although créditos coefficient shows the expected sign according to the theory in all models, it was only significant when the number of skilled migrants remained constant (models 1 and 2 of Table 2), and its effect was small: a 1% increase in the credit to the companies of a city on average increases the number of certified companies by 0.05%. Contrary to what happened with the FDI and the establishment of large companies, it is clear that this variable would have the expected effect, given that the resource is susceptible to be destined to the payment of the certification rather than to the creation of new companies. Furthermore, the productivity coefficient presented the expected sign and was very significant in all the models; in this case, a one percentage point increase in productivity on average leads to an increase of 0.25% in the number of certified companies.

The models of Table 2 show that the number and the ratio of skilled migrants in a city has an effect according to the theory in both functional forms (linear and quadratic); however, the ratio of migrants was not significant (models 3 and 4), indicating that in the case of the certified companies, the presence of skilled migrants is more relevant, regardless of how many non-migrant skilled workers there are in the city. When the number of skilled workers (model 1) is linear and show an increase, the estimation shows that the number of certified companies also increases by .15% on average. This is interesting and supports the hypothesis of the study regarding the effect of skilled labor migration on isomorphism and the findings of Figure 5. Furthermore, the quadratic model (model 2) complements the findings, indicating that if the number of skilled migrants is low in a city, then an increase of these migrants causes a decrease in the number of certified companies per every thousand in almost the same magnitude, possibly due to its greater effect on the creation of new companies. However, when the city has crossed a threshold of migrants (calculated to be 75 thousand), the effect is reverted, increasing the number of certified companies.

It can be inferred from Figure 5 that the threshold is related to the number of certified companies in a city, that is, when this number is small in a city, then skilled migrants cause a decrease in the number of certified companies, possibly because they contribute to the creation of new companies more than to the certification of them. In the same sense, based on Figure 4, it can be inferred that the threshold is related to the size of the city, thus suggesting that in a small city, skilled migrants have a greater influence in the creation of new companies than on the certification of them. Under this reasoning, corporate isomorphism manifested in the ISO 9000 and ISO 14000 certification is affected by skilled labor migration (normative mechanism); however, this isomorphism should be conceived as a long-term process that is also affected by other coercive, normative, and mimetic isomorphism mechanisms.

Conclusions

This study empirically addresses theoretical argument that the organizations, as open systems, are affected by the environment (and vice versa) through skilled labor migration. On the one hand, the study starts, from the idea that the environment causes organizations to reach homogenization by being pressured by environmental mechanisms; on the other hand, skilled labor migration (as an environmental element) is strongly linked to the company and its search for greater competitiveness.

Thus, two indicators were addressed. These, according to the theory presented by ^{Powell and Di Maggio (1999)}, can reflect the organizational homogenization or isomorphism process of the business sectors of Mexican cities: the first, related to competitive isomorphism, is the diversity of economic sectors in one city, which is established under the argument that competitive advantages cause the development of clusters or specialized groups of companies that in turn reduces the number of classes of economic sectors. The second indicator, related to institutional isomorphism, is quality certified companies, which is established under the argument that companies seek to “stay in the game”, meaning that they aim to maintain themselves in a level of quality that allows them to compete with other companies, for which they obtain the certifications.

In order to obtain consistent estimators for the isomorphism mechanisms, the empirical analysis is based on a convenient methodology. Thus, the analysis of skilled labor migration and the environmental organizational theory consider the potential endogeneity of migration through the use of instrumental variables. By using other normative, mimetic and coercive isomorphism mechanisms as control variables, the evidence suggests that the effect of skilled labor migration has an effect on both a decrease of the economic sectors and the quality certification of companies. In other words, evidence was found regarding a causal effect of the migrant workforce towards the business sector of Mexican cities. The evidence based on both indicators show, that the relation between isomorphism and skilled labor migration is not linear, but that the levels or ratio of skilled migrants requires the crossing of a threshold to cause isomorphism; therefore, in this case, skilled labor migration is a mechanism that intervenes in isomorphism in the long-term.

Although the results suggest that skilled labor migration causes a greater competitive and institutional isomorphism in the business sectors of the Mexican economies, it is convenient to carry out this analysis for longer periods of time, as well as using more isomorphism indicators to strengthen the findings of this study.