Printer Friendly

Gross flows of formal and informal workers in the Mexican labor market.

1. Introduction

Changes in absolute levels of unemployment and employment are usually small in most economies. In the Mexican labor market, aggregate unemployment has remained relatively low, and has shown little change over time compared with other developed countries. When looking at the official unemployment statistics from 2005 to 2011, the average rate of unemployment in Mexico was relatively stable, about 4.4% with a standard deviation of 0.88. Canada had a higher mean unemployment of about 7% with standard deviation of 0.99, and US had a mean unemployment of 6.85% with a 3.2 standard deviation in the same period. The US had a larger unemployment rate and greater fluctuation during this period due to the financial crisis in 2008. Just looking at these simple statistics it seems that the Mexican official unemployment is lower and fluctuates much less than the US and Canada. However, just examining the stability of the aggregate unemployment rate does not offer information about how often workers enter the labor force and stay employed or unemployed.

The absolute numbers (stocks) of workers employed, unemployed or outside the labor force are the most common labor statistics available. Using these statistics, the unemployment rate and other important indexes can be easily estimated. But gross flows of workers also provide important information for policy makers in charge of macroeconomic policy and stability. The flows and transitions help us to understand how often people change their labor status over time, and this information has important macroeconomic implications. The information on the cyclical components of gross flows of workers is relevant for researchers as well as policy makers because it helps us to understand the dynamic characteristics of the labor market during periods of economic downturn and recovery. Information like this may help to construct anticyclical policies at macro-level, but also to understand workers' behavior and heterogeneity on a more micro level. For example, understanding the labor flows of young and female workers may help policymakers to develop specific policies to promote stable employment for these groups.

This paper contains an analysis on the gross flows of Mexican workers using the microdata from the Mexican National Survey of Employment and Occupations (ENOE) from 2005 to the first quarter of 2012. The ENOE began in 2005 as the result of merges and improvements in some previous labor surveys. This new survey is representative for the whole country rather than only urban workers as in previous labor surveys.

Although the ENOE is reliable and consistent, we must also keep in mind that this survey continues to differ from labor surveys in other countries. For example, the differences between Mexico and the US-Canada labor aggregates go beyond the mere methodologies of the labor surveys: it has to do with the different methods of household accounting. The tax system in US and Canada request adult population to keep a basic accounting of their households, making easier to separate household income from that of a family business. In Mexico, not all households are required to file income tax returns, and most family businesses do not separate their accounts from that of the household itself. This leads to a problem often referred in the economic literature as the informal economy. The reason for, and nature of, informality is still an ongoing debate. One explanation might be that the Mexican labor market has different institutional arrangements which are said to produce an informal sector and reproduces other types of jobs (workers) commonly known as "informal workers". Gong and Van Soest (2002) summarize the two views about informal sector. The first view is that informal jobs are secondary jobs and workers might prefer to work in the formal sector, which is rationed. In this view informal employment is a buffer between formal employment and unemployment. The second view is that workers are heterogeneous but choose optimally to be informal or formal depending in their marginal productivities.

The Instituto Nacional de Estadistica, Geografia e Informatica, INEGI, defines informal employment (occupation) as employment in an economic unit that uses resources from the household but without legally becoming a firm, so that its activities cannot be differentiated from the activities of the household itself. This is the official concept of informality that is included in the ENOE and used throughout this paper to separate formal and informal employment.

Literature about gross flows of workers in Mexico is scarce. Bosch and Maloney (2007) wrote a well-known work of gross flows in Mexico. They include information on gross flows and transitions, which are estimated for urban workers, and consider the option of employment in the informal sector. They estimate the transition of work categories using longitudinal data from the Mexican National Survey of Urban Employment (ENEU) from 1987 to 2002, which is a predecessor of the ENOE. The main problem of their work is that the data used was only representative for urban workers in some large cities.

Levy (2008) conducted a comprehensive study on worker conditions in the labor market with the presence of informality. He offers an analysis of worker mobility using data from the Mexican Institute of Social Security (IMSS). His analysis shows how often workers remain in the formal sector, considering income and individual differences. He also analyzes aggregate employment data, in order to compare the percentage of formal and informal workers, using the ENEU from the years 1998 and 2001 as well as the ENOE from 2006. He does not estimate any kind of flows or transitions in the labor market. Similarly, Kaplan, Martinez and Robertson (2007) studied the employment dynamics in Mexico also using the IMSS data. They matched workers and firms to develop a measure of access and job separation in order to identify net job creation and destruction. The limitation of this analysis is that the IMSS data is only available for the formal sector.

Blanchard et al. (1990) wrote a pioneering analysis of gross labor flows for US. They estimate the gross flows using adjusted data and estimate the seasonal and cyclical properties of the flows. They use the Current Population Survey (CPS) dataset from 1968 to 1986 and disaggregate data by age and sex. They also include a simple model to analyze the cyclical behavior of workers. The estimation of adjusted gross flows follows the methodology of Abowd and Zellner (1985) to estimate the missing observations in the CPS every month. Jones (1993) analyzed of gross flows of labor for Canada using the Labor Force Survey, which contains monthly data from 1976 to 1991. He employs unadjusted data to estimate the flows and captures the seasonal and cyclical properties of the Canadian gross flows of workers. Jones and Riddell (1998) conducted a similar analysis of gross flows using comparable unadjusted data for US and Canada. (1) Other authors, such as Davis and Haltiwanger (1998), and Faberman and Haltiwanger (2006) analyze gross flows but the focus is on workers turnover and the forces that determinate the creation and destruction of jobs in order to analyze the influences of the flows of workers as well as job creation.

The present paper follows a similar approach as in Blanchard et al. (1990) and Jones (1993). The advantage of matching workers' labor status over time is useful to understand the transitions inside the labor market. We are not analyzing job creation or destruction but the cyclical analysis intuitively shows how gross worker flows behave during times when both job creation and destruction are high. Our objective is twofold: First, to estimate the gross flows and transitions of workers in the Mexican labor market using unadjusted data; (2) and second, to test empirically possible cyclical properties of these flows.

In addition, we established an additional objective: to highlight the dynamic characteristics of the Mexican labor market in the presence of informal employment. To this end, we use twelve flows instead of the traditional six flows analysis of Blanchard et al. (1990) and Jones (1993). We denote four labor market categories as: Not-in-the-labor force (N), formal employment (EF), informal employment (El) and unemployment (U). (3) We also disaggregate the total flows by age and sex in order to consider some important individual characteristics. The period of analysis includes a short expansion phase of the Mexican GDP from the first quarter of 2005 to the first quarter of 2006, a contraction phase from the second quarter of 2006 to the second quarter of 2009 and a recovery phase from the third quarter of 2009 to 2012. Figure 1 shows the evolution of the real GDP growth rate of Mexico during this period.


The organization of this work is as follow: the first section is a brief introduction, the second contains the analysis of the gross flows for a quarterly dataset from 2005 to 2012; the third presents an analysis of the cyclical properties of the gross flows. Section four contains the conclusions and final comments.

2. Gross flows from 2005 to 2012

Labor flows of workers are affected by several factors on both the supply and the demand side of the labor market. The move from one labor category to another could be due to individual motives or to external forces governing the macroeconomy or the business environment. On the supply side, we have several individual factors that induce people to change labor status. These individual factors could be health issues, schooling, marriage, child birth, job satisfaction, family reallocation, etc. Informal workers may have their own reasons for moving into this category including factors such as the cost of compliance with different regulations, flexible working hours and job conditions, low cost of mobility and reallocation, etc.

Factors from the demand side include demand segmentation, foreign trade policy, macro-financial stability, economic crisis, bankruptcies of large corporations, etc. For these sorts of factors, there are several possible reasons for accepting work in the informal sector, including the segregation caused by unions and trades, asymmetric information, etc. As there is no consensus about the nature of informal employment, we do not intend to give an explanation for these specific flows, but to show each of the flows and their cyclical properties.

We must also consider some other problems facing the estimation of the gross flows like misclassification errors and missing observations. Blanchard et al. (1990) and Jones (1993) give a full explanation of these problems in the case of obtaining gross flows from panel data of labor statistics. Missing observations may occur in the survey itself and the degree of correlation of this problem with the labor market activity is not well known. Thus, we cannot infer how employment status has changed for people whose information is missing. The loss of information from non-random reasons is about 20% in the ENOE Sample. Another problem we face is that of spurious transitions due to the misclassification errors in the survey. The degree to which this is a problem in our sample is not known, but we assume that the errors in different stages offset each other. (4) Although we use unadjusted data, the information obtained is still of much relevance and allows us to estimate the changes in the level of each group and to use weights to get very close to the population totals.

The Mexican labor force contains approximately 45 million people, and a little more than 2 million of these people are unemployed every quarter. As already observed, the unemployment level is relatively low but we do not know how often workers move in and out of this category. We use the official definitions of unemployment and informality formulated by the INEGI. The period under study is January 2005 to March 2012, using the convention that the flow is dated from the beginning of the immediate quarter. Labor statistics from a nation-wide representative sample became available only with the ENOE in 2005, which brings our period of analysis to only 28 consecutive quarters. A positive aspect of this sample is that during this period a major financial crisis occurred with a short but distinctive business cycle, (5) which offers the opportunity to observe some cyclical characteristics of gross flows.

We also disaggregate total flows and stocks by sex and age groups including also the respective hazards (6) (see Table 1). We set the upper limit of age for the group of young workers to 29 years old rather than the 25 year-old limit of Jones (1993). The main reason for this is to include late-comers to the labor market, in order to better capture the dynamic characteristics of this group. (7)

The absorption capacity of the labor market is limited, and asymmetric information is also a problem in the Mexican labor market, so it is possible that proper matching might not be efficient (Levy, 2008:94).

Table 1 shows mean gross labor flows in four groups of workers. The most interesting aspect of the total gross flows in Mexico is perhaps the size of the flows and the heterogeneity. The category not-in-the-labor force (N) (8) has the largest flow into and out of employment, either formal or informal. Every quarter, more than 5 million workers move in and out of inactivity. The monthly average is relatively large compared to those in the US or Canada. Although the absolute flows for formal workers are larger than for informal, this has to do with the fact that informal workers are account for only 25 percent of the total labor force. As can be seen in table 1, the PEFN hazard, i.e. the probability (transition) of becoming Not-in-the-labor force after being employed in the formal sector, is lower than the PEIN hazard, which is the probability (transition) of becoming Not-in-the-labor force after being employed in the informal sector, meaning that informal workers are more likely to become economically inactive than formal workers. On the other hand, PNEP hazard, the probability (transition) of becoming employed in the formal sector after being Not-in-the-labor force is larger than PNEI, the probability (transition) of becoming employed in the informal sector after being Not-in-the-labor force, meaning that is more likely for inactive individuals move to formal jobs rather than to informal. This is especially true for male workers.

Other large flows are those inside the category of employment itself: EPEI, the flow from employed in the formal sector to employed in the informal sector, and EIEP, the flow from employed in the informal sector to employed in the formal sector. These categories show that workers move from formal to informal employment and back very often. The category PEIEF, the probability (transition) of becoming employed in the formal sector after being employed in the informal sector is almost three times larger than the PEFEI, the probability (transition) of becoming employed in the informal sector after being employed in the formal sector, which means that the probability of moving to a formal job after being in the informal sector is three times larger. This also means that either the family business has become a formal business, or some family members have become independent, and have established a formal firm, following the definition of the INEGL Moreover, the hazard category PEFEI is very large, even among sub-groups, with the only exception being the sub-group of female workers, which shows the lowest probability of moving to the informal sector after having been employed in the formal sector. The PEIEF hazard category, the probability (transition) of becoming employed in the formal sector after being employed in the informal sector is larger for male and young workers than for other categories of workers, which means that male and young workers have a greater probability of becoming employed in the formal sector after having been employed in the informal sector.

The absolute values of gross flows in and out of unemployment are generally low but the hazards of these flows show large differences. For example, the hazard category PEIU, the probability (transition) of becoming unemployed after being employed in the informal sector, is high for all workers except for female workers. With respect to the probability of becoming economically inactive, female workers show higher rates than men. The PUEF hazard, that is, the probability (transition) of becoming employed in the formal sector after being unemployed is almost the same for young and old workers but larger for male workers, and smaller for female workers. Interestingly, the PUEI hazard, i.e. the probability (transition) of becoming employed in the informal sector after being unemployed, is larger for old and male workers. This result is rather puzzling. If the informal jobs are secondary, it might be logical to think that the groups associated with informal work are young people and women. In this case, informal jobs would be considered a middle ground between unemployment and formal work. Table 1 shows that the probability of moving to the informal sector after being unemployed (PUEI) is larger for male and old workers and much lower for younger and female workers. A suggested interpretation might be that informal jobs are not temporary or marginal, but are instead a real and optimal choice similar to formal jobs.

3. Cyclical components of Mexican gross flows

One traditional approach to analyzing seasonality of time series is the dummy variable model, which assumes that the observed series is the sum of three principal parts: a seasonal component, a trend, and an error or noise. This approach is explained in Plosser (1979) as follows:

[y.sub.t] = [y.sup.c.sub.t] + [n.summation over (q = 1)][[alpha].sub.q][d.sub.qt] + [[epsilon].sub.t]

where [y.sup.c.sub.t] is the trend, usually represented by a polynomial function in t, [[alpha].sub.q] represents an estimated mean of the quarter q, [d.sub.qt] represents the dummy variable which capture the seasonal component, and [[epsilon].sub.t] is the error which also can be interpreted as a non-seasonal component. The above equation is the traditional approach for seasonal analysis of times series. It can also be used for decomposing series into seasonally adjusted series. We can also include additional functions in order to analyze business cycles using this approach as benchmark.

This section contains information about the seasonal (short cycles) and cyclical features of the gross labor flows. First we performed an analysis of seasonality of the gross flows using the above equation. The ENOE survey is dated quarterly instead of monthly, limiting our ability to see seasonal components in detail. Despite this loss of information, we still used quarterly flows to capture on average the seasonal mobility of workers. We estimated the seasonal components in a similar way as in Jones (1993), but using the regressed values of every gross flow on quarterly dummies rather than the traditional monthly dummies. The standard error of the predicted values is used to capture the seasonal components and the standard error of the residuals is used to capture the non-seasonal part. These standard deviations are shown in Table 2. The analysis shows that the seasonal components as are usually larger than non-seasonal components an in most cases. Only the flows of young workers from formal employment to informal employment (EFEI) and from informal employment to formal employment, (EIEF), and the flow of female workers from formal employment to unemployment (EFU) and from inactivity to formal employment (NEF), show larger non-seasonal components.

Another way to look at the effect of seasonality on gross flows is as a percentage of the variance of the seasonal and non-seasonal components over the variance of the predicted value of the flows: [[sigma].sup.2.sub.s]/[[sigma].sup.2] and [[sigma].sup.2.sub.n]/[[sigma].sup.2]. Table 3 shows the relative size in percentage of the seasonal effects on the gross flows by groups. In general, the gross flows of young and female workers are the least affected by seasonality. The flows from employed in the informal sector to unemployed (EIU) and from unemployed to employed in the formal sector (UEF) are strongly seasonal for young workers and the flows from employed in the informal sector to Not-in-the-labor force (EIN) and from unemployed to employed in the informal sector (UEI) are the strongest for female workers. For male and old workers all flows are strongly seasonal. The largest seasonal effects of gross flows for old workers are in the categories of employed in the informal sector to Not-in-the-labor force (EIN) and of Not-in-the-labor force to employed in the formal sector (NEF) while the EIN and UEF are the largest for male workers. We cannot specify the month of the year for which the flows are the highest or the lowest, since the data is quarterly. But it is clear that the aggregate quarterly data also show some seasonality which deserve analysis, especially for those in charge of public policy.

Figure 2 is a bar plot of the seasonal coefficients from the OLS regressions by groups. It shows that moving from formal to informal employment and back (EFEI and EIEP) is more frequent in the first quarter. It shows that leaving employment for inactivity (EFN and EIN) also happens more often in the first quarter. Here the decision to change labor status might be facilitated by the yearly bonus (aguinaldo) and other fringe benefits that allow workers to change jobs or just become inactive. The flows from unemployment into the formal and informal sectors, UEF and UEI, peak in the last quarter of the year, as do flows from economic inactivity to the informal sector, NEI, perhaps following the seasonal pattern of retail sales due to high demand at the end of every year. Business hire more personnel in this quarter, due to increased sales projections for the end of the year along with higher household consumption. The flows from informal employment and economic inactivity into unemployment (EIU and NU) are the highest in the third quarter, usually a time where some workers begin looking for new positions.

Other useful tools for cyclical analysis are scatter plots, presented here in figures 3 and 4 with the gross flows and hazards plotted along a cubic spline. This spline is allowed to change at the end of the recession. A vertical line is set at the first quarter of 2009 to divide the recession from the recovery. From figure 3 we observe that all flows in and out of unemployment (EFU, EIU, NU, UEF, UEI and UN) have a distinctive italic-style "S" shape spline. An intuitive interpretation of these figures is that the flows in and out of unemployment increase during a downturn in economic activity and decrease during economic recovery. A similar shape is shown for the PEFU, PEIU and PNU, the hazards associated with moving into unemployment from the other sectors, in figure 4.



We can also express the flows and hazards net of seasonal dummies, as in figures 5 and 6. In addition to the vertical line that divides the recession and the recovery, we also added a horizontal line to divide the negative and positive flows and hazards net of seasonal components. In this way we constructed a graph with four quadrants to examine the properties of the flows and hazards along the business cycle. For example, if the flows are counter cyclical then flows must be negative but increasing (third quadrant) during the recession and positive but decreasing during the recovery (first quadrant). Furthermore, we made a separate linear fit during the recession and during the recovery. Using this simple analysis, we observe in figure 5 that the flows into unemployment (EFU, EIU, NU), and the flows out of unemployment (UEF, UEI and UN) are negative net of seasonal components in the third quadrant and positive flows in the first quadrant. The linear fit of the first three flows have positive slope during recession and negative during recovery, while the last three have positive slopes in both periods. A similar analysis is performed on the hazards net of seasonal dummies in figure 6. The dispersion is much larger with hazards than with flows and only the PEFU, PEIU and PNU, the hazards of moving into unemployment, show evidence of counter cyclical components. But figure 6 also shows some peculiarities, including a pro-cyclical PUEF hazard that follows the business cycle, positive but decreasing flows net of seasonal components during recession (second quadrant) and negative but increasing flows during recovery (fourth quadrant).



A standard statistical analysis is needed to confirm these findings. The period under analysis includes a downturn and then a recovery of economic activity related to a short business cycle that might allow us to make statistical tests to detect cyclical properties. So we added an additional variable to detect such properties. We used the growth rate of real GDP as a cyclical variable. We made an OLS regression of the flows on the quarterly dummies, the cubic time variable and a variable that contains information about the business cycle. The estimates of these regressions on flows and hazards are shown in tables 4 and 5.


Most estimates of the cyclical component of flows associated with unemployment are negative and highly significant. The flows from employment in the formal and informal sectors to unemployment, EFU and EIU, are highly significant and countercyclical for all groups and the same happens for the flows from unemployment to employment in the informal sector and to economic inactivity (UEI, UN, respectively), and from economic inactivity to unemployment (NU). The only exception is the UEF flow, from unemployment to employment in the formal sector, which is only significant for the total, for male and for old workers. This is a standard result which shows the counter-cyclical behavior of unemployment.

The EIEP flows, from informal to formal jobs, for young and male workers are the only pro-cyclical ones in the entire regression. Thus, we would expect that in an expansion phase, more workers move from informal to formal jobs. Also the EFEI flow, from the formal to the informal sector, is pro-cyclical for young workers, meaning that less young workers will move to the informal sector during the recession, and more will do so during a recovery.

The EFN flow, from the formal sector to economic inactivity, is significant and countercyclical for old and male workers, and EIN, from the informal sector to economic inactivity, and NEF, from inactivity to formal sector employment, are significant and countercyclical only for male workers. It is not clear why these flows are countercyclical. One hypothesis might be that the distinction between unemployment and not-in-the-labor force is irrelevant. In this case, we might be counting as economically inactive those workers that were indeed unemployed.

The cyclical coefficients of the hazards are shown in table 5. All the hazard categories of unemployment, PEFU, PEIU and PNU, are counter-cyclical and highly significant for all groups, with the only exception being the PEIU for female workers. PUEF is pro-cyclical for all groups, as depicted in the graphical analysis. The hazard of moving from employment in the formal sector to employment in the informal sector, PEPEI, is pro-cyclical for young workers and the hazard associated with movements in the opposite direction, PEIEF, is pro-cyclical for the total, young and male workers. The hazards of workers moving from different forms of employment to economic inactivity, PEPN, PEIN, and from inactivity to the informal sector, PNEI, are all significant and countercyclical for male workers while PNEF, the hazard associated with moving from inactivity to employment in the formal sector, is pro-cyclical and significant for male workers, and for the total. The PEIN hazard, of becoming economically inactive after working in the informal sector, is countercyclical and significant for young workers, and PNEI, the probability (transition) of becoming employed in the informal sector after being Not-in-the-labor force, is also counter-cyclical and significant for total and for old workers.

4. Concluding remarks

This analysis of gross flows of workers in the presence of informal employment in Mexico offers interesting information about mobility of workers inside the labor market. It is important to note that the Mexican labor market is different from the US and Canadian markets, and that the Mexican labor survey reflects the particularity of this market. Although our analysis is not comparable with other countries due to different periodicity, the Mexican labor market is dynamic with large flows of workers that change labor status every quarter. One important feature to emphasize is the relative large flows in and out of economic inactivity (Not-in-the labor Force). There are institutional and economic factors that encourage workers to become inactive, which also implies an economic loss in productive activities for the entire economy. Although the reasons for becoming inactive are not explored in depth here, further research to evaluate the reasons and economic costs of this large inactive population is recommended.

The traditional view that informal jobs are marginal, and just a middle ground between unemployment and formal employment cannot be directly supported by the analysis of gross flows and hazards. The groups of young and female workers have higher hazards of becoming economically inactive, while male and veteran workers have higher hazards of working in the informal sector. This is a puzzling situation if we believe that informal jobs are just a buffer between unemployment and formal employment. The usual hypothesis is that informal employment might be supplementary for younger workers and for women. The hazards involved with movements into and out of informal work are the highest for older and male workers. This finding does not fit well with the idea that informal jobs are just supplementary. Other works in line with these findings that cast some doubts about this duality view are Maloney (1999, 2003), although further research is needed to confirm this.

Although gross flows and probabilities cannot be compared with other countries like the US or Canada, it is possible to find some other similarities. The cyclical properties of gross flows in Mexico are in part similar to those in US and Canada. All three countries have strong seasonal components. Furthermore, these countries have gross flows into and out of unemployment that are countercyclical.

The analysis shows that seasonal components have less of an effect on gross flows of young workers. On the other hand, gross flows of female workers are much less affected by the business cycle. Gross flows related to unemployment are countercyclical for almost all groups which is a standard result. The implications of the cyclical analysis of gross flows are important for public policy. First, if we want to decrease the size and importance of the informal sector we must improve labor conditions for male and older workers, many of them head of households and main income earners. These workers are moving often into and out of informality, even more often than young and female workers. This conclusion has important implications because it suggests that informal employment might be a real substitute for formal employment. Furthermore, male and older workers are strongly affected by seasonal components and the business cycle. Thus, it might be worth considering public policies that promote stability for male and old workers.

In the other hand, younger and female workers are more likely to move into economic inactivity, which hinders the potential product of the entire economy. If the reasons for becoming inactive are retraining or formal education then it is an investment in human capital, but if younger and female workers become inactive dependents and discouraged workers (9) then, there is a loss in economic activity. In that case, public policies to keep young and female workers in the labor market should be considered.

Fecha de recepcion: 26 IV 2012

Fecha de aceptacion: 01 VII 2013

Roberto Gallardo del Angel

Universidad Veracruzana


Abowd, J.M., and A. Zellner. 1985. Estimating gross labor-force flows, Journal of Business and Economic Statistics, 3: 254-283.

Akay, A. and M. Khamis. 2011. The persistence of informality: Evidence from panel data, Institute for the Study of Labor, Discussion paper series, IZA DP No. 6163.

Bargain, O. y P. Kwenda. 2010. Is informality bad? Evidence from Brazil, Mexico and South Africa, Institute for the Study of Labor, Discussion paper series, IZA DP No. 4711.

Blanchard, O.J., P. Diamond, R.E. Hall and K. Murphy. 1990. The cyclical behavior of the gross flows of U.S. workers, Brookings Papers on Economic Activity, pp. 85-155.

Bosch, M. and W. Maloney. 2007. Gross worker flows in the presence of informal labor markets: Evidence from Mexico, 1987-2002, Institute for the Study of Labor, Discussion paper series, IZA DP No. 2864.

Davis, S. J. and J. Haltiwanger. 1998. Measuring gross workers and job flows, in J. Haltiwanger, M. Manser and R. Topel (eds.) Labor Statistics Measurement Issues, NBER, pp. 77-122.

Gong, X. and A. Van Soest. 2002. Wage differentials and mobility in the urban labor market: a panel data analysis for Mexico, Labor Economics, 9: 513-529.

Jones, S.R.G. 1993. Cyclical and seasonal properties of Canadian gross flows of labor, Canadiom Public Policy, 19: 1-17.

--and W.C. Riddell. 1998. Gross flows of labor in Canada and the United

States, Canadiom Public Policy, 24, Supplement 1: CERF/CSLR Conference on the Canada-US Unemployment Rate Gap, S103-S120.

Kaplan, D.S., G. Martinez-Gonzalez and R. Robertson. 2007. Mexican employment dynamics: Evidence from matched firm-worker data, World Bank Policy Research Working Paper, No. 4433.

Kanbur, R. (2009). Conceptualizing informality: Regulation and enforcement, Institute for the Study of Labor, Discussion paper series, IZA DP No. 4186.

Levy, S. (2008). Good Intentions bad outcomes: Social policy, informality and economic growth in Mexico, Brooking Institution Press.

Maloney, W.F. 1999. Does informality imply segmentation in urban labor markets? Evidence from sectorial transitions in Mexico, The World Bank Economic Review, 13: 275-302.

--. 2003. Informality revisited, World Bank Policy Research Working Paper, No. 2965.

Plosser, C.I. 1979. A time series analysis of seasonality in econometric models, in A. Zellner (editor) Seasonal Analysis of Economic Time Series, NBER, pp. 365-410.

Shimer, R. 2005. The cyclical behavior of equilibrium unemployment and vacancies, The American Economic Review, 95: 25-49.

--. 2012. Reassessing the ins and outs of unemployment, Review of Economic Dynamics, 15: 127-148.

Acronym   Original Phrase

EP        Employed Formal flow

El        Employed Informal flow

EFEI      Employed Formal to Employed Informal

EPN       Employed Formal to Not-in-the-labor
          force flow

EPU       Employed Formal to Unemployed flow

EIEF      Employed Informal to Employed Formal

EIN       Employed Informal to Not-in-the-labor
          force flow

EIU       Employed Informal to Unemployed flow

NEP       Not-in-the-labor force to Employed
          Formal flow

NEI       Not-in-the-labor force to Employed
          Informal flow

NU        Not-in-the-labor force to Unemployed

UEF       Unemployed to Employed Formal flow

UEI       Unemployed to Employed Informal flow

UN        Unemployed to Not-in-the-labor force

PEFEI     Probability (transition) of becoming
          Employed Informal after being Employed

PEPN      Probability (transition) of becoming
          Not-in-the-labor force after being
          Employed Formal

PEPU      Probability (transition) of becoming
          Unemployed after being Employed Formal

PEIEF     Probability (transition) of becoming
          Employed Formal after being Employed

PEIN      Probability (transition) of becoming
          Not-in-the-labor force after being
          Employed Informal

PEIU      Probability (transition) of becoming
          Unemployed after being Employed Informal

PNEF      Probability (transition) of becoming
          Employed Formal after being
          Not-in-the-labor force

PNEI      Probability (transition) of becoming
          Employed Informal after being
          Not-in-the-labor force

PNU       Probability (transition) of becoming
          Unemployment after being
          Not-in-the-labor force

PUEF      Probability (transition) of becoming
          Employed Formal after being Unemployed

PUEI      Probability (transition) of becoming
          Employed Informal after being Unemployed

PUN       Probability (transition) of becoming
          Not-in-the-labor force after being

* I appreciate the advices and suggestions from M. Veall, B. Spencer, A. Sweetman, S. Jones, M. Dooley and all the department of Economics of McMaster University for valuable ideas during my research visit of 2011 and 2012,

(1) In longitudinal surveys there is a loss of information due to non-random factors like people dropping out of the sample voluntarily. The difference between adjusted and unadjusted data is that adjusted data contains estimations of the data loss using statistical techniques. Abowd and Zellner (1985) offer a statistical procedure to estimate de loss of information for the American CPS.

(2) Jones and Riddell (1998) give a full explanation of how the unadjusted flows are useful for labor market analysis. Blanchard et al. (1990) also shows how the unadjusted data can render acceptable results.

(3) The subdivision into twelve flows can be turned back into six if we sum up the flows of informal employment (EI) and formal employment (EF) to get total employment (E). But this is not true for the transitions probabilities, which must be re-estimated.

(4) Misclassification is problematic because one single error can produce two bad records, producing spurious transitions. Misclassification might also be caused by the way the interview is carried out within every survey. Interviews by phone or in person might render different levels of misclassification errors. For that reason the CPS in US requires re-interviews to respondents in order to correct for misclassification. The ENOE uses direct personal interviews by qualified personnel, so we assume that misclassification errors are a minor problem related only to the interviewer's skills.

(5) The definition of business cycle here is perhaps closest to the Juglar-fixed investment cycle.

(6) A "Hazard" contains information on the transition probabilities of each flow. For example, PUEF is roughly the probability of becoming employed in the formal sector (EF) after being unemployed (U). The concept of hazard is similar to that of Jones (1993).

(7) Levy (2008) in his chapter of mobility defines young workers as those who are up to 30 years old and concludes that these workers have a higher rate of mobility than older workers. He uses an index called Frequency of Mobility in order to analyze aggregate entries and exits of workers using IMSS data.

(8) All the original phrases for the acronyms used in this section can be found in the appendix.

(9) People who are not working in the formal or informal sectors, and who have decided to stop looking for a job.
Table 1
Mean quarterly flows and hazards 2005-2012

                Total        Young         Old


Employed      44 098 668   14 722 600   29 376 068
EF            33 368 569   11 724 462   21 644 107
El            10 730 098   2 998 137    7 731 961
Unempl.       2 056 991    1 164 161    892 830
Not in the    36 547 084   18 896 878   17 650 207
labor force


EFEI          2 610 799    763 844      1 846 955
EFN           3 304 510    1 293 371    2 Oil 139
EFU           570 189      294 728      275 461
EIEF          2 618 372    798 017      1 820 355
EIN           2 067 727    676 117      1 391 610
EIU           278 459      117 294      161 165
NEF           3 404 730    1 428 181    1 976 549
NEI           2 123 565    721 010      1 402 555
UN            540 942      337 717      203 224
UEF           581 079      309 641      271 437
UEI           288 677      118 404      170 273
UN            495 283      284 686      210 597


PEFEI         0.096        0.089        0.098
PEFN          0.121        0.151        0.107
PEFU          0.021        0.034        0.015
PEIEF         0.246        0.267        0.237
PEIN          0.194        0.226        0.181
PEIU          0.026        0.039        0.021
PNEF          0.122        0.120        0.124
PNEI          0.076        0.060        0.088


PNU           0.019        0.028        0.013
PUEF          0.352        0.353        0.350
PUEI          0.172        0.133        0.216
PUN           0.297        0.323        0.268

                 Male        Female


Employed      27 612 031   16 486 636
EF            21 098 895   12 269 674
El            6 513 136    4 216 962
Unempl.       1 261 898    795 093
Not in the    10 295 249   26 251 836
labor force


EFEI          1 831 657    779 142
EFN           1 180 095    2 124 414
EFU           385 747      184 442
EIEF          1 832 372    785 999
EIN           614 326      1 453 401
EIU           221 646      56 813
NEF           1 559 312    1 845 419
NEI           759 853      1 363 712
UN            236 403      304 539
UEF           382 379      198 700
UEI           228 025      60 652
UN            189 843      305 439


PEFEI         0.107        0.076
PEFN          0.069        0.207
PEFU          0.023        0.018
PEIEF         0.288        0.183
PEIN          0.096        0.337
PEIU          0.035        0.013
PNEF          0.209        0.090
PNEI          0.102        0.067


PNU           0.031        0.015
PUEF          0.386        0.301
PUEI          0.227        0.090
PUN           0.189        0.458

Table 2
Seasonal components of gross flows 2005q2-2012ql

Flow               Total

       [sigma]   [[sigma].   [[sigma].
                  sub.s]      sub.n]

EFEI     122        102         68
EFN      218        185         115
EFU      118        91          75
EIEF     130        99          84
EIN      190        171         84
EIU      73         60          42
NEF      182        149         104
NEI      201        170         108
NU       129        104         76
UEF      116        101         58
UEI      79         67          42
UN       116        99          61

Flow               Young

       [sigma]   [[sigma].   [[sigma].
                  sub.s]      sub.n]

EFEI     40         24          32
EFN      88         68          56
EFU      51         41          30
EIEF     39         21          33
EIN      59         48          34
EIU      28         24          15
NEF      104        78          70
NEI      54         41          35
NU       73         60          41
UEF      57         48          30
UEI      31         26          17
UN       57         48          30

Flow                Old

       [sigma]   [[sigma].   [[sigma].
                  sub.s]      sub.n]

EFEI     103        83          60
EFN      150        127         80
EFU      69         51          47
EIEF     106        78          72
EIN      139        125         61
EIU      47         37          29
NEF      128        115         57
NEI      156        134         81
NU       60         48          37
UEF      63         55          31
UEI      50         41          29
UN       61         52          32

Flow               Male

       [sigma]   [[sigma].   [[sigma].
                  sub.s]      sub.n]

EFEI     84         67          51
EFN      124        104         67
EFU      87         68          54
EIEF     87         64          58
EIN      78         67          39
EIU      59         48          35
NEF      123        108         60
NEI      85         73          43
NU       61         50          35
UEF      84         73          42
UEI      64         52          38
UN       52         45          27

Flow              Female

       [sigma]   [[sigma].   [[sigma].
                  sub.s]      sub.n]

EFEI     50         42          27
EFN      108        85          66
EFU      35         24          25
EIEF     55         41          37
EIN      120        105         58
EIU      15         12           8
NEF      77         48          60
NEI      125        100         75
NU       73         58          44
UEF      35         30          19
UEI      18         16           8
UN       67         55          39

Notes: [sigma] is the standard deviation of the flow,
[[sigma].sub.s] shows the standard deviation of the predicted
values which is considered the seasonal components and
[[sigma].sub.n] is the standard deviation of the residuals which
shows the non-seasonal components.

Table 3
Seasonal components of gross flows 2005q2-2012ql

Flow        Total            Young             Old

       Sea-     Non     Sea-     Non     Sea-     Non
       sonal   seaso-   sonal   seaso-   sonal   seaso-
                nal              nal              nal


EFEI    69       31      35       65      66       34
EFN     72       28      59       41      72       28
EFU     60       40      64       36      54       46
EIEF    58       42      30       70      54       46
EIN     81       19      66       34      81       19
EIU     67       33      73       27      61       39
NEF     67       33      55       45      80       20
NEI     71       29      58       42      73       27
NU      65       35      68       32      63       37
UEF     75       25      73       27      75       25
UEI     72       28      71       29      67       33
UN      73       27      72       28      73       27

Flow        Male            Female

       Sea-     Non     Sea-     Non
       sonal   seaso-   sonal   seaso-
                nal              nal


EFEI    63       37      72       28
EFN     71       29      62       38
EFU     62       38      49       51
EIEF    54       46      55       45
EIN     75       25      77       23
EIU     65       35      71       29
NEF     76       24      39       61
NEI     74       26      64       36
NU      67       33      63       37
UEF     75       25      72       28
UEI     65       35      82       18
UN      74       26      67       33

Notes: The column with seasonal component shows the percentage of
seasonality on the variation of the predicted values of the flows
[[sigma].sup.2.sub.s]/[[sigma].sup.2]. The non/seasonal column
shows the percentage of non/seasonal component on the total
variation of the predicted values of the flows

Table 4
Cyclical components of gross flows 2005-2012

Flow          Total                Young                 Old

EFEI     -1.479 (3.104)      2.882 *** (1.171)      -4.362 (2.878)
EFN      -7.449 (5.086)        1.064 (2.323)      -8.514 *** (3.276)
EFU    -12.315 *** (2.322)   -4.236 *** (1.035)   -8.079 *** (1.394)
EIEF      7.176 (4.962)      4.922 *** (1.724)      2.254 (4.086)
EIN      -2.325 (3.202)        -0.770 (1.222)       -1.556 (3.343)
EIU    -6.638 *** (1.794)    -2.165 *** (0.574)   -4.473 *** (1.268)
NEF      -2.709 (4.008)        0.789 (2.624)        -3.498 (2.132)
NEI    -11.073 *** (3.786)     -1.270 (0.952)     -9.803 *** (3.395)
NU     -10.464 ** (4.536)    -5.691 *** (2.349)   -4.773 ** (2.250)
UEF     -4.766 * (2.877)       -1.095 (1.631)     -3.671 *** (1.373)
UEI    -4.169 *** (1.548)    -1.316 *** (0.508)   -2.854 ** (1.227)
UN     -6.312 *** (2.501)    -3.315 *** (1.186)   -2.997 ** (1.387)

Flow          Mole                Female

EFEI     -0.676 (2.123)       -0.803 (1.378)
EFN    -5.509 *** (2.207)     -1.940 (3.653)
EFU    -9.223 *** (1.367)   -3.092 *** (1.077)
EIEF    4.994 * (2.868)       2.182 (2.305)
EIN    -3.591 ** (1.623)      1.266 (1.997)
EIU    -5.825 *** (1.322)     -0.812 (0.541)
NEF     -4.427 * (2.275)      1.719 (2.725)
NEI    -6.714 *** (1.116)     -4.359 (3.001)
NU     -4.432 *** (1.887)   -6.031 ** (3.052)
UEF    -4.095 ** (2.068)      -0.671 (0.897)
UEI    -3.489 *** (1.392)   -0.680 *** (0.276)
UN     -2.809 *** (0.973)   -3.503 ** (1.748)

Notes: Every entry corresponds to the coefficient of the real GDP
growth rate of Mexico taken from regressions where the gross
flows are regressed on this growth rate, quarterly dummies and a
cubic time trend. The ***, ** and * symbols represent
coefficients that are statistically significant different than
zero at 2%, 5% and 10%, respectively. Standard errors are in

Table 5
Cyclical components of hazards 2005-2012

Transitions         Total                Young

PEFEI           0.005 (0.011)       0.018 * (0.011)
PEPN            -0.015 (0.013)       -0.015 (0.019)
PEFU          -0.044 *** (0.009)   -0.057 *** (0.014)
PEIEF          0.072 ** (0.033)     0.098 * (0.053)
PEIN            -0.026 (0.024)     -0.092 ** (0.044)
PEIU          -0.065 *** (0.017)   -0.088 *** (0.020)
PNEF           0.020 * (0.012)       0.031 (0.021)
PNEI           -0.022 * (0.013)      0.002 (0.007)
PNU           -0.033 ** (0.016)    -0.042 ** (0.019)
PUEF          0.219 *** (0.043)    0.302 *** (0.060)
PUEI            -0.029 (0.034)       -0.012 (0.050)
PUN             0.029 (0.040)        -0.006 (0.048)

Transitions          Old                  Male

PEFEI          -0.00001 (0.014)      0.003 (0.010)
PEPN            -0.021 (0.013)     -0.029 *** (0.010)
PEFU          -0.041 *** (0.008)   -0.054 *** (0.009)
PEIEF           0.059 (0.040)      0.084 *** (0.030)
PEIN            -0.005 (0.031)     -0.059 *** (0.024)
PEIU          -0.058 *** (0.016)   -0.094 *** (0.021)
PNEF            0.012 (0.010)      0.069 *** (0.025)
PNEI          -0.039 ** (0.020)    -0.029 *** (0.011)
PNU           -0.027 ** (0.014)     -0.042 * (0.023)
PUEF           0.129 ** (0.055)    0.243 *** (0.065)
PUEI            -0.022 (0.060)       0.010 (0.051)
PUN             0.048 (0.051)        0.010 (0.039)

Transitions        Female

PEFEI           0.006 (0.014)
PEPN            0.019 (0.031)
PEFU           -0.028 (0.011)
PEIEF           0.054 (0.043)
PEIN            0.023 (0.026)
PEIU           -0.020 (0.013)
PNEF            0.019 (0.012)
PNEI           -0.014 (0.015)
PNU           -0.028 * (0.015)
PUEF          0.217 *** (0.051)
PUEI           -0.021 (0.022)
PUN            -0.066 (0.057)

Notes: Every entry corresponds to the coefficient of the real GDP
growth rate of Mexico taken from regressions where the hazards
are regressed on this growth rate, quarterly dummies and a cubic
time trend. The coefficients from the hazards regressions as well
as the standard deviations are multiplied by 100. The ***, ** and
* symbols represent coefficients that are statistically
significant different than zero at 2%, 5% and 10%, respectively.
Standard errors are in parenthesis.
COPYRIGHT 2013 El Colegio de Mexico, A.C.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2013 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:articulo en ingles
Author:Gallardo del Angel, Roberto
Publication:Estudios Economicos
Date:Jul 1, 2013
Previous Article:El rezago social en areas metropolitanas de Mexico.

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |