Thesis of the MSc in Economics in Universitat de Barcelona. Title: Formal-informal sector transitions in the Ecuadorian labor market

Thesis of the MSc in Economics in Universitat de Barcelona Title: Formal-informal sector transitions in the Ecuadorian labor market Author: Adriana Vega Supervisor: Raúl Ramos Abstract: The study offers an analysis of the transitions between the formal and informal labor markets using longitudinal data for Ecuador. The transition matrix provides evidence of the movements among the sectors, where flows between formal and informal labor markets present similar volumes. The differential in earnings can justify in some extend movements from the informal to formal sector. The results from multinomial logit models show that the level of education, years of experience and differential in earnings have a significant effect on worker s transitions between these two sectors. These results suggest the presence of an integrated point of view of the relationship between formal and informal labor markets and not the traditional segmented view. Key words: informal sector, labor mobility, Ecuador 1

1. INTRODUCTION The widespread informality is a characteristic of developing and transition economies 1. The informal economic activities include a diverse range of people in these countries and a variety of effects on them. This phenomenon has been persistent and remains as a major challenge in many countries seeking to reduce it 2. Jutting and Laiglesia (2009) study shows that informal employment has existed since it was firstly defined in the middle 1970s. From the middle 1970s to 2000s, informal employment has presented an upward oriented trend in the regions shown in Figure 1. On average, in these last 30 years, informal employment had accounted for more than 47% of total non-agricultural employment in developing regions and around 24% in transition economies. Specifically, Latin America presented on average more than 50% of informality. Figure 1. Share of informal employment in non-agricultural employment Source: Jütting and Laiglesia (2009), Is Informal Normal? Toward more and better jobs in developing countries. Note: the box chart shows the range of informal employment as share of total non-agricultural employment by region, based on the latest available observation for each country. The edges of each box correspond to the upper and lower quartiles, with the vertical line inside indicating the median value for each region. The whiskers outside the box show the upper and lower adjacent values of the data. The outlier value in Latin America is Haiti. 1 Independently the way of measuring informality, this phenomenon is high in Latin America (Perry et al., 2007). Informality has increased over time in Sub-Saharan Africa and Asia (Jütting, Parlevliet and Xenogiani, 2008). 2 According to Bacheta et al., (2009), based on broad definitions of informality, countries in Africa, Asia and Latin America present persistent informality rates. 2

The term informality varies among researchers, but most of the time is associated with negative things: unprotected workers, tax evasion, illegal activities, low productivity, low investment rates, etc. However, in the literature different schools of thoughts have emerged dealing with the causes of informal employment. The main question in this sense is whether individuals or firms voluntary exit the formal sector or are omitted from it. There are two dominant schools of thoughts discussed in the World Bank publication: Informality: Exit and Exclusion (Perry et al., 2007). The exclusion view, which coincides with the traditional way of thinking, points to a segmented market where people choose informality as their only available option of occupation (Soto, 2000). In the same sense, informal workers would prefer a formal job because higher wages are paid and labor protection is offered. In contrast, the voluntary view of the informal labor market follows the opposite direction, where the labor market offers jobs with various characteristics from where workers can choose voluntary to join into informal work. Maloney (2004) in this line gives evidence for Latina America and settles that the informal sector is an unregulated micro-entrepreneurial sector, which does not correspond to a disadvantaged segment of labor market. Furthermore, Bosch and Maloney (2005) suggest that a considerable part of the informal sector, particularly the selfemployed correspond to voluntary entry to this segment of the labor market. Accordingly, there are diverse alternatives of informal employment that can offer desired qualities, such as independence or possibilities for training. In this scope, the aim of this study is to examine the dynamics in the labor market across the different labor status and observe the characteristics that determine the probability of staying or moving across the formal and informal sectors in Ecuador. Hence, the objective of this analysis is to identify if the characteristics of the labor market suggest that workers from both formal and informal sectors form an integrated market or present a traditional segmented market. 3

The empirical exploration is based on the Ecuadorian National Survey of Employment and Unemployment (ENEMDU) for the period of 2011-2012. The investigation consists in examining the flows among the labor sectors: formal, informal, unemployed or out of the labor force. For this purpose, a transition matrix is constructed to schematize the dynamics of the market and interactions among the sectors. Additionally, to have information about the differentials in the real earnings during the transitions, we performed a mean comparison test between the initial and final wages of workers. Finally and since the transition matrix does not consider observable characteristics of the workers, a multinomial logit analysis was implemented to identify these characteristics and determine the probabilities of choosing the different labor sectors. The study is organized as follows. The next section discusses a brief summary of empirical literature on the dynamics in the formal and informal labor market. Section 3 describes the data and definition of informality and other main variables used in the study. Section 4 presents the econometric methodology and models, and results are reported in section 5. Finally, section 6 discusses the main findings and remarks. 2. LITERATURE REVIEW AND CONTRIBUTION As previously mentioned, there are two perspectives used to explain informality regarding segmentation in the labor market. The first alternative implies that workers would prefer formal wage employments, but there is a restriction and it is not always possible to find one, so the only option would be to enter in the informal sector. In this sense, Harris and Todaro (1970) set in their model a minimum wage above the equilibrium wage resulting in a limitation of formal jobs leading us to segmentation in the labor market. In the second stream, Maloney (1999) perceives the formal and informal sector as an integrated market. In other words, workers choose among the different job offers their occupations according to their preferences, abilities and needs. Thus, workers who prefer informal employment to formal one are due to desirable 4

characteristics that this part of the labor market offers. These two schools of thoughts examine in different ways the diverse flows in the labor market and the involving sectorial wage differentials. If we consider a segmented market, flows from informal to formal jobs should be larger than the opposite movement. Comparing with an integrated market, the flows between formal and informal jobs should be in both directions and of a similar volume. (Fields, 2009). Recent evidence suggests that emerging countries show important dynamics in the labor market. Moving of workers across jobs, or changes from unemployment to employment, and even the entering and exiting the labor market indicates the mobility of the sector. In this scope, Maloney (1999) analyzed the workers transitions between sectors using panel data from Mexico, finding that both earnings differentials and patterns of mobility in the labor market implied that a big proportion of the informal sector is a desirable destination and that the different types of works, formal and informal, are well integrated. Duryea et al., (2006) examined workers flows and the possible associated changes in earnings on three countries in Latin America: Argentina, Mexico and Venezuela. The authors found that workers who moved from formal wage to informal wage experienced on average a decline in wages, while the inverse movement produced the opposite effect. As well, Cea et all., (2008) using a panel data analysis, provided us with evidence for Chile. The results reveled that the dynamic of the labor status of the individuals have an important permanence. On the other hand, the study suggests that age, schooling, and no labor income are significant characteristics that determine the probability of being in a certain labor status. In another study for Argentina, Jimenez (2011) found evidence for the segmentation of the formal sector of the labor market. It is important to remark that workers from this group were confronted with unfavorable job conditions. In the case of transition economies, Lehman and Pignatti (2008) in their panel data analysis in Ukraine found the presence of segmentation in the labor market and evidence of earnings premiums associated with voluntary moves. Slonimczyk and Gimpelson (2013) using a multinomial logit model and allowing for individual 5

heterogeneity in preferences, revealed the existence of an integrated labor market in Russia. In general, informality is a prominent feature of transition and emerging economies, as is the mobility of workers across sectors in the labor market. These topics have been studied in many countries and over many years and the evidence provide us with diverse results. Previous studies of the Ecuadorian labor market are mainly focused in cross section analysis, which allows exploring some issues regarding the different labor status without giving rise to encounter the movements between the different sectors of this market. The understanding of these transitions is required to determine specific and correct economic and labor policies for the country. 3. DATA The source of the data used in this study is the National Survey of Employment and Unemployment (ENEMDU), conducted by the Ecuadorian National Institute of Statistics and Census (INEC). The ENEMDU is a household rotating panel survey. The panel does not follow individuals continuously, but it is constructed from four reports, spread over two consecutive years. Households are interviewed for two consecutive quarters then, in the next two consecutive quarters the interviewed households are substituted by a new sample, and finally the first group of households returns to the sample for the two last quarters. The analyzed panel is 2011: 4-2012: 4. The weights used are the expansion factors specified by the INEC. In the analysis we consider workers who are aged 15 years or older. We consider four different status in the labor market: (1) formal sector, (2) informal sector, (3) unemployed and (4) out of the labor force. Individuals are classified as unemployed if they did not work in the reference week but had 6

searched for a job. Out of the labor force individuals are those who do not work, or seek some type of employment. We can find diverse definitions of informality. There are various criteria regarding the level of production units or firms, or the level of workers. A frequently criteria to define the informal sector at the level of firms are the size and the registration status. In this sense, the International Labor Organization (ILO) (2002a) defines as an informal worker all the individuals who work in enterprises with less than five paid employees, are not registered, and are engaged in non-agricultural activities. As mentioned before, the informal sector can also be defined according to the worker level, basically established on the employment relationship. Under this principle, ILO (2002b) considers as an informal employment both self employment and wage employment, who are not recognized, regulated or protected by existing legal or regulatory labor benefits. Two definitions for the informal sector are used in this study. The first one focuses on the firm characteristics: salaried workers employed by small establishments with less than ten employees and which are not registered, plus all independent and self-employed workers. The second definition focuses on social security coverage. Workers are considered as informal if they are not covered by social security. Figure 2 plots the informality rates based on the two above stated definitions during the observation period 2011 and 2012. As can be seen, the country presents a share of informal workers between 32% and 38%, respectively, of the labor force in this period. In general, in Ecuador the percentage of informality can be considered high, but it is among the common fraction of workers, which belong to this segment of the labor market in most of Latin American countries. 3 Since both informality definitions present similar, we are going to use the first definition for the rest of the analysis. 3 Between 1970s and 2000s, on average, informal employment accounts for more than 47 percent of total non-agricultural employment in West Asia and in North Africa, and more than 70 percent in sub-saharan Africa, more than 50 percent in Latin America, nearly 70 percent in South and Southeast Asia and 24 percent in transition economies (Jütting and Laiglesia, 2009). 7

Figure 2.. Share of informal employment in Ecuador 2011-2012 40 35 30 25 20 15 Informal sector 1 Informal sector 2 10 5 0 2011 2012 Source: Author s calculations based on the Ecuador National Survey of Employment and Unemployment, 2011-2012. Note: Informal sector 1 is form by salaried workers employed by small establishments with less than ten employees and are not registered, plus all independent and self-employed workers. Informal sector 2 is form by workers who are not covered by social security. Given the four labor status, we defined ten kind of transitions, which represent the flow in the labor market between sectors during the analyzed period. Additionally, we included four permanencies, which characterize the permanency of labor status compared to the previous period. Tables A1 and A2 in the appendix present the mean values for the sample using the two informality definitions across the different transitions and permanencies, respectively, for age, level of schooling, work experience, and the initial real wage. The initial and final real wage denotes the monthly real wage of workers during December 2011 and December 2012, respectively. Thus, to construct this variable we considered exclusively the earnings of the main job as the only source of labor income. Finally the differential in earnings corresponds to the difference between the final and initial real wage whether the worker experienced a transition or permanency among the labor sectors. 8

4. METHODOLOGY In this section we describe the econometric techniques implemented in the study in order to provide empirical evidence of the patterns of mobility between the mentioned sectors of the Ecuadorian labor market. To identify the directions and volume of the diverse flows in the labor market and the implicit differential in earnings, we used three methods: transition matrices, the mean sample of the real wage differentials from transitions between sectors and a multinomial logit analysis of movements between labor sectors. Transition matrices The transition matrices allow us to determine the flows of workers between the considered labor sectors by calculating the conditional probability of finding a worker in sector j at the end of the period, given that the worker began in sector i, Pij. The sum of each row of the transition matrix is equal to 100% and the totals at the end of the columns and rows represent the share of workers in each category at the end of the period Pi and Pj. The components in the main diagonal reveal the share of workers who staid in the same labor category at the end of the period. The information of the transition matrices give us a first intuition of the different movements of workers among the established sectors. Since this methodology is only descriptive, we will also implement a multinomial logit analysis in order to determine which workers characteristics affect the probability of a worker choosing sector j relative to the probability of staying in sector i. Real wage differentials from transitions between labor sectors In order to see the changes of the real wage during the transitions of the workers across the labor market sectors; we employed the sample mean comparison test between the initial and final real wage. With the sample mean we were able to identify the tendency of the sample data and significance of the different movements and demonstrate if workers experienced an improvement or a decline 9

in their earnings. Additionally, the variations in the real wages can be a possible reason for the movements. Multinomial Logit Analysis of transitions between labor sectors The multinomial logit model is a probabilistic discrete model, which can explain transitions and permanencies in the different sectors of the labor market. We can find the effect for each workers characteristics on the probability of choosing a sector. We model flows among four different labor market states: formal sector (j=1), informal sector (j=2), unemployment (j=3), and out of the labor force (j=4). We use the standard exponential form for the multinomial logit analysis: where the vector measures the degree to which an increase in worker characteristics increase the probability of a worker going to sector j relative to the probability of staying in sector i. The workers characteristics are age, gender, marital status, level of education, years of experience, regions, and the logarithms of the differential in earnings. We can calculate the probability of making a transition where the explicative variables determine the increase or decrease of these probabilities. 5. RESULTS Transition matrices In this first exploration with transition matrices we describe labor mobility by calculating the conditional probabilities of finding a worker in status j, at the end of the period, conditional on being in status i at the beginning. This yields to a four by four annual matrix for the analyzed period December 2011 to December 2012 (table 1). Since the purpose of this investigation is to identify the patterns of 10

Table 1. Worker transitions among sectors of the labor market in Ecuador. 2011-2012 Transition probabilities: Pij (percent) Initial sector Formal sector Informal sector Final sector Unemployed Out of labor force Formal sector 79 12 2 6 100 25 Informal sector 12 72 2 15 100 35 Unemployed 23 36 19 22 100 3 Out of labor force 4 12 3 81 100 36 Total Pj 26 34 3 37 Source: Author s calculations based on the Ecuador National Survey of Employment and Unemployment, 2011-2012. Note: P i. is the relative size of a sector at the initial period; Pj is the relative size of a sector at the final period. Pi 11

workers transitions that moved from and to the formal and informal sector, we are going to focus mainly in the transition from the formal sector to each of the other labor status and from the informal sector to the other sectors. Additionally, in the matrix we can observe some facts about the dynamics of the market and the interactions among all the defined sectors. Regarding the movements from the employees of the formal sector, we found a 79% of permanency in this sector. An interesting result is that the largest outflow of this sector is into the informal category. This kind of flows revel a decrease of job conditions for workers, in terms of formality. This disadvantage situation of labor transition could be explained by the rationed out of formal employment opportunities (Fields, 1975; Perry et al., 2007) or voluntary decisions of individuals or firms due to inefficiencies in formal sector protections, and low levels of labor productivity (Maloney, 1999). As well, workers of formal firms who leave their jobs, can move to unemployment (2%) or leave the labor activity (6%). Turning to the moves from informal employees, we see a similar picture as moves from the formal relationship. First, the probability of staying as informal workers is high, meaning that 72% of the workers remain in this part of the market. If we look at the outflows from the informal sector, we find that 12% of workers raised their job quality by changing to the formal sector. This sort of moves can be understood as workers improvements in their job conditions and furthermore; informal workers are typically subject to lower remuneration than similar workers in the formal sector (Günther and Launov, 2006). The probability of moving from an informal job to a formal one is higher than the probability of moving to unemployment. The flow from informal jobs led out of the labor force corresponds to 15%. Concerning the movements from unemployment, we observed a high mobility, since 19% of individuals remained unemployed during the observation period. In this sense, most of unemployed migrated to the informal sector, followed by the formal sector and finally out of the labor force, respectively 36%, 23%, and 22%. 12

Most of people who found a job were in the informal sector; this result can be attributed to an insufficient human capital or other individual characteristics and preferences of the workers. Traditionally, if we consider unemployment as the lowest state in the labor market, people from this segment who had restrictions and difficulties in getting into formal employments, will probably be induced to rapidly accept employments with lower labor conditions and earnings. Jütting, et al., (2008) argue that it can also be the case that working in the informal sector gives some extra advantages to individuals compared to the formal sector. After all, informality cannot be considered the last resort of a worker. Workers may voluntary choose to work in the informal sector because in there they would have the chance to accumulate experience or training and preparation in the case of being a low skilled young worker or unskilled older individual; and also they would find greater flexibility and autonomy. In the World Bank publication (2012) Ecuador: Las caras de la informalidad stresses that the owners of small companies value the independence and flexibility of having their own business and the main reason to have these small business is because they want to achieve autonomy. When we refer to people out of the labor force we observe a high permanency in this state. A high proportion of the movements of this group are to the informal sector (12%). On the other hand just 4% respectively 3% moved to the formal sector and into unemployed. Again the flows of people pointed to the informal sector are higher than the ones to the formal sector. In this context, we can see that there are some attractive characteristics that the informal sector offers to the workers. Real wage differentials The characteristics of each sector of the labor market and as well the specificities of a work determine different wages among workers. These diverse work structures make it difficult to have a specific magnitude and sign of how the differential in wages could be. According to Cornelißen and Hübler (2007), important determinants of earning levels are the unobserved characteristics of 13

workers and firm heterogeneity. Furthermore, Tansel and Oznur (2012) test if informal workers get lower remunerations compared to similar workers in the formal sector. The authors found that unobserved fixed effects joint to observable workers characteristics explained the differentials in payments between formal and informal employment. Thus the unobserved characteristics such as personal abilities of production, character traits and quality of management could affect the workers productivity and therefore the differential in wages. The effect of firms heterogeneity on wage differentials can be measured by the different needs of firms to motivate their workforce, insure workers against market fluctuation or compensate the workforce for undesirable working conditions. In general, informal workers should be compensated with higher wages because of the lower labor conditions they face. But on the other hand, informal workers should be penalized for not being registered and evading taxation. Using data from ENENMDU, table 2 presents the absolute sample mean difference in real earnings of the individuals in their transitions across sectors. As we said before, we are going to concentrate in the movements from the formal and informal sector to the other labor states, and paying special interest in the fluctuation from the formal to the informal sector and the reverse movement. The movements from the formal sector into informal employment lead in a significant reduction in remuneration. By contrast, movements from the informal to formal sector increase the salaries. Considering the remaining transitions from formal or informal employment to unemployment or out of the labor force, there are obvious and significant declines in remuneration. The outcomes presented in table 2 do not capture any unobserved workers characteristic or firm heterogeneity that can affect the resulting earning differentials in the transitions across the labor sectors. Even though, the results let us suppose that movements from the informal to the formal sector are due in part to the increase in wages for the transition workers, including as well the improvement in the labor conditions that they can experience. Taking into account the opposite movement, wage differentials cannot be considered as an explanation of this transition, because moving from the formal to the informal sector reflects a decline in remuneration. 14

Thus, there should be other explanations for this kind of flows, which are not specifically earning improvements. The multinomial logit analysis will provide us with additional information of the individual s behavior for each of the four labor sectors and their alternative options to enter in a new sector or to stay in the same one. Table 2. Real wage differentials in the workers transition between sectors, Ecuador, 2011-2012 Workers transitions between sectors Mean Differential From formal sector to Informal sector -58.6 *** (19.02) Unemployed -283.2 *** (53,77) Out of labor force -379.24 *** (37,69) Workers transitions between sectors Mean Differential From informal sector to Formal sector 79.94 *** (19.78) Unemployed -118.4 *** (19.88) Out of labor force -103.1 *** (6.13) Source: Author s calculations based on the Ecuador National Survey of Employment and Unemployment, 2011-2012. Note: Standard errors in parenthesis; *p<0.1, **p<0.05, ***p<0.01 Multinomial Logit Analysis The multinomial logit analysis of movements between labor sectors allows to determine in statistically terms, if workers are more or less likely to move to another sector compared to the initial sector, according to the their specific characteristics (table 3). 15

Table 3. Multinomial Logit Analysis of transitions between sectors, Ecuador, 2011-2012 Workers transitions between sectors Constant Age Female (d) Married (d) Schooling Experience Regions (d) Coast Centre South Log. Diff. earnings From the formal sector to Informal sector -0.538 0.007-0.002 0.059-0.231-0.028 0.258 0.316 0.249-0.111 (1.50) (1.09) (0.01) (0.42) (6.90)*** (3.20)*** (1.37) (1.31) (1.09) (2.86)*** Unemployed -3.290-0.019-0.006-1.048-0.279-0.077 1.448 1.538 1.157-0.779 (2.82)*** (0.94) (0.02) (1.96) (2.68)*** (1.92) (1.87) (1.71) (1.26) (10.16)*** Out of the labor force -1.332 0.004 0.854-0.524-0.457-0.006-0.560-0.660 0.120-0.898 (2.13)** (0.35) (3.55)*** (1.89) (7.19)*** (0.45) (1.81) (1.59) (0.31) (17.72)*** From the informal sector to Formal sector -1.542-0.022-0.392 0.144 0.223-0.015-0.585-0.043-0.311 0.067 (4.98)*** (4.22)*** (3.24)*** (1.14) (7.18)*** (2.10)** (3.81)*** (0.23) (1.61) (2.31)** Unemployed -3.304-0.045-0.240-0.856 0.171-0.022 0.226 0.215 0.697-0.668 (4.54)*** (3.55)*** (0.86) (2.45)** (2.27)** (1.12) (0.56) (0.42) (1.41) (12.06)*** Out of the labor force -3.168 0.005 1.133-0.478-0.035-0.016 0.231-0.273 0.300-0.726 (7.90)*** (0.91) (7.76)*** (3.23)*** (0.91) (2.49)** (1.17) (1.06) (1.20) (24.01)*** Source: Author s calculations based on the Ecuador National Survey of Employment and Unemployment, 2011-2012. Note: The coefficients reflect how the different worker characteristics and the percentage change in the real wage affect the probability of moving from the initial sector to the final sector relative to the probability of staying in the initial sector. The informal sector is formed by salaried workers employed by small establishments with less than ten employees and are not registered, plus all independent and self-employed workers. Z statistics are in parenthesis; *p<0.1, **p<0.05, ***p<0.01. (d) Dummy variables; default categories are: male, other civil status except married and north region. 16

The main objective of this study is to obtain an overview of the labor force dynamics in Ecuador. Thus the informal sector comprises the largest source of employment. As mentioned before, this segment of the labor force not necessarily corresponds to a lower job status compared to the formal sector, since it is observed that workers transitions from formal employment into informal one are significant in the labor market. The regression results show that workers become less likely to leave formal status for informal status, as their level of education increases. In average a worker who did the transition in this direction has 13 years of studies. Concerning the experience of the worker, the probability of moving into informal sector decreases as the experience increases. It is important to stress that the mean of the years of experience is 8, so this is a considerable number of years of training. This figure can be related to people who started a business. Aroca and Maloney (1998) found that the informal self-employment sector is a desirable destination for workers, but it requires accumulating financial and human capital. Thus the mean number of years of experience suggests that workers first work for some years accumulating savings and knowledge, which afterwards can be used to start a business. As expected, the logit results show that the probability of moving from the formal sector to the informal sector declines as the percentage difference of the real wage between the initial and final sector rises. The sample mean difference in real earnings of the individuals who moved in this direction had a reduction in their remuneration. Thus if this difference increases workers would be less likely to move into the informal employment. Paralleling, we follow the transitions between the formal employment and the two sectors of unemployment: unemployment and out of the labor force. In both cases, the probability of making these transitions decreases as the level of education increases. Moving to the second logit analysis, the flows from the informal sector to the formal sector present some interesting patterns. In the first place, better educated workers push up the mean years of schooling in the formal sector, so comparing the mean years of schooling of individuals who persisted as informal and individuals who moved to the formal status is 10 and 13, respectively (table A1 17

and A2). The logit results suggest that better educated people are more likely to enter the formal sector employment. This association may propose that while workers are increasing their level of education they start in the informal sector as an option of employment and after raising their level of instruction and skills they will try to get better labor conditions. The probability of moving from the informal sector to the formal sector is associated with the percentage difference in the real wage. The mean differences in the real wage of the workers who moved to the formal sector is positive and significant, thus if this differences increases the probability of moving to this sector will also increase. The age is a factor that affects in a negative way the probability of passing from informal to formal employment. Finally the model picks up some gender dissimilarities. Females are less likely to flow in this direction compared to males. The main finding regarding the transition from the informal sector to unemployed and out of the labor force are related with the percentage difference of real wage, which influences in a negative way these two transitions. We have already noted that the coefficients from the multinomial logit are interpreted relative to the base line, which is to remain in the same sector formal or informal respectively. Another way to evaluate the effect and magnitude of covariates is to examine the marginal effect of changing their values on the probability of observing a transition or permanency (table A3) 4. Additionally, using the two multinomial logit models and in order to have a better idea of the economic significance of the effect of observable characteristics of the workers in the different transitions across the labor sectors, we run simulation exercises assigning specific workers characteristics. 5 We first analyze the effect of years of experience in the probabilities of moving from the formal sector to the other segments and the probability of permanency in the very same sector. For this 4 Margins estimate the average effect over the estimation sample. In the case of factor variables, the margin is the discrete change from the base level. 5 Simulations results for other characteristics are omitted to save space. 18

exploration we fixed the age to 38 years and secondary degree 6 and we then compare the transition or permanency probabilities for both men and women (figure A1: figure A4). Both for females and males, more years of experience led to a significant increase in the probabilities of retention in formal employment. The transition probabilities show that more years of experience reduce exit rates to the informal and unemployed sectors. We also analyzed the effect of the different levels of education, for the second logit model. In this case we again fixed the age to 38 years and assigned 10 years of experience (figure A5: figure A8). Some interesting results are that while the level of education increases, the probability of passing from the informal to the formal sector increases as well. Specifically, having a college degree leads to a 11 (8) percentage point increase in the fraction of men (women) changing to a formal job compared to having a secondary degree. Given the previous characteristics of the workers the probabilities of staying an informal employee is high, reaching 80% (74%) for men (women) respectively, with non-education. However, this probability decreases as the individuals acquire more years of instruction. 6. FINAL REMARKS Informality is an important phenomenon, which comprises a significant share of the labor force employment in many developing and transition economies. The study provides an overview of the dynamics of the formal and informal sectors and some specific patterns of the transitions within the labor sectors. We specify a transition matrix and a multinomial logit model to identify the movements across the sectors and the effect of each worker s characteristics on the probability of choosing a sector. Moreover, we analyzed the differential in earnings of the workers during the transition to a different labor sector or permanency in the same one. 6 Secondary education corresponds to high school education. 19

The results observed in the transition matrix suggest a quite high mobility not only in and out of the labor market but also across the sectors. Nonetheless, these results can be considered in the normal percent range of movements found in similar analysis of labor mobility in Latin America. The flows in the labor sectors regarding employed people suggest that individuals are searching for job opportunities in both formal and informal sectors, since the transitions in the formal and informal employment in both directions are similar among each other. The analysis also offers additional information about the differential in earnings at the mean of the diverse transitions. Workers who moved from formal to informal employment experienced negative consequences in their earnings. By contrast, the reverse movement improves worker s remuneration. This factor suggests that one of the reasons for movements from informal employment to a formal one is the improvement in earnings. The multinomial logit analysis was applied since the transition matrices and the mean differential in earnings do not consider observable characteristics of the workers that can affect their choice in which sector to work. This approach indicated that education, years of experience, and other characteristics influence the selection of employment and therefore the transitions or permanencies in the different sectors. The findings presented in the study sustain an integrated labor market in terms of formality, because of the interaction between the formal and informal sectors. The patterns of mobility imply that informal employment should be viewed as a desirable destination as it is the formal sector. Hence, the two ways of employments present various advantages and costs, which are depending on the individual s preferences and abilities. Finally, the voluntary element (so the voluntary change from formal to informal employment) that the labor market might present leads to important policy implications that vary from the traditional point of view. If the informal workers are part of this segment involuntary, then policy makers have to focus on aspects 20

such as the rigidities of wages and be aware of the formality of the enterprises in terms of social protection. But in the case individuals choose voluntary to become informal, then policy makers need to take into account the inefficiencies in the labor codes and the low levels of formal sector productivity. 7 7 See Maloney (2004). 21

7. REFERENCES Aroca, P, and Maloney, W. (1998). Logit Analysis in a Rotating Panel Context and an Application to Self-Employment Decisions. Latin America and the Caribbean Region, Poverty Reduction and Economic Management Unit, World Bank, Washington, D.C. Processed. Bacheta, M., Ernest, E. and Bustamante, J. (2009), Globalization and informal jobs in developing countries, A joint study of the International Labor Office and the Secretariat of the World Trade Organization. Bosch, M. and Maloney, W. (2005); Labor Market Dynamics in Developing Countries. Comparative Analysis using Continuous Time Markov Processes, Policy Research Working Paper 3583. The World Bank. Cea, S. and Contreras, M. (2008), Transiciones Laborales: Evidencia para datos de panel. Tesis de Economía, Universidad de Chile. Cornelißen, T. and Hübler, O. (2007), Unobserved Individual and Firm Heterogeneity in Wage and Tenure Functions: Evidence from German Linked Employer-Employee Data, IZA DP No. 2741. Duryea, S., Marquez, G., Pagés, C. and Scarpetta S. (2006), For Better or for Worse? Job and Earnings Mobility in Nine Middle and Low-Income Countries, Brookings Trade Forum, Global Labor Markets, pp. 187-203. Fields, G. S. (1975), Rural-Urban Migration, Urban Unemployment and Underemployment, and Job-Search Activity in LDC s, Journal of Development Economics, 2, 165 187. Fields, G. (2009), Segmented Labor Market Models In Developing Countries. In Ross, D. and Kinkaid, H., editors, The Oxford Handbook of Philosophy of Economics, pp. 476 510. Günther, I. and A. Launov (2006), Competitive and Segmented Informal Labor Markets, IZA Discussion Papers No. 2349. Harris, J. R. and Todaro, M. P. (1970), Migration, Unemployment and Development: A Two- Sector Analysis. The American Economic Review, 60(1), pp. 126 142. International Labor Office (2002a). Women and men in the informal economy: A statistical picture, Geneva, International Labor Office. International Labor Organization (2002b). Decent work and the informal economy. International Labor 408 Conference, 90th Session. Geneva: International 409 Labor Office. Jimenez, M. (2011), La Economía Informal y el Mercado Laboral en la Argentina: Un Análisis desde la Perspectiva del Trabajo Decente, Centro de Estudios Distributivos, Laborales y Sociales, Documento de trabajo Nro. 116 22

Informal Employment Re- Jütting, J., Parlevliet, J., and Xenogiani, T. (2008), loaded, IDS Bulletin, 39(2), pp. 28 36. Jütting, J., and Laiglesia, J. (2009), Is Informal Normal? Towards more and better Jobs in developing countries, Organization for Economic Co-operation and Development (OECD). Lehmann, H., and Pignatti, N. (2008), Informal Employment Relationships and Labor Market Segmentation in Transition Economies: Evidence from Ukraine, IZA Discussion Paper No. 3269. Maloney, W., (1999), Does Informality Imply Segmentation in Urban Labor Markets? Evidence from Sectorial Transitions in Mexico, The World Bank Economic Review, Vol. 13, No.2, pp. 275-302. Maloney, W., (2004), Informality Revisited, World Development, vol. 32(7), pp. 1159-1178. Perry, G., Maloney, W., Arias, O., Fajnzylber, P. Mason, A. and Saavedra-Chanduvi J. (2007). Informality: Exit and Exclusion, The World Bank. Slonimczyk, F., and Gimpelson, V. (2013), Informality and Mobility: Evidence from Russian Panel Data, IZA Discussion Paper No. 7703. Soto, H. DE. (2000), The Mystery of Capital: Why Capitalism Triumphs in the West and Fails Everywhere Else, Basic Books, New York, NY. Tansel, A., and Oznur, E. (2012), The Formal/Informal Employment Earnings Gap: Evidence from Turkey, IZA DP No. 6556. World Bank. (2012), Ecuador: Las caras de la Informalidad, Washington, DC., Inform No. 67808-EC. 23

APPENDIX Table A1. Summary statistics using the first definition of the informal sector, Ecuador, 2011-2012 Mean Workers transitions between sectors Number of observations Age (years) Schooling (years) Experience (years) Initial real wage (USD) From the formal sector to Formal sector 1841 39.2 15.0 9.75 466.2 (0.30) (0.09) (0.24) (11.78) Informal sector 292 39.1 13.2 7.79 324.5 (0.75) (0.24) (0.53) (18.87) Unemployed 38 29.7 14.6 3.29 338.3 (1.57) (0.64) (0.82) (48.09) Out of the labor force 156 40.9 13.6 10.7 375.1 (1.61) (0.35) (1.09) (37.31) From the informal sector to Formal sector 396 38.4 13.1 8.77 250.6 (0.71) (0.21) (0.51) (14.78) Informal sector 2184 45.4 10.9 12.5 207.4 (0.30) (0.09) (0.25) (5.72) Unemployed 72 31.3 13.4 5.9 145.0 (1.38) (0.45) (1.11) (16.12) Out of the labor force 456 45.3 10.7 9.81 99.7 (0.92) (0.21) (0.63) (5.96) Source: Author s calculations based on the Ecuador National Survey of Employment and Unemployment, 2011-2012. Note: The informal sector is formed by salaried workers employed by small establishments with less than ten employees and are not registered, plus all independent and self-employed workers. Standard errors in parenthesis. 24

Table A2. Summary statistics using the second definition of the informal sector, Ecuador, 2011-2012 Mean Workers transitions between sectors Number of observations Age (years) Schooling (years) Experience (years) Initial real wage (USD) From the formal sector to Formal sector 2193 41.8 14.3 10.7 436.8 (0.29) (0.09) (0.23) (10.26) Informal sector 236 39.4 13.1 8.02 293.3 (0.90) (0.27) (0.63) (22.03) Unemployed 35 31.3 15.0 3.91 339.3 (1.56) (0.66) (0.83) (53.90) Out of the labor force 156 49.4 12.4 12.6 306.8 (1.34) (0.35) (1.01) (29.85) From the informal sector to Formal sector 421 40.5 12.51 8.69 259.5 (0.72) (0.22) (0.50) (21.76) Informal sector 2139 42.8 10.9 11.6 198.9 (0.31) (0.09) (0.26) (4.76) Unemployed 72 31.2 13.4 5.6 153.9 (1.36) (0.44) (1.13) (15.28) Out of the labor force 456 43.0 11.1 9.12 106.1 (0.97) (0.23) (0.66) (7.02) Source: Author s calculations based on the Ecuador National Survey of Employment and Unemployment, 2011-2012. Note: The informal sector is formed by no covered social security workers. Standard errors in parenthesis. 25

Table A3. Multinomial Logit Analysis of transitions between sectors - marginal effects, Ecuador, 2011-2012 Workers transitions Regions (d) Log. Diff. Age Female (d) Married (d) Schooling Experience between sectors Coast Centre South earnings From the formal sector to Formal sector -0.001-0.020 0.012 0.035 0.003-0.016-0.021-0.030 0.037 (0.87) (1.25) (0.75) (9.25)*** (3.37)*** (0.83) (0.78) (1.19) (10.79)*** Informal sector 0.001-0.005 0.011-0.021-0.003 0.028 0.034 0.023-0.005 (1.11) (0.36) (0.73) (6.19)*** (3.05)*** (1.55) (1.39) (1.03) (1.59) Unemployed -0.0002-0.005-0.010-0.001-0.001 0.015 0.017 0.008-0.006 (1.09) (0.93) (1.88)* (0.75) (1.77)* (3.27)*** (1.86)* (1.17) (5.97)*** Out of the labor force 0.0001 0.030-0.013-0.013 0.0003-0.026-0.031-0.001-0.030 (0.53) (3.70)*** (1.49) (6.48)*** (0.70) (2.50)** (2.27)** (0.06) (22.74)*** From the informal sector to Formal sector -0.022-0.048 0.020 0.022-0.001-0.066-0.004-0.041 0.014 (4.13)*** (4.10)*** (1.55) (7.26)*** (-1.86)* (3.64)*** (0.16) (1.85)* (5.34)*** Informal sector 0.002-0.031 0.022-0.020 0.003 0.042 0.017 0.009 0.043 (3.52)*** (2.03)** (1.35) (5.08)*** (3.11)*** (1.96)* (0.62) (0.33) (15.09)*** Unemployed -0.001-0.013-0.012 0.003-0.0002 0.004 0.006 0.013-0.008 (3.43)*** (2.38)** (2.19)** (2.14)** (0.71) (0.57) (0.63) (1.28) (8.08)*** Out of the labor force 0.001 0.092-0.030-0.006-0.001 0.020-0.019 0.019-0.049 (2.27)** (8.50)*** (2.90)*** (2.01)** (1.93)* (1.48) (1.15) (1.07) (35.08)*** Source: Author s calculations based on the Ecuador National Survey of Employment and Unemployment, 2011-2012. Note: The coefficients reflect the marginal effect of each covariate on the probability of observing each sector. The informal sector is formed by salaried workers employed by small establishments with less than ten employees and are not registered, plus all independent and self-employed workers. Z statistics are in parenthesis; *p<0.1, **p<0.05, ***p<0.01. (d) Dummy variables; default categories are: male, other civil status except married and north region. 26

Figure A1. Years of experience effect (Formal to formal) Figure A2. Years of experience effect (Formal to informal) Source: Author s calculations based on the Source: Author s calculations based on the Ecuador National Survey of Employment Ecuador National Survey of Employment and Unemployment, 2011-2012. and Unemployment, 2011-2012. Figure A3. Years of experience effect (Formal to unemployed) Figure A4. Years of experience effect (Formal to out of labor f.) Source: Author s calculations based on the Source: Author s calculations based on the Ecuador National Survey of Employment Ecuador National Survey of Employment and Unemployment, 2011-2012. and Unemployment, 2011-2012. Figure A5. Level of education effect (Informal to formal) Figure A6. Level of education effect (Informal to informal) Source: Author s calculations based on the Source: Author s calculations based on the Ecuador National Survey of Employment Ecuador National Survey of Employment and Unemployment, 2011-2012. and Unemployment, 2011-2012. 27

Figure A7. Level of education effect (Informal to unemployed) Figure A8. Level of education effect (Informal to out of labor) Source: Author s calculations based on the Source: Author s calculations based on the Ecuador National Survey of Employment Ecuador National Survey of Employment and Unemployment, 2011-2012. and Unemployment, 2011-2012. 28