Arepas are not Tacos: On the Labor Markets of Latin America

Board of Governors of the Federal Reserve System International Finance Discussion Papers ISSN 1073-2500 (Print) ISSN 2767-4509 (Online) Number 1396 September 2024 Arepas are not tacos: On the labor markets of Latin America Maria Aristizabal-Ramirez, Cezar Santos, and Alejandra Torres Please cite this paper as: Aristizabal-Ramirez, Maria, Cezar Santos, and Alejandra Torres (2024). “Arepas are not tacos: On the labor markets of Latin America,” International Finance Discussion Papers 1396. Washington: Board of Governors of the Federal Reserve System, https://doi.org/10.17016/IFDP.2024.1396. NOTE: International Finance Discussion Papers (IFDPs) are preliminary materials circulated to stimulate discussion and critical comment. The analysis and conclusions set forth are those of the authors and do not indicate concurrence by other members of the research staff or the Board of Governors. References in publications to the International Finance Discussion Papers Series (other than acknowledgement) should be cleared with the author(s) to protect the tentative character of these papers. Recent IFDPs are available on the Web at www.federalreserve.gov/pubs/ifdp/. This paper can be downloaded without charge from the Social Science Research Network electronic library at www.ssrn.com.

Arepas are not Tacos: On the Labor Markets of Latin America* Maria Aristizabal-Ramirez† Cezar Santos‡ Alejandra Torres§ July 17, 2024 Abstract This paper examines labor markets across Latin American countries,revealingsubstantialdifferencesinunemployment,informality, and worker transitions. Using surveys from eight countries, weconstructcomparablestatisticsonemploymentstocksandmobility patterns. Notable cross-country differences emerge, with economies mostly clustered into high unemployment-low informality or low unemployment-high informality groups. Transition probabilitiesanddirectionalflowsalsovarysignificantly. Wehighlight the importance of using country-specific parameters when simulatinglabormarketandaggregateoutcomes. Finally,wecompareourmainresultswiththosebysexandeducationgroups. JELclassifications: E24,E26,J46,O54 Keywords: LatinAmerica,Labormarkets,Informality,Unemployment,Transitions *We have benefited from helpful comments from Matias Busso, Joaqu´ın Garc´ıa- Cabo, John Leahy, Renata Narita, Pablo Ottonello and Todd Schoellman. The views expressedinthisarticlearethoseoftheauthorsanddonotnecessarilyrepresentthose of the Inter-American Development Bank or the views of the Board of Governors of the Federal Reserve System or any other person associated with the Federal Reserve System. †FederalReserveBoard. Email: maria.aristizabalramirez@frb.gov ‡Inter-AmericanDevelopmentBank&CEPR.Email: cezarsantos.econ@gmail.com §Inter-AmericanDevelopmentBank. Email: aletorres6277@gmail.com

1 Introduction HowsimilarareLatinAmericanlabormarkets? Areinformalityandunemployment high everywhere? What about transition rates? Do workers move between the formal and the informal sector similarly? While there is consensus that developing economies have larger informal sectors than developed economies (Ulyssea, 2020), there are striking differences in labor market stocks and transitions across Latin American countries. In this paper, we provide a set of comparable labor market statisticstoshedlightonthespecificsofLatinAmericanlabormarkets. We first show that countries can generally be grouped as high unemployment and low informality or low unemployment and high informality. Second, we document country-specific labor market transitions thatuncoverdifferencesintermsofmobilityandsegmentationbetween the formal and informal sectors. Further, when we decompose the sample certain demographics, we observe that differences across countries remain in terms of sex. However, with regards to education, highly educated workers on average have lower unemployment and lower informalityeverywhere. Finally,usingasimplemodelandcalibratingitwith the provided transition rates, we show that simple differences in labor marketpatternshavemeaningfulaggregateimplicationsintermsofconsumptionandsavings. We start by constructing a comprehensive dataset of labor market variables from 2012 to 2019 using surveys from eight countries: Argentina, Bolivia,Brazil,Chile,CostaRica,Ecuador,Mexico,andParaguay. Aunifieddefinitionofinformalityclassifiesworkersashavinganinformaljob ifemployedbutnotreportingtheirstatustosocialsecurityortaxauthorities. An examination of average unemployment and informality rates reveals striking cross-country differences, with unemployment ranging from3%inMexicoto14%inBrazil,andinformalityfrom15%inChileto 80%inBolivia. Tosystematicallyanalyzethesedifferences,wedepictthe distribution of countries in a two-dimensional matrix of unemployment andinformalitymiddlevalues’. Ourfirstresultisthatmostcountriesfall into two groups that we will denominate Group I or Group IV. Group I 1

corresponds to countries with high unemployment and relatively low informality (Argentina, Brazil, Chile, Costa Rica). Group IV are countries with low unemployment and high informality (Bolivia, Ecuador, Mexico,Paraguay). We also analyze labor market transitions across formal employment, informal employment, and unemployment. We break down the analysis into two parts. First, we distinguish countries by the absorbing states: where are workers more likely to remain? Second, we focus on how workers move across states. The main takeaway is the emergence of substantialcross-countrydifferencesinmobilitypatternsanddirectional flows. First, regarding the absorbing states, formal to formal transitions are the most prevalent state, and high-unemployment countries (Group IV) tend to exhibit high unemployment to unemployment rates. Similarly, high-informality countries (Group I) have high informal to informal rates. Second, in terms of mobility, we find heterogeneity in the speed of transitions across both groups of countries. For example, some countries like Chile and Mexico exhibit higher worker mobility across employment states, driven by either a higher propensity to exit unemployment(Mexico)orinformaljobs(Chile). Incontrast,morerigidmarkets like Paraguay and Costa Rica see workers frequently remaining in informal employment or unemployment. In terms of direction, highinformality countries witness a greater likelihood of transitioning from formal to informal jobs upon job loss, while a trend of shifting to unemploymentprevailsinhigh-unemploymenteconomies. Leavinginformality is the primary driving force in low-informality high-unemployment countries (Group I), whereas leaving unemployment is more common in high-informality, low-unemployment economies (Group IV). As a final result of this analysis, certain labor markets appear segmented in terms of formal and informal opportunities. For instance, Bolivia and Paraguay exhibit restricted mobility back into formality for informal workers. Moreover, we assess the importance of these labor market transitions when modeling economic outcomes in Latin American countries. To do this,wewriteasimplemodelofconsumptionandsavingswithtwodif- 2

ferent sectors, formal and informal, and unemployment. Earnings followa Markovprocess. First,we comparethe implied pathsof informalityandunemploymentinBrazilandColombiausingtransitionprobabilities from Mexico and Brazil. Borrowing parameters from one country to simulate another can lead to quite different and potentially misleading results. Second, when we calibrate the model to Brazil or Colombia using both Brazil’s and Mexico’s transitions, we find that the observed labormarkettransitionscanamounttolargedifferencesintheassets-toincome ratio. However, as differences in unemployment and informalityearningsdisappear,thesetransitionsplayasmallerroleinexplaining consumption-and-savings decisions. Hence, differences in labor market transitions can be important aspects to take into account when thinking about policies in terms of consumption, savings, and ultimately taxes. Not considering the subtleties of the labor market could lead to meaningful mismeasurements or aggregate outcomes. These results are in linewithRestrepo-Echavarria(2014),whoshowsthatmismeasurements ofinformalitycanleadtohigherbusinesscyclevolatility. Finally, we explore differences in unemployment and informality rates and transitions across employment states by sex and education level. Whiletheoverallcountryclassificationsintogroupsbasedonhigh/low unemployment and informality holds on average across demographics, there are notable variations. First, female workers exhibit higher unemployment and informality rates compared to males, with the sex unemployment gap being larger than the informality gap. Second, the differences across countries are less pronounced for highly educated workers who predominantly fall into the low unemployment-low informality group. Conversely, less-educated workers experience starker crosscountry disparities, with higher unemployment in high-unemployment countries and higher informality in high-informality countries. Third, there is greater labor mobility for men and less-educated individuals. Maintaining formal employment remains the likeliest outcome across groups, though men and less-educated workers tend to persist in informalitymorethanwomenandhighlyeducatedmen. Related Literature. Our paper makes two contributions to the litera- 3

ture. First, by generating comparable measures of labor market dynamics across Latin American countries, our paper produces ready-to-use labor market parameters. Our paper systematically compares informality and different labor market transitions across Latin American countries,somewhatsimilarlytowhatHobijnandS¸ahin(2009)didforOECD countries. Creating a compelling and comparable data base of Latin American countries was challenging until recently. For this reason, previous papers had to focus on particular countries, like Brazil (Gerard and Gonzaga, 2021; Gomes et al., 2020) or Mexico (Gong and Van Soest, 2002; Maloney, 1999). Alternatively, when cross-country comparisons were made, they were restricted to certain demographics (Bosch and Maloney,2010;Funkhouser,1996). Second, we show that, despite general similarities (e.g., relatively high informalityrates),LatinAmericancountriesexhibitsubstantialvariation in labor market outcomes. In this regard, we build on the analysis of labormarketsanddevelopmentbyDonovanetal.(2023),butinsteadfocus on the diversity of experiences across Latin American labor markets. By looking at high-informality countries, our work relates to Meghir et al. (2015), Ulyssea (2018), Herren˜o and Ocampo (2023), and Menezes-Filho and Narita (2023), among others. We innovate by documenting that unemployment, informality and employment transition rates exhibit substantial variation in Latin America. We also show that this substantial variationhasimportantimplicationsforsimulationsoflabormarketand aggregateoutcomes(asin,e.g.,Chodorow-ReichandCoglianese,2021). 2 An Overview of Unemployment and Informality in Latin America To analyze the specifics in the labor market characteristics across Latin American countries, we create a comparable data set of average labor market stocks and transitions between 2002 and 2022 using data from Argentina,Bolivia,Brazil,Chile,Colombia,CostaRica,Ecuador,Mexico, andParaguay. Toconstructthisdataset,weuselabormarketsurveysfor 4

Figure1: Informalityvs. Unemployment: Twodifferentforces Note: Theverticalaxisshowstheaverageunemploymentrate,andthehorizontalaxis shows the average informality rate. Dot sizes represent GDP per capita, where Chile is the largest and Bolivia the smallest. The graph is divided into four groups using theaverageunemploymentandinformalityratesacrosscountries. GroupIinthetop left corner shows high unemployment with low informality, Group II in the bottom left corner shows low unemployment and low informality, Group III in the top right corner shows high unemployment and high informality, and Group IV in the bottom rightcornershowslowunemploymentandhighinformality. Datasource: Countries’ labormarketsurveysandauthors’calculations. SeeAppendixAfordetails. each country similar to Donovan et al. (2023) and unify the definition of informality. Workers are classified as having an informal job if they are employed and do not report their working status to the corresponding socialsecurityortaxoffice. Workersareclassifiedasunemployedifthey report not having a job but are actively looking for one. Both informal and formal workers can transit to unemployment. Finally, workers are classified as out of the labor force if they do not report having a job and are not actively looking for one. The unemployment rate is computed as the ratio of unemployment to total employment (formal and informal)plusunemploymentandtheinformalityrateasinformalitytototal employment. SeeAppendixAforfurtherdetails. 5

We start by describing the differences in average unemployment and informality rates. We choose these two stocks to describe the markets because unemployment is the most common statistic used in the literature to characterize a labor market, and there is a common belief that informality is high across all Latin America. In addition, participation rates are similar across most of these countries, as we show in Figure B1 intheAppendix. Figure 1 shows on the horizontal axis the average informality rate per country and the average unemployment rate on the vertical axis. The size of each dot represents GDP per capita, with Chile being the largest, and Bolivia the lowest. The horizontal dashed line shows the average value of unemployment and the vertical dashed line shows the average value of informality across our sample. The graph then displays four potential regions in which to group countries. The top-left region represents high unemployment and low informality (Group I), the bottomleft region represents low informality and low unemployment (Group II),thetop-rightregionrepresentshighinformalityandhighunemployment (Group III), and the bottom-right region represents high informalityandlowunemployment(GroupIV).Wefixthesecut-offsandregions fortherestofouranalysis. Thedifferencesintherangeofbothunemploymentandinformalityrates are striking. The average unemployment rate varies from 3% in Mexico to almost 14% in Brazil, and informality varies from around 15% in Chile to almost 80% in Bolivia. Notably, virtually all countries in our sample fall into Groups I and IV. Countries in Group I are Argentina, Brazil, Chile, and Costa Rica, whereas in Group IV, we have Bolivia, Ecuador, Mexico, and Paraguay. The sole exception is Colombia, which falls in Group III, with both high unemployment and high informality. We will drop Colombia from most of the analyses of the next few sections, however, as the country does not have panel data that allow us to compute transitions across employment states. We include it in Figure 1 to highlight the varied experiences in labor markets across Latin America. Moreover,wewilluseColombiainSection3.1tohighlightthe importanceofusingcountry-specificparametersinsimulations. 6

There are three important characteristics of groups I and IV. First, the average GDP per capita of Group I countries with high unemployment and low informality is $19,500, which is, on average, higher than that of Group IV countries with $12,700 GDP per capita. This is consistent with Donovan et al. (2023) who find that income and informality are negatively correlated. However, Mexico is the one case with a GDP per capitaof$18,700,whichishighbyLatinAmericanstandards,yethaslow unemploymentandhighinformality.1 Second,GroupI’sunemployment rateisashighasthatinSpain,thehighestunemploymentratewithinthe European Union (EU). Group IV’s unemployment rate is comparable to countriesliketheUnitedStatesorCanada. Third,theinformalityratein Group I is somewhat comparable with the informality rate in some EU countriesaccordingtotheInternationalLaborOrganization(ILO,2020), while Group IV’s informality rate is higher than that in all advanced economies. Most countries, however, do not produce informality measures,andiftheydo,thedefinitionisnotnecessarilycomparable. 3 Labor Market Transitions across Employment States Inthissectionwefocusonlabormarkettransitions. Table1showslabor market transitions across formal jobs (F), informal jobs (I) and unemployment (U). The sum of transitions from a given state including out of the labor force sum up to one. For simplicity of notation we exclude from the main analysis values out of the labor force. 2 Each column OD shows the average quarterly transition rate per country from the origin state O to the destination state D. For example, FI is the probability of moving from a formal job to an informal job in the next quarter. Countries at the top of the table are those in Group I as defined in Section 2, and countries in the bottom part of the table are those in Group IV. The bottom two rows are the middle value (average between the maximum 1WeuseGDPandpopulationdatain2019fromPennWorldTable9.1. 2WeprovidethetablewithforthenoparticipationratetransitionsinTableB2inthe Appendix. 7

Table1: LaborMarketTransitions: MobilityandDirection Country FF II UU FI FU IF IU UF UI Argentina 90.41 61.75 37.13 5.58 1.53 16.16 7.83 8.51 21.96 Brazil 82.61 69.41 45.67 10.47 3.20 16.67 6.25 12.10 17.96 I Chile 90.58 50.06 32.73 3.34 2.89 24.54 7.62 25.66 13.55 CostaRica 90.18 66.69 33.93 5.76 2.05 12.37 6.39 14.35 22.96 Bolivia 60.09 73.28 19.32 18.23 2.32 6.50 2.53 10.34 28.76 Ecuador 83.29 72.57 27.90 10.20 2.10 12.86 2.26 20.00 21.28 IV Mexico 79.54 68.35 20.08 14.05 1.96 15.38 2.49 22.90 30.11 Paraguay 88.62 79.05 34.59 8.30 1.88 6.90 4.44 8.77 34.49 Middlevalue 75.34 64.55 32.50 10.79 2.36 15.52 5.05 17.09 24.02 Coeffofvar. 0.12 0.13 0.28 0.51 0.25 0.42 0.47 0.44 0.29 Note: Each column OD shows the average labor market transition rate per country from the origin state O to the destination state D. F denotes formal, I informal, and U unemployment. These transitions plus values from each state to out of the labor force sum up to one. Countries at the top of the table are those in Group I as defined in section 2, and countries in the bottom part of the table arethoseinGroupIV.Thebottomtworowsarethemiddlevalueofeachcolumnandthecoefficient ofvariation. Highlightedcellsrepresenttransitionsabovethemiddlevalue. Datasource: Countries’ labormarketsurveysandauthors’calculations. SeeAppendixAforfurtherdetails. and minimum value of each column), and the coefficient of variation (standarddeviationdividedbythemeanofeachcolumn).3 Highlighted cellsarethoseaboveeachcolumn’smiddlevalue. To study these transitions, we break down the analysis into two parts. First, we distinguish countries by the absorbing states: where are workers more likely to remain? Second, we focus on how workers move acrossstates,andpointoutthatlabormarketsdifferintermsofmobility anddirection. Thesethreecharacteristics,theabsorbingstates,direction and mobility, are driven by the dominant state, unemployment (Group I)orinformality(GroupIV)asdiscussedinSection2. Start by focusing on the probability of staying in the same state, the first three columns of Table 1. Across all Latin American labor markets,remaininginaformaljob(FF)isthemostlikelyoutcome,followed by remaining in an informal job (II), with Bolivia being the only exception, where remaining informal is more likely than remaining for- 3Weusethemiddlevaluetocapturelargedifferencesacrossthesample. 8

mal. Keeping a formal job (FF) has both the highest probability and the lowest coefficient of variation. Not surprisingly, the probability of keeping an informal job is high in countries with high-informality and low-unemployment (Group IV) and the probability of remaining unemployed (UU) is high for low-informality and high-unemployment countries. Theprobabilityofremaininginformal(II)islowonlyforChileand Argentina. Intermsofmobility,ChileandMexicohavehigherprobabilitiesofmoving into and out of a particular state. However, two distinct forces are at play: a higher likelihood of exiting unemployment and a higher likelihood of exiting informality. In Mexico, workers transition more frequently across states due to a higher rate of leaving unemployment. In Chile,highermobilityisdrivenbyworkersleavinginformaljobs. Incontrast, there are more rigid markets like Paraguay and Costa Rica, where workers do not return to the formal sector as frequently. In Costa Rica, theytendtoremainunemployed,whileinParaguay,theymostlyremain informal. In this sense, the informal sector plays different roles across countries. In Chile, for example, the informal sector is more temporary, and it is a state where the worker goes while looking for a formal job. In contrast, in Mexico or Bolivia it is a more permanent state. That is, it is equivalent to having a formal job, with the difference that workers do notreporttheirstatustothecorrespondingsocialsecurityortaxoffice. In terms of direction, when a formal worker loses their job, they can get an informal job or remain unemployed. High-informality countries have a higher probability of moving from a formal to an informal job (FI). Similarly, high-unemployment countries have a higher probability ofmovingtounemployment(FU).Theprobabilityofmovingfromaformal job to informality is always greater than the probability of moving to unemployment. However, moving from a formal to an informal job has the highest variability. A worker who loses their formal job moves on average 68% of the time to an informal job and 1% of the time to unemploymentinMexico,whereasinChiletheprobabilityofmovingtoan informal job is only 35.5% and the probability of moving to unemploy- 9

ment is 30%.4 It is worth mentioning that informality is generally not a typeofunemploymentorunderemployment,that,ifcombinedwithunemployment, will eliminate labor market differences across countries. Ulyssea (2020) highlights that formal and informal firms co-exist within thesameindustriesandevenwithsimilarproductivitylevels. A worker in a low-informality high-unemployment country (Group I) who ended up in an informal job is more likely to leave informality (either to formal employment, IF, or unemployment, IU) than a worker in a high-informality country. There are two observations worth noting, however. First, the probability of leaving an informal job for a formal job is always higher than moving to unemployment, but the probability of moving to unemployment in Group I is about two times larger than the probability of moving to unemployment in Group IV. This means that leaving informality is the driving force guiding the direction labor marketflowsinthelow-informalityandhigh-unemploymentcountries, and not entering unemployment is the driving force in high-informality low-unemployment countries. Second, some markets in both groups of countriesaresegmented,meaningthatworkersdonoteasilytransitback to formality. For instance, IF in Costa Rica is substantially lower than IF in the rest of Group I, and IF in Bolivia and Paraguay is half of Mexico’s orEcuador’spercentage. Finally, unemployed workers in high-informality low-unemployment countries(GroupIV)haveahigherflowoutofunemployment(eitherto formal employment, UF, or informal employment, UI) and, unsurprisingly, the flow is greater towards informality. There are also important distinctions to note when a worker leaves unemployment. First, Chile stands out because it has the lowest levels of informality across all the region. In this country, not only are returns to formality after unemployment (UF) the highest, but also the probability of moving into informality is the lowest. Second, in Argentina, Paraguay and Bolivia, it is very unlikely that an unemployed worker returns to the formal sector, whereas in Mexico or Ecuador, the worker is equally likely to get a 4Thesepercentagesarecalculatedconditionalonworkersleavingtheformalsector, thatis,consideringFI,FU andFN. 10

formaloraninformaljob. 3.1 The Importance and Implications of Country-specific Parameters Inthissection,usingdatafromBrazilandColombiaandasimplemodel, weshowhowandwhythesedifferencesacrosslabormarketsmatterfor aggregate outcomes. The main takeaway is that labor market heterogeneityacrossgroupsIandIVrequirespolicymakersandresearchersto take into account these differences before arriving at savings, consumption and ultimately welfare conclusions. To do so, we first show the differences in the paths to steady state of unemployment and informality if countries had different transition probabilities. Second, we show that the paths not only imply different steady state labor market stocks, but alsothattheyhavepotentiallymeaningfulaggregateimplications. First,howwouldinformalityorunemploymentlookinaBrazil(aGroup I country) if we assume Mexico’s labor market structure (Group IV)? To answer this question, we use data from Brazil and Colombia and comparetheimpliedpathsofinformalityandunemploymenttosteadystate using Mexico’s and Brazil’s transition probabilities from Table 1. In Figure 2, we take a state vector s in 2012Q1 for Brazil and apply its own 1 transition matrix (solid dark line). Alternatively, we take the same initialstatebutusethetransitionmatrixfromMexicoinstead,givingusan alternative time series for {F,I} (dashed light line). We compare these twoscenarioswithBrazil’sactualdata(thedotsinthegraph). Figure2a shows informality, and Figure 2c shows unemployment. If we use Mexico’s transition matrix to simulate Brazil’s informality, we overstate informalityastheinitialstockisbelowMexico’saverageinformality. Similarly,simulatingBrazil’sunemploymentwithMexico’stransitionunderstates Brazil’s unemployment. Moreover, this would suggest a counterfactually quick transition from an unemployment rate of 12 percent to anunemploymentrateof3percent. Second, how much can we infer for a country for which we do not have 11

complete data, like Colombia, from other countries’ experiences? Figures 2b and 2d use Mexico’s and Brazil’s transition matrices to generate paths for informality and unemployment for Colombia. Recall that it is not feasible to construct quarterly transitions using Colombia’s labor market survey, which is not a panel dataset. The results of these simulations are telling. Both Brazil’s and Mexico’s parameters understate Colombia’s informality rate. If applied to Colombia, Brazil’s transition probabilitiessuggestapaththat,withintwoyears,movesfroma70percent informality rate to 40 percent. Using Mexico’s transitions, the implied path reduces informality from 70 to 50 percent. With respect to unemployment (Figure 2d), the gap between Colombia’s data and the path with Brazil’s transition is not as significant. Brazil’s transition matrixwouldsuggesthigherunemployment,whereasMexico’stransitions wouldsuggestanunemploymentrateabout10percentagepointslower thanthedata. Third, what are the aggregate implications of these differences? To answer this question, we now set up a very simple model. Suppose a worker has preferences represented by the utility function v(c) = c1−σ/1 − σ, with a time discount factor of β. The worker faces shocks to which sector s they participate in: formal f, informal i or unemployment u. The probability of switching from sector s in the current period to sector s′ in the next period is given by π(s,s′) and the income for sectorsisgivenbyy(s). Theworkercaninsureagainsttheseincome-sector shocksbysavinginaone-periodbondthatpaysanon-contingentinterestrater. Thevaluefunctionfortheworker’sproblemreads: (cid:88) V(a,s) = maxv(c)+β π(s,s′)V(a′,s′) c,a′ s′ s.t. c+a′ = y(s)+(1+r)a. We calibrate the model setting β to 0.98, σ to 2, and r to 0.01, standard values. Inaddition,weuserelativeearningsdataforBrazil,andColombia. Informal earnings relative to formal earnings in Brazil are 0.74 and unemployment earnings are 0.30. In contrast, in Colombia informal 12

earnings are 0.37 while unemployment earnings are 0.27. In this simpleexercise,wecomputeassets-to-earningsratiosforBrazilandColombia using Brazil’s and Mexico’s transitions and show them in Figures 2e and 2f. We draw two important conclusions. Given the earnings differences in Brazil between formal and informal workers, changing the probability of having a formal job has large implications for savings, as the assets-to-earnings ratio decreases from 175% to 115%. In contrast, in Colombia the difference is almost nonexistent. This suggests that as long as there are productivity differences between informality and unemployment, and returns across these states differ, these labor market transitions may play determinant roles in shaping aggregate dynamics intermsofsavingsandconsumption. 3.2 The Takeaway As the title of this paper quips, although there are similarities in the labormarketsofLatinAmerica(e.g.,relativelyhighinformalityrates),the exercises of this section provide a cautionary tale. Despite certain similarities, several differences persist. These differences arise from the role that unemployment and informality play in shaping the direction and mobility of labor markets in these countries. Further, policies that interact with sectoral shocks (e.g., taxes or retirement plans) should consider these labor market nuances, as they influence both savings and consumptionpaths. 4 Unemployment, Informality and Demographics In this section we explore the differences in labor market stocks and transitions by demographics. We split our sample by sex and education level, and we classify workers with college or higher degrees as highly educated and workers with less than college as low educated. We start by showing differences in informality and unemployment, and then we 13

Figure 2: Implied Informality and Unemployment Paths in Brazil and ColombiaUsingBrazil’sandMexico’sTransitions (a)InformalityBrazil (b)InformalityColombia (c)UnemploymentBrazil (d)UnemploymentColombia (e)AssetstoearningsBrazil (f)AssetstoearningsColombia Note: The vertical axis in panels (a) and (b) show informality rate in percentage, and inpanels(c)and(d)showsunemploymentrateinpercentage. Panels(a)and(c)show informality and unemployment rates for Brazil. Dots show data. The solid dark line shows the implied informality or unemployment rate for Brazil using Brazil’s transition probabilities in Table 1. The dashed light line shows the implied informality or unemploymentrateforBrazilusingMexico’stransitionprobabilitiesinTable1. Panels (b)and(d)showanequivalentscenarioforColombia. Panels(e)and(f)showthethe assetstoearningsratiosimpliedbythemodelusingBrazil’sandMexico’stransitions. Data: Countries’labormarketsurveysandauthors’calculations. 14

Figure3: Informalityvs. UnemploymentbyDemographics (a)Sex (b)Education Note: Theverticalaxisshowsaverageunemploymentrates, andthehorizontalaxisshowsaverageinformality rates.ThegraphisdividedintofourgroupsasdescribedinSection2.Panel(a)showsdifferencesacrosssex.Diamondsrepresentwomenandcirclesmen. Panel(b)showsdifferencesacrosseducationlevel. Trianglesrepresent low-educationworkersandsquareshigh-educationworkers. Higheducationiscollegeormore,loweducation less than college. Data source: Countries’ labor market surveys and authors’ calculations. See Appendix A for furtherdetails. discussdifferencesinthetransitions. Figures 3a and 3b show the differences in unemployment and informality. TheanalysisissimilartothatofSection2. Thehorizontalaxisshows informality, the vertical axis shows unemployment and the lines separate the graph in four groups according to the aggregate averages. Diamonds represent women and circles men in Figure 3a, while triangles represent low-education workers and squares represent high-education workersinFigure3b. Start with sex. Female workers have higher unemployment and higher informality. But on average, the classification of countries into the four groups (high/low unemployment and high/low informality) does not change when considering sex differences. This result is consistent with previous work documenting sex unemployment and informality gaps (Albanesi and Sahin, 2013; Azmat et al., 2006; Galiani and Weinschelbaum,2012;Ulyssea,2020). However,whenwecomparetheunemployment and informality gaps, we find that they are statistically different with 95% confidence intervals and that the unemployment gap is on 15

average higher than the informality gap. In particular, women’s unemployment is on average 36% larger than men’s unemployment, and women’s informality rate is on average 13% higher than men’s informalityrate. There are three points to highlight about this result. First, Chile is the only country where the informality gap is larger than the unemployment gap. Second, Mexico is the only country where male’s informality ishigherthanfemale’sinformality,andtheunemploymentgapisalmost zero. Third, the compositional effect in Paraguay and Costa Rica is pronounced, and the group classification changes. Men in Costa Rica are in Group II (low informality and low unemployment), while women in ParaguayareinGroupIII(highinformalityandhighunemployment). Now turn to education. There is a strong compositional effect. Differences across countries in labor market outcomes are less apparent for high-education workers. Not surprisingly, high-education workers are on average in Group II (low informality and low unemployment) (Mincer, 1991). On the other hand, for low-education workers, the differences are more pronounced. Low-education workers have on average higher unemployment for the countries in Group II or higher informality for those in Group IV. Two important nuances materialize. Bolivia is the only country where high-education workers experience on average moreunemploymentthanlow-educationworkersandinformalitylevels are above the cross-country average. Unemployment of high-education workers in Chile, although lower, is closer to the aggregate unemploymentratethanitisforallothercountries. Finally, turn to transitions across employment states. Table 2 shows the corresponding transitions for each demographic group. The table follows the same structure we used before. This time, however, the second column defines the demographic group. M denotes men, W women, H high education, and L low education. There are three points to note. First, the labor market exhibits higher worker mobility, on average, for men and low-education workers. Second, keeping a formal job remains the most likely outcome for all demographic categories. The countries 16

Table2: LaborMarketTransitionsbySexandEducation Country Type FF II UU FI FU IF IU UF UI Sex M 91.80 65.60 40.81 5.08 1.63 16.59 9.21 10.13 26.57 Argentina W 88.61 58.02 33.83 6.23 1.39 15.69 6.50 7.13 17.81 M 83.30 70.15 45.45 10.46 3.15 17.67 6.42 14.61 22.13 Brazil W 81.66 68.59 45.82 10.48 3.26 15.56 6.06 10.27 14.86 I M 90.75 48.78 33.27 3.39 3.31 28.65 8.80 31.74 14.68 Chile W 90.35 51.29 32.12 3.27 2.32 20.54 6.47 19.48 12.36 M 90.16 68.82 33.63 6.16 2.07 15.66 7.23 17.95 27.32 CostaRica W 90.22 64.37 34.16 5.09 2.00 8.77 5.51 10.74 18.49 M 62.46 75.91 19.04 18.48 2.37 7.62 3.00 12.95 35.71 Bolivia W 56.25 70.16 19.64 17.91 2.22 5.19 1.97 7.70 21.73 M 83.91 76.43 28.51 11.67 2.05 15.74 2.66 24.36 28.10 Ecuador W 82.42 67.97 27.38 8.15 2.16 9.47 1.80 16.07 15.29 IV M 80.35 75.24 21.80 15.25 2.22 16.26 2.98 25.52 36.52 Mexico W 78.28 57.33 17.50 12.18 1.57 13.97 1.72 18.98 20.44 M 88.78 80.95 33.89 8.58 1.71 7.73 4.87 8.98 39.48 Paraguay W 88.37 76.90 35.20 7.88 2.14 5.91 3.96 8.79 30.72 Education H 94.76 52.40 42.46 2.84 0.73 34.53 5.30 16.81 13.92 Argentina L 88.50 62.50 36.62 6.78 1.87 14.73 8.03 7.69 22.69 H 84.91 77.21 50.87 10.83 1.92 17.64 1.81 17.18 12.49 Brazil L 82.04 67.61 45.22 10.39 3.51 16.44 7.29 11.63 18.45 I H 94.25 53.74 42.99 2.28 1.95 30.71 7.25 26.87 10.97 Chile L 88.60 49.15 29.00 3.92 3.40 23.08 7.72 25.26 14.42 H 94.99 52.49 46.49 2.70 0.88 28.63 5.53 19.70 12.50 CostaRica L 88.95 67.27 33.11 6.56 2.35 11.64 6.48 13.92 23.67 H 67.06 62.80 29.04 11.79 2.63 12.77 3.20 15.42 18.17 Bolivia L 56.01 74.80 15.23 22.01 2.12 5.72 2.45 8.33 33.10 H 91.62 66.79 37.27 5.23 1.32 21.57 3.47 24.07 17.01 Ecuador L 80.02 73.12 25.86 12.14 2.40 12.01 2.15 19.13 22.25 IV H 87.73 51.12 29.81 7.21 1.61 31.43 3.95 27.65 18.78 Mexico L 75.56 69.86 17.22 17.38 2.14 13.97 2.36 21.53 33.41 H 92.49 77.76 39.74 5.97 1.33 14.94 3.18 25.44 30.34 Paraguay L 86.69 79.22 34.54 9.40 2.46 5.84 4.64 7.74 34.87 Middlevalue 75.34 64.55 32.50 10.79 2.36 15.52 5.05 17.09 24.02 Note: Each column OD shows the average labor market transition per country and type of demographicsfromoriginO todestinationD. Mdenotesmen, Wfemale, Hhigh-educationworkersand Llow-educationworkers. Fdenotesformal,Iinformal,andUunemployment. Countriesonthetop of each panel are those in Group I as defined in Section 2, and countries in the bottom part of the tablearethoseinGroupIV.Thebottomrowisthemiddlevalueofeachcolumn,asusedinSection3. Highlightedcellsrepresenttransitionsabovethemiddlevalue. Datasource: Countries’labormarket surveysandauthors’calculations. SeeSectionAintheAppendixforfurtherdetails. 17

previously classified in Group I (high unemployment, low informality) tend, on average, to have workers remaining more in unemployment across all demographics. Conversely, countries in Group IV (low unemployment,highinformality)tendtohaveworkersremainingmoreinthe informal sector across all demographics. However, there are some differenceswithindemographics. Menandhigh-educationworkerstendto leave informality faster (IF and IU), while high-education workers and women have lower chances of leaving unemployment to get a job (UF, andUI).Third,thedirectionoftheflowsdoesnotreversecomparedwith the aggregate outcomes. For countries where leaving informality was thedominatingforce,itcontinuestobethecase,andforcountrieswhere avoiding unemployment was driving the direction, it also remains the same. Nonetheless, switching from unemployment to informality is the highestformen,particularlyformenincountriesinGroupIV. 5 Conclusion This paper’s analysis of labor markets across Latin American countries reveals substantial differences in unemployment, informality, and worker transitions. These results highlight the importance of using country-specific parameters when modeling economic outcomes in the region. Despite sharing some broad similarities, persistent differences emergeinthelevelsofunemploymentandinformality,themobilitypatterns across employment states, and the directional flows between formal,informal,andunemployedsectors. This paper provides a rich set of comparable labor market statistics and highlights that workers move differently across labor market states across Latin American countries. It is possible to divide countries into twogroups,onewhereunemploymentplaysamajorroleandonewhere informality is key in explaining the mobility and direction of the labor market. Moreover, the role of informality is heterogeneous across countries. For some, like Chile, it is a temporary state that workers transit before they find a different formal job. For others, like Bolivia, the for- 18

malandtheinformalsectorsarealmostmutuallyexclusiveandworkers donoteasilytransitbetweenthetwosectors. Thecross-countryvariationsinlabormarketstructures,workforcecompositions, and economic conditions require an approach that accounts for the unique characteristics of each national labor market. We also show that the differences in these labor market transitions have important implications for aggregate outcomes such as labor market stocks, consumption and savings decisions. It is then an interesting avenue of futureresearchtostudyhowthesedifferencesimpacttheroleofpolicies thatinteractwithsuchdecisions. Bibliography ALBANESI, S. AND A. SAHIN (2013): “The gender unemployment gap: Trend and cycle,” Staff Reports (613). New York: Federal Reserve Bank of NewYork. AZMAT, G., M. GU¨ELL, AND A. MANNING (2006): “Gender gaps in unemployment rates in OECD countries,” Journal of Labor Economics, 24, 1–37. BOSCH, M. AND W. F. MALONEY (2010): “Comparative analysis of labor market dynamics using Markov processes: An application to informality,”LabourEconomics,17,621–631. CHODOROW-REICH, G. AND J. COGLIANESE (2021): “Projecting unemployment durations: A factor-flows simulation approach with application to the COVID-19 recession,” Journal of Public Economics, 197, 104398. DEPARTAMENTO NACIONAL DE ESTAD´ISTICA, COLOMBIA(2011-2024): “GranEncuestaIntegradaHogares(GEIH),”. DONOVAN, K., W. J. LU, AND T. SCHOELLMAN (2023): “Labor market dynamics and development,” The Quarterly Journal of Economics, 138, 2287–2325. FUNKHOUSER, E.(1996): “TheurbaninformalsectorinCentralAmerica: Householdsurveyevidence,”Worlddevelopment,24,1737–1751. 19

GALIANI, S. AND F. WEINSCHELBAUM (2012): “Modeling informality formally: householdsandfirms,”EconomicInquiry,50,821–838. GERARD, F. AND G. GONZAGA (2021): “Informal labor and the efficiency cost of social programs: Evidence from unemployment insurance in brazil,” American Economic Journal: Economic Policy, 13, 167– 206. GOMES, D. B., F. S. IACHAN, AND C. SANTOS (2020): “Labor earnings dynamicsinadevelopingeconomywithalargeinformalsector,”JournalofEconomicDynamicsandControl,113,103854. GONG, X. AND A. VAN SOEST (2002): “Wage differentials and mobility in the urban labour market: a panel data analysis for Mexico,” Labour economics,9,513–529. HERREN˜O, J. AND S. OCAMPO (2023): “The macroeconomic consequences of subsistence self-employment,” Journal of Monetary Economics,136,91–106. HOBIJN, B. AND A. S¸AHIN (2009): “Job-finding and separation rates in theOECD,”EconomicsLetters,104,107–111. ILO (2020): “Overview of the informal economy in the European Union,”. INSTITUTO BRASILEIRO DE GEOGRAFIA E ESTAT´ISTICA (2012–2024): “Pesquisa Nacional por Amostra de Domic´ılios Cont´ınua (PNADC),” AccessedMay1,2019–September15,2021. INSTITUTO NACIONAL DE ESTAD´ISTICA, BOLIVIA (2015–2024): “EncuestaContinuadeEmpleo(ECE),”AccessedAugust31,2018–March 31,2021. INSTITUTO NACIONAL DE ESTAD´ISTICA DE CHILE (1986–2024): “Encuesta Nacional de Empleo (ENE),” Accessed November 29, 2017 – October19,2021. INSTITUTO NACIONAL DE ESTAD´ISTICA, PARAGUAY (2010–2017): “EncuestaContinuadeEmpleo(ECE),”AccessedAugust8,2016–July13, 2019. INSTITUTO NACIONAL DE ESTAD´ISTICA Y CENSOS, ARGENTINA (2003– 2024): “Encuesta Permanente de Hogares (EPH),” Accessed November24,2009–October18,2021. INSTITUTO NACIONAL DE ESTAD´ISTICA Y CENSOS, COSTA RICA (2010– 2024): “EncuestaContinuadeEmpleo(ECE),”Version1. 20

INSTITUTO NACIONAL DE ESTAD´ISTICA Y CENSOS, ECUADOR (2007– 2024): “Encuesta Nacional de Empleo, Desempleo y Subempleo,” AccessedMay21,2016–October25,2021. INSTITUTO NACIONAL DE ESTAD´ISTICA Y GEOGRAF´IA, MEXICO (1995– 2024): “Encuesta Nacional de Ocupacio´n y Empleo (ENOE),” AccessedDecember1,2014–August25,2021. MALONEY, W. F. (1999): “Does informality imply segmentation in urbanlabormarkets? EvidencefromsectoraltransitionsinMexico,”The WorldBankEconomicReview,13,275–302. MEGHIR, C., R. NARITA, AND J.-M. ROBIN (2015): “Wages and Informality in Developing Countries,” American Economic Review, 105, 1509–46. MENEZES-FILHO, N. A. AND R. NARITA (2023): “Labor market turnover and inequality in Latin America,” Tech. rep., IDB Working PaperSeries. MINCER, J. (1991): “Educationandunemployment,”. RESTREPO-ECHAVARRIA, P. (2014): “Macroeconomic volatility: The role oftheinformaleconomy,”EuropeanEconomicReview,70,454–469. ULYSSEA, G. (2018): “Firms,Informality,andDevelopment: Theoryand EvidencefromBrazil,”AmericanEconomicReview,108,2015–47. ——— (2020): “Informality: Causes and consequences for development,”AnnualReviewofEconomics,12,525–546. 21

Appendix A Data For our analysis, we build a comprehensive dataset to provide comparable statistics across countries. To build labor market variables, we use labor surveys that have a rotating panel structure. This way, we can follow individuals for at least a quarter. Table A1 shows the surveys and periodsusedforeachcountry. To build income variables for our model, we use supplementary data to include different sources of income. For Brazil, the PNADC main dataset only includes information on first and second-activity income, which means that we do not have information on income for the unemployed. To address this, PNADC has a supplementary survey that includes information on other sources of income. Data for PNADC are collected in a 1-2(5) rotation scheme, where each household is surveyed for a month, leaves the survey for two months, and then is surveyed again in the following month. The data for the supplementary survey are collected at the first and fifth visits. Our earnings variable is the one defined by the Brazilian statistical office (IBGE) as total income from all sources. ForColombia,GEIHalsoincludesinformationonincomeonlyformain and second activity. We use the “Medicio´n de Pobreza Monetaria y Desigualdad” survey, which includes data on other sources of income to calculate poverty statistics. This survey follows the same methodology usedinGEIH,sowecanmergethetwodatasetsdirectly. A-1

TableA1: DataSources Country Dataset Source Datesused Argentina EPH InstitutoNacionaldeEstad´ısticayCensos,Argentina(2003–2024) 2012q1-2019q4, Exc2015q3-2016q1 Bolivia ECE InstitutoNacionaldeEstad´ıstica,Bolivia(2015–2024) 2016q1-2019q4 Exc2015q1-2015q3,2018q2-2018q3 Brazil PNADC InstitutoBrasileirodeGeografiaeEstat´ıstica(2012–2024) 2012q1-2019q4 Chile ENE InstitutoNacionaldeEstad´ısticadeChile(1986–2024) 2012q1-2019q4 Colombia GEIH DepartamentoNacionaldeEstad´ıstica,Colombia(2011-2024) 2012q1-2019q4 CostaRica ECE InstitutoNacionaldeEstad´ısticayCensos,CostaRica(2010–2024) 2012q1-2019q4 Ecuador ENEMDU InstitutoNacionaldeEstad´ısticayCensos,Ecuador(2007–2024) 2012q1-2019q3 Exc2019q3-2019q4 Mexico ENOE InstitutoNacionaldeEstad´ısticayGeograf´ıa,Mexico(1995–2024) 2012q1-2019q4 Paraguay ECE InstitutoNacionaldeEstad´ıstica,Paraguay(2010–2017) 2012q1-2017q4 Exc2014q4,2017q3,2017q4 A-2

A.1 Details on the Definition of Informality We start with the definition of informality in Donovan et al. (2023). To arrive at a common definition of informality across countries, we make someadjustmentsasfollows: • Argentina: Donovan et al. (2023) use job benefits of the worker. Inthiscase,informalisdefinedasbeingemployedandnothaving anyjobbenefit,whileformalityincludeshavingatleastonetypeof job benefit. Job benefits include paid leave, Christmas bonus, paid sick leave, and social security. We use contributions to pension fundsinstead. • Bolivia: Informality is defined as workers either contributing to a pensionfundorthefirmhavingataxID. • Brazil: The survey asks if the worker has a ’carteira de trabalho,’ which is an employee record. Formality in this case is defined for employed workers who have a ’carteira de trabalho,’ are domestic workers,orcontributetosocialsecurity. • Chile: Includes contributions to pension, to the health system, to unemployment insurance. Also, if workers have annual vacation leave,sickleave,maternityleaveordaycare. • Colombia: Workers are formal if they work at a firm with more than5workersandcontributetopensionfunds. • Costa Rica: Workers are formal if the place they work in is registered and if they have deductions for social security and income tax. • Ecuador: Workers are formal if they work at a firm that is registered and has more than 100 workers. The survey also includes as informalworkerswithnoincomeandhelpersofwageemployees. • Mexico: There is a variable on the survey that indicates whether theworkerisformalorinformal. However,theofficialcalculations A-3

of the informality rate consider informal workers who are vulnerable due to the nature of the economic unit for which they work, as well as those whose labor dependence does not recognize their source of work. This includes the population working in unregistered micro businesses, as well as self-employed persons in subsistence agriculture, those who work without social security, and those whose services are used by registered economic units. We excludedjobbenefitsfromthedefinition. • Paraguay: Donovan et al. (2023) use the variable that indicates whether workers’ contract has either a defined or undefined term. Weaddcontributiontopensionfundstoourdefinition. B Extra Tables and Figures TableB2: TransitionProbability: OutoftheLaborForce Country NN NF NI NU FN IN UN Argentina 85.49 6.05 6.02 2.44 2.48 14.26 32.40 Brazil 76.59 8.25 10.43 4.72 3.72 7.67 24.27 I Chile 86.05 4.56 5.35 4.04 3.18 17.78 28.06 CostaRica 80.87 10.02 6.79 2.32 2.01 14.55 28.77 Bolivia 71.24 17.51 3.50 7.75 19.36 17.69 41.57 Ecuador 80.77 11.20 3.24 4.79 4.42 12.31 30.82 IV Mexico 81.29 12.59 2.18 3.93 4.44 13.78 26.91 Paraguay 80.03 12.75 5.93 1.30 1.20 9.61 22.15 Middlevalue 78.65 11.04 6.31 4.52 10.28 12.73 31.86 Coeffofvariation 0.06 0.40 0.48 0.51 1.15 0.26 0.20 Note:EachcolumnODshowstheaveragelabormarkettransitionratepercountryfromthe originstateOtothedestinationstateD. Ndenotesoutofthelaborforce,Fdenotesformal, Iinformal,andUunemployment. CountriesatthetopofthetablearethoseinGroupIas defined in Section 2, and countries in the bottom part of the table are those in Group IV. Thebottomtworowsarethemiddlevalueofeachcolumnandthecoefficientofvariation. Highlighted cells represent transitions above the middle value. Data source: Countries’ labormarketsurveysandauthors’calculations. SeeAppendixAforfurtherdetails. A-4

FigureB1: ParticipationRate Note: The vertical axis shows the average participation rate per country. Each bar representsacountryfromoursample,andthebarsareorderedfromlefttorightfrom the lowest to the highest GDP per capita. Bars in light dark blue represent countries inGroupIandbarsinlightbluerepresentcountriesinGroupIVasusedinFigure1. Datasource: Countries’labormarketsurveysandauthors’calculations. SeeAppendix Afordetails. A-5

Table B3: Labor Market Transitions out of the Labor Force by Sex and Education Country Type NN NI NU NF FN IN UN Sex M 81.97 7.47 7.32 3.24 1.49 8.60 22.49 Argentina W 86.88 5.49 5.50 2.13 3.77 19.79 41.23 M 73.60 9.55 11.62 5.23 3.08 5.76 17.80 Brazil W 77.77 7.75 9.96 4.52 4.60 9.80 29.05 I M 81.84 5.79 6.48 5.89 2.55 13.76 20.32 Chile W 87.98 4.01 4.83 3.19 4.06 21.70 36.04 M 78.07 10.14 8.05 3.75 1.60 8.29 21.10 CostaRica W 82.08 9.96 6.26 1.71 2.69 21.35 36.61 M 68.30 17.58 3.68 10.44 16.69 13.47 32.30 Bolivia W 73.42 17.29 3.39 5.89 23.62 22.69 50.93 M 80.98 9.96 3.68 5.37 2.37 5.17 19.03 Ecuador W 80.70 11.66 3.07 4.58 7.27 20.76 41.26 IV M 74.49 15.63 3.99 5.89 2.19 5.52 16.17 Mexico W 83.04 11.81 1.72 3.43 7.96 26.99 43.09 M 78.08 13.74 6.48 1.70 0.93 6.46 17.64 Paraguay W 80.91 12.35 5.63 1.11 2.15 13.23 25.29 Education H 84.61 3.58 6.35 5.45 1.67 7.78 26.81 Argentina L 85.55 6.20 5.99 2.26 2.84 14.75 33.00 H 78.18 6.88 8.73 6.20 2.34 3.34 19.46 Brazil L 76.46 8.36 10.57 4.60 4.06 8.65 24.71 I H 79.86 3.95 9.72 6.47 1.52 8.29 19.16 Chile L 86.71 4.62 4.89 3.78 4.08 20.06 31.33 H 79.59 9.01 5.82 5.95 1.42 14.05 22.09 CostaRica L 80.96 10.06 6.84 2.14 2.15 14.61 29.30 H 65.64 14.69 5.87 13.81 18.52 21.23 37.37 Bolivia L 72.61 18.03 3.08 6.28 19.85 17.03 43.34 H 70.60 13.26 6.52 9.86 1.83 8.16 22.25 Ecuador L 81.20 11.12 3.10 4.58 5.44 12.72 32.76 IV H 77.89 9.62 3.87 8.62 3.45 13.51 23.77 Mexico L 81.67 12.92 1.99 3.42 4.93 13.80 27.84 H 75.70 14.38 9.40 9.34 1.42 5.16 18.06 Paraguay L 80.36 12.71 5.86 1.07 1.45 10.30 22.85 Middlevalue 78.65 11.04 6.31 4.52 10.28 12.73 31.86 Note: EachcolumnODshowstheaveragelabormarkettransitionpercountryandtypeof demographicsfromoriginOtodestinationD.Mdenotesmen,Wfemale,Hhigh-education workersandLlow-educationworkers. Ndenotesoutofthelaborforce,Fformal,Iinformal, and U unemployment. Countries at the top of each panel are those in Group I as defined in Section 2, and countries in the bottom part of the table are those in Group IV. Thebottomrowisthemiddlevalueofeachcolumn,asusedinSection3. Highlightedcells representtransitionsabovethemiddlevalue.Datasource:Countries’labormarketsurveys andauthors’calculations. SeeSectionAintheAppendixforfurtherdetails. A-6