Unemployment and Development Data Appendix

Online Appendices A Data Appendix Among the 199 surveys listed below, there are 11 from earlier than 1990, 59 from the 1990s, 88 from the 2000s, and 41 from 2010 and later. Among the 84 countries, there are 55 for which we have at least two surveys. Table A.1: Tier 1, Most Comparable Surveys Tier 1a: Searched for work last week Country Azerbaijan Bangladesh Bolivia Botswana Brazil Burkina Faso Burkina Faso Canada Chile Colombia Costa Rica Cuba Dominican Ecuador El Salvador Fiji Ghana Ghana Ghana Greece Hungary India Indonesia Indonesia Jamaica Kenya Year 1995 2000, 2005, 2010 1992, 2001 2001, 2011 2010 2014 2006 2011 1992, 2002 1993, 2005 2000, 2011 2002 2002 1990, 2001, 2010 1992, 2007 2007 1984, 2000 1998 2010 2001, 2011 2011 1983, 1987, 1993, 1999, 2004 1990, 1995, 2010 2014 1991, 2001 2009 Source Survey of Living Conditions Household Income-Expenditure Survey (HIES) IPUMS-I IPUMS-I IPUMS-I LSMS IPUMS-I IPUMS-I IPUMS-I IPUMS-I IPUMS-I IPUMS-I IPUMS-I IPUMS-I IPUMS-I IPUMS-I IPUMS-I Living Standards Survey IPUMS-I IPUMS-I IPUMS-I IPUMS-I IPUMS-I Indonesia Family Life Survey IPUMS-I IPUMS-I Malawi Malaysia Mexico Mongolia Mozambique Nigeria Pakistan Panama Paraguay Peru Peru Philippines Poland Portugal Romania Rwanda Saint Lucia South Africa South Sudan Spain Sudan Tajikistan Tanzania Trinidad and Tobago Uganda Venezuela Zambia 2008 1991, 2000 1990, 1995, 2000, 2010, 2015 2000 1997, 2007 2010 1973 1990, 2000, 2010 1992 2007 1994 1990 2002 1991, 2001 1992, 2002, 2011 2002 1980, 1991 1993 2008 2011 2008 1999 2002, 2012 1970, 1980, 1990, 2000, 2011 1991, 2002 2001 1990, 2010 IPUMS-I IPUMS-I IPUMS-I IPUMS-I IPUMS-I IPUMS-I IPUMS-I IPUMS-I IPUMS-I IPUMS-I Living Standards Survey IPUMS-I IPUMS-I IPUMS-I IPUMS-I IPUMS-I IPUMS-I Integrated Household Survey IPUMS-I IPUMS-I IPUMS-I LSMS IPUMS-I IPUMS-I IPUMS-I IPUMS-I IPUMS-I Tier 1b: Searched for work in the last 4 weeks Argentina 1991 IPUMS-I Armenia 2011 IPUMS-I Belarus 2009 IPUMS-I Bosnia and Herzegovina 2004 Living in Bosnia and Herzegovina Survey Brazil 1997 Survey of Living Conditions Brazil 2000 IPUMS-I Bulgaria 2007 Multi-topic Household Survey Canada 1991, 2001 IPUMS-I Dominican Republic Iran Iraq Italy Jordan Malawi Paraguay Serbia South Africa Tanzania Uganda United States United States 2010 2011 2012 2001 2004 2013 2002 2007 2007, 2001, 2011 2010 2011 1980, 1990, 2000 2001-2014 IPUMS-I IPUMS-I Household Socio-economic Survey IPUMS-I IPUMS-I Integrated Household Panel Survey IPUMS-I LSMS IPUMS-I National Panel Survey National Panel Survey IPUMS American Community Survey (ACS) Table A.2: Tier 2, Comparable Search Questions, Less Comparable Duration Questions Country Year Source Seeking window Armenia 2001 IPUMS-I Current Bangladesh 1991, 2001 IPUMS-I 7 days main Bangladesh 2011 IPUMS-I Current status Brazil 1980 IPUMS-I Current Burkina Faso 1996 IPUMS-I At least three out of the last week Cambodia 1998, 2008 IPUMS-I 6 month Egypt 2006 IPUMS-I current France 2006, 2011 IPUMS-I Current Haiti 2003 IPUMS-I Last month Hungary 1990 IPUMS-I Current Iran 2006 IPUMS-I Past 30 days Iraq 1997 IPUMS-I Current Ireland 1991, 1996, 2002, 2006, 2011 IPUMS-I Current Kyrgyz Republic 1999, 2009 IPUMS-I Current Mali 1998, 2009 IPUMS-I 4 weeks Morocco 1994, 2004 IPUMS-I Current Nicaragua 2005 IPUMS-I 2 weeks Rwanda 1991 IPUMS-I Most of the week Senegal 2002 IPUMS-I Continuously for at least 3 months Sierra Leone 2004 IPUMS-I 4 weeks South Africa 1996 IPUMS-I Current Switzerland 2000 IPUMS-I Current Turkey 1990 IPUMS-I Current Uruguay 2006, 2011 IPUMS-I 4 weeks Venezuela 1990 IPUMS-I Current Zambia 2000 IPUMS-I Primary activity 7 days Table A.3: Tier 3, Least Comparable Search or Activity Questions Country Argentina Year 2001, 2010 Source IPUMS-I Austria 1991 IPUMS-I Austria Austria Belarus Botswana 2001 2011 1999 2011 IPUMS-I IPUMS-I IPUMS-I IPUMS-I Activity Exclude: for self-consumption A minimum average of 12 hours per week 7 days No text Exclude: for self-consumption 4 Weeks Cameroon 2005 IPUMS-I 7 Days 44 China Ethiopia Fiji France Hungary India Liberia Netherlands 1990 IPUMS-I 2007 IPUMS-I 1996 IPUMS-I 1990, 1999 IPUMS-I 2001 IPUMS-I 2009 IPUMS-I 2008 IPUMS-I 2001 IPUMS-I No text Standard Worked for money Current Current Standard 12 Months No Text Palestine 1997, 2007 IPUMS-I 7 Days Peru Portugal Portugal 1993 1981 2011 IPUMS-I IPUMS-I IPUMS-I Not comparable 7 Days No text Slovenia 2002 IPUMS-I Current Search 4 weeks Current Only previously employed No text Yes Last 7 days for worked before; now for looking for the first job No text No text Not comparable Enrollment ANPE Unemployment benefit Only 12 months main activity available 12 months Not comparable Included did not seek but want to work Not comparable Text not available No text Registered as unemployed at the employment service of Slovenia Spain Switzerland Turkey Ukraine United States Vietnam 1991, 2001 1990 2000 2001 1960 2009, 1999 IPUMS-I IPUMS-I IPUMS-I IPUMS-I IPUMS-I IPUMS-I 7 Days Principal occupation Earn cash or income in kind Status Last week Earn income Unemployed, worked previously Current Last week Unemployment allowances, unemployed Looking for work or laid off 4 weeks 45 B Model Derivations In this appendix, we develop the expressions for Ui (x) and wi (x), and show the intermediate steps to develop equation (12). We start by simplifying equations (7) - (10) to   (1 − δ)Ui (x) = AM b(AM )x + δηθi1−α Ei (x) − Ui (x)   (1 − δ)Ei (x) = wi (x) + δsi Ui (x) − Ei (x) AM x − wi (x) 1 − δ(1 − si ) Z x̄
−α ∗ Ji (x)gi (x)dx. 1 − Gi (xi ) AM c = δηθi Ji (x) = (B.1) (B.2) (B.3) (B.4) x∗i The firm receives (1 − β)Si (x) = (1 − β)[Ei (x) − Ui (x) + Ji (x)] = Ji (x) when a vacancy is filled. Combining this division of surplus with equation (B.3) gives Ei (x) − Ui (x) = β AM x − wi (x) . 1 − β 1 − δ(1 − si ) (B.5) Substituting equation (B.5) into equation (B.1) yields Ui (x) = 1  β AM x − wi (x)  AM b(AM )x + δηθi1−α . 1−δ 1 − β 1 − δ(1 − si ) (B.6) We can then solve for wi (x) by combining equations (B.6) and (B.5) with equation (B.2): wi (x) = AM b(AM )x ki (θi ) β(δηθi1−α + 1 − δ + δsi ) + AM x, where ki (θi ) = . 1 + ki (θi ) 1 + ki (θi ) (1 − β)(1 − δ + δsi ) Substituting this solution into equations (B.3) and (B.6) gives us, respectively,
AM x 1 − b(AM ) (1 − β) Ji (x) = βδηθi1−α + 1 − δ + δsi (B.7)
AM x 1 − b(AM ) (1 − β)  1  1−α β Ui (x) = AM b(AM )x + δηθi . 1−δ 1 − β βδηθi1−α + 1 − δ + δsi (B.8) Equation (B.8) appears as equation (11) in the text. Finally, substituting equation (B.7) into equation (B.4) and dividing both sides by 1 − Gi (x∗i ) yields equation (12) that determines θi for any given level of x∗i : c=
(1 − β)δηθi−α 1 − b(AM ) E(x|x > x∗i ). 1−α βδηθi + 1 − δ + δsi C Model Extension with Worker Churn For simplicity, our model has abstracted from any worker “churn” between the traditional and modern sectors. Donovan, Lu, and Schoellman (2020) assemble data on quarterly labor market transition rates for 42 countries, including transition rates out of self-employment. In their sample, the poorest countries by a wide margin are Nicaragua, Palestine, and the Philippines, for which the average share of low-skill occupations in self-employment in our data is 84 percent. Hence, the self-employment data for these three countries can approximately represent our traditional sector. For these countries, the average probability that a self-employed worker will be unemployed or a wage worker next quarter is 13 percent. This is not a trivial rate of transition to the modern sector, so we would like our model to be robust to incorporating such transitions. Accordingly, we extend the model of Section 4 to allow worker movement between the traditional and modern sectors at a given level of modern sector productivity. Specifically, we introduce idiosyncratic shocks to worker productivity in the traditional sector. This volatility could be caused by weather or commodity price fluctuations, for example. In the model of Section 4, workers below a cutoff ability level always remained in the traditional sector and workers above the cutoff always remained in the modern sector. In the stationary equilibrium of the modified model, there will still exist a cutoff ability level below which workers always remain in the traditional sector, and a cutoff above which workers will always remain in the modern sector. However, these cutoffs will no longer be equal. Instead, the cutoff above which workers always remain in the modern sector will be strictly greater than the cutoff below which workers always remain in the traditional sector, and workers with abilities between these cutoffs will move between the modern and traditional sectors in response to shocks to their traditional sector productivities. This change does not affect the qualitative responses of unemployment to increases in modern sector productivity or unemployment benefits. Environment and Stationary Equilibrium. For ease of exposition we develop the new features of our model for only one type of labor, rather than two types as in the model of Section 4. The extension to two types is mechanical. We therefore drop the subscripts i = h, l wherever they occur, otherwise retaining the notation and equations from Section 4 except where specified. We assume that the efficiency of any worker in the traditional sector fluctuates between a high state and a low state, taking the value zH > 1 in the high state and zL < 1 in the low state. The marginal product of any worker in the traditional sector therefore fluctuates between AT zH and AT zL . Worker transition between the high and low states follows a Markov chain  1−π with transition matrix Π = π  π 1 − π . It follows that in the stationary equilibrium the share of workers of any ability in each state equals 0.5.8 It is easily shown that, if zL is close enough to 1, a worker with sufficiently low ability will prefer the traditional sector regardless of which state he is in. Similarly, it is easily shown that if zH is close enough to 1, a worker with sufficiently high ability will prefer the modern sector regardless of which state he is in. Flows between the two sectors can therefore occur only for workers with abilities between these two cutoffs. A worker with ability between the two cutoffs prefers to enter the modern sector as unemployed when his traditional sector productivity is low, but not when his traditional sector productivity is high. Now suppose such a worker has entered the modern sector, and transitions from the low to the high traditional productivity state. If he is unemployed at the time, he returns to the traditional sector. If he is employed, however, his decision will depend on how much he is earning, i.e., on his ability. Thus there is a third ability cutoff, above which a worker in the high state prefers the modern sector if he is employed but moves to the traditional sector if he is unemployed. Figure 8: Equilibrium Labor Division across Traditional and Modern Sectors High State 𝑧! Region 0 T Region 1 π 𝑥#∗ U Region 2 T π 𝑠 𝑓 T π E 𝑥%∗ π π 𝑠 U 𝑓 E Region 3 𝑥&∗ U 𝑠 𝑓 E Low State 𝑧" Label the three ability cutoffs x∗1 , x∗2 , and x∗3 , where x∗1 < x∗2 < x∗3 . These three cutoffs divide labor market behavior into four ability regions as shown in Figure 8. Each blue circle represents a set of workers. Workers on the horizontal axis can be in the high or low state of traditional sector productivity, and workers above (below) the horizontal axis are in the high (low) state. Workers with abilities x ∈ [x, x∗1 ) always stay in the traditional sector and workers with abilities x ∈ [x∗3 , x̄] always stay in the modern sector. Workers with abilities 8 If we assume 0.5zH + 0.5zL = 1, then in the stationary equilibrium traditional sector output produced by workers who always remain in the traditional sector is the same as what would be produced by these workers in the model of Section 4. x ∈ [x∗1 , x∗2 ) prefer the traditional sector when their traditional sector productivity is high and the modern sector when their traditional sector productivity is low. Finally, workers with abilities x ∈ [x∗2 , x∗3 ) prefer the modern sector when their traditional sector productivity is low, and when their traditional sector productivity is high they prefer the traditional sector to modern sector unemployment but modern sector employment to the traditional sector. The arrows in Figure 8 show the labor flows between employment and unemployment in the modern sector, the flows between the traditional sector and modern sector employment and unemployment, and the probabilities with which these flows occur. We can use these arrows to compute gross labor flows and unemployment rates in the stationary equilibrium of our modified model. For workers with abilities between x∗1 and x∗2 , gross labor flows in any period equal (0.5π + 0.5π)[G(x∗2 ) − G(x∗1 )] = π[G(x∗2 ) − G(x∗1 )], since half the workers are in each state, all workers in the high (low) state are in the traditional (modern) sector, and all workers in both sectors flow between sectors when they change states. For workers between x∗2 and x∗3 , we must solve for the division of the labor force between the traditional sector, modern sector employment, and unemployment in order to compute the gross labor flows between the modern and traditional sectors and the unemployment rate. Note that only workers in the high state will be in the traditional sector, and only workers in the low state will be unemployed. In contrast, workers employed in the modern sector may be in either the high or low state. Using these facts, we show in Appendix Section C.1 that, for workers with abilities between x∗2 and x∗3 , gross labor flows in any period equal 2π s[0.5(f + s) + π(1 − f − s)] [G(x∗3 ) − G(x∗2 )], (f + s)(π + s − 2sπ) + sπ where f denotes the job finding rate and s denotes the separation rate. The job finding rate f equals ηθ1−α , where θ is the market tightness that applies to all modern sector job-seekers and is determined in general equilibrium as in the model of Section 4. Let u1 , u2 , and u3 be the shares of unemployed workers with abilities within x∗1 and x∗2 , x∗2 and x∗3 , and above x∗3 , respectively. We derive in Appendix Section C.1 that the total unemployment rate u is given by:


u =u1 G(x∗2 ) − G(x∗1 ) + u2 G(x∗3 ) − G(x∗2 ) + u3 1 − G(x∗3 )


0.5[π + s(1 − π)] G(x∗2 ) − G(x∗1 ) s(π + 0.5s − sπ) G(x∗3 ) − G(x∗2 ) s 1 − G(x∗3 ) = + + . π + (f + s)(1 − π) (f + s)(π + s − 2sπ) + sπ f +s To solve the model with stochastic productivity, note that since we have three ability cutoffs, the indifference condition (13) in the model of Section 4 must be replaced by three indifference conditions. To express these we need some additional notation. We denote the value of traditional sector work by T , retaining E and U for the values of modern sector employment and unemployment, respectively. Letting J = H, L denote the state of traditional sector productivity, and k denote region of Figure 8, we can more specifically denote the value functions for traditional sector work, modern sector employment, and unemployment respectively by Tk (zJ ; x), k = 0, 1, 2; Ek (zJ ; x), k = 1, 2, 3; and Uk (zJ ; x), k = 1, 2, 3. The three indifference conditions can then be expressed as T0 (zL ; x∗1 ) = U1 (zL ; x∗1 ), T2 (zH ; x∗2 ) = E2 (zH ; x∗2 ), T2 (zH ; x∗3 ) = U3 (zH ; x∗3 ). At x∗1 , a worker in the low state is indifferent between traditional sector work and entering the modern sector as unemployed. (A worker in the high state with ability x∗1 strictly prefers traditional sector work.) At x∗2 , a worker in the high state is indifferent between modern sector employment and traditional sector work. Finally, at x∗3 a worker in the high state is indifferent between traditional sector work and entering the modern sector as unemployed. In Appendix Section C.2, we also write out the value functions in these three indifference conditions, taking account of the potential labor flows and probabilities in the relevant regions of Figure 8. We then complete the characterization of the solution of the stationary equilibrium. Under the same conditions as in the model of Section 4, skill-biased technological progress will increase the labor force share of the modern sector (reduce x∗3 ) and decrease the labor force share of the traditional sector (reduce x∗1 ). The impact on the share of the labor force that churns between the traditional and modern sectors is ambiguous: some workers who chose the traditional sector when their traditional sector productivity was low will now choose the modern sector and be subject to churning, but some workers who chose the traditional sector when their traditional sector productivity was high will now always choose the modern sector. C.1 Stationary u1 , u2 , u3 and Total Gross Labor Flows For workers with abilities between x∗1 and x∗2 , denote by e1 the share of them who are employed in the modern sector. To solve for the unemployment rate u1 we use the fact that in the stationary equilibrium the inflow and outflow of employed workers are equal: (1 − π)f u1 = (1 − π)se1 + πe1 . Substituting for e1 using the fact that e1 + u1 = 0.5 yields u1 = 0.5[π + s(1 − π)]/[π + (f + s)(1 − π)]. For workers between x∗2 and x∗3 , we have eH + t2 = 0.5 and eL + u2 = 0.5, where eH , eL , t2 , and u2 respectively denote the shares of workers with abilities between x∗2 and x∗3 who are employed in the high state, employed in the low state, in the traditional sector, or unemployed. Within this region, labor flows out of the traditional sector then equal πt2 , and gross labor flows between the modern and traditional sectors equal 2πt2 . To solve for t2 and u2 we use the facts that in the stationary equilibrium, outflows from t2 , eL , and eH equal inflows into t2 , eL , and eH , respectively: πt2 = s(1 − π)eH + sπeL + (1 − f )πu2 , eH − (1 − s)(1 − π)eH = (1 − s)πeL + f πu2 , eL − (1 − s)(1 − π)eL = (1 − s)πeH + f (1 − π)u2 . These conditions automatically guarantee the condition that the inflows into and outflows from the unemployed are equal: πt2 + s(1 − π)eL + sπeH = f u2 + (1 − f )πu2 . Substituting for eH and eL yields two equations in the two unknowns t2 and u2 :   s(π+0.5s−sπ) (f + s)u2 + st2 = s u2 = (f +s)(π+s−2sπ)+sπ ⇒ +s)+π(1−f −s)] (f + s − 1)πu2 + (π + s − πs)t2 = 0.5s t2 = s[0.5(f (f +s)(π+s−2sπ)+sπ We can now compute total gross labor flows between the modern and traditional sectors for workers with ability between x∗1 and x∗3 : Total gross labor flows =π[G(x∗2 ) − G(x∗1 )] + 2πt2 [G(x∗3 ) − G(x∗2 )]. =π[G(x∗2 ) − G(x∗1 )] + 2π s[0.5(f + s) + π(1 − f − s)] [G(x∗3 ) − G(x∗2 )]. (f + s)(π + s − 2sπ) + sπ The labor market for workers with abilities above x∗3 is exactly like the modern sector in the s model of Section 4, hence u3 = f +s . C.2 Value functions and solutions The solution of the stationary equilibrium is characterized by four unknowns: {x∗1 , x∗2 , x∗3 , θ}. Accordingly, we have four equations to solve the four unknowns, including the three indifference conditions T0 (zL ; x∗1 ) = U1 (zL ; x∗1 ), T2 (zH ; x∗2 ) = E2 (zH ; x∗2 ), and T2 (zH ; x∗3 ) = U3 (zH ; x∗3 ), plus one free entry condition. We now write out the value functions in these three indifference conditions, beginning with the two value functions that do not involve flows between the modern and traditional sectors, T0 (zL ; x) and U3 (zH ; x). We have T0 (zL ; x) = AT zL + δ[(1 − π)T0 (zL ; x) + πT0 (zH ; x)], T0 (zH ; x) = AT zH + δ[(1 − π)T0 (zH ; x) + πT0 (zL ; x)]. Solve to obtain T0 (zL ; x) ≡ T0 (zL ) = AT zL δπAT (zH − zL ) + . 1 − δ (1 − δ)(1 − δ − 2δπ) U3 (zH ; x) ≡ U3 (x) is obtained through the same derivation that yields equation (11) in the text: 1 β AM x(1 − b(AM ))(1 − β) U3 (x) = [AM b(AM )x + δf ]. 1−δ 1−β βδf + 1 − δ + δs We solve for U1 (zL ; x) using the following five equations: U1 (zL ; x) = AM bx + δ[(1 − π)f E1 (zL ; x) + (1 − π)(1 − f )U1 (zL ; x) + πT1 (zH ; x)], E1 (zL ; x) = w1 (x) + δ[(1 − π)sU1 (zL ; x) + (1 − π)(1 − s)E1 (zL ; x) + πT1 (zH ; x)], T1 (zH ; x) = AT zH + δ[(1 − π)T1 (zH ; x) + πU1 (zL ; x)], J1 (zL ; x) = AM x − w1 (zL ; x) + δ(1 − s)(1 − π)J1 (zL ; x), J1 (zL ; x) = (1 − β)[E1 (zL ; x) − U1 (zL ; x) + J1 (zL ; x)], where the last two equations use the free entry condition (V = 0) and Nash bargaining, respectively. E2 (zH ; x) and T2 (zH ; x) must be solved simultaneously. We use the following eight equations: E2 (zH ; x) = w2 (zH ; x) + δ[(1 − π)sT2 (zH ; x) + (1 − π)(1 − s)E2 (zH ; x) + πsU2 (zL ; x) + π(1 − s)E2 (zL ; x)], T2 (zH ; x) = AT zH + δ[(1 − π)T2 (zH ; x) + πU2 (zL ; x)], E2 (zL ; x) = w2 (zL ; x) + δ[(1 − π)sU2 (zL ; x) + (1 − π)(1 − s)E2 (zL ; x) + πsT2 (zH ; x) + π(1 − s)E2 (zH ; x)], U2 (zL ; x) = AM bx + δ[(1 − π)f E2 (zL ; x) + (1 − π)(1 − f )U2 (zL ; x) + πf E2 (zH ; x) + π(1 − f )T2 (zH ; x)], J2 (zH ; x) = AM x − w2 (zH ; x) + δ(1 − s)[πJ2 (zL ; x) + (1 − π)J2 (zH ; x)], J2 (zL ; x) = AM x − w2 (zL ; x) + δ(1 − s)[πJ2 (zH ; x) + (1 − π)J2 (zL ; x)], J2 (zH ; x) = (1 − β)[E2 (zH ; x) − T2 (zH ; x) + J2 (zH ; x)], J2 (zL ; x) = (1 − β)[E2 (zL ; x) − U2 (zL ; x) + J2 (zL ; x)].
Finally, we complete the equations needed for the solution by expressing E J(x|x ∈ Modern)
in the simplified free entry condition AM c = δηθ−α E J(x|x ∈ Modern) :
E J(x|x ∈ Modern) = u1 (G(x∗2 )−G(x∗1 ))EJ1 (zL ;x)+u2 (G(x∗3 )−G(x∗2 ))EJ2 (z;x)+u3 (1−G(x∗3 ))EJ3 (x) u1 (G(x∗2 ) − G(x∗1 )) + u2 (G(x∗3 ) − G(x∗2 )) + u3 (1 − G(x∗3 )) , R x∗2 where EJ1 (zL ; x) = R x∗3 x∗ 2 J3 (x)g(x)dx 1−G(x∗3 ) . x∗ 1 J1 (zL ;x)g(x)dx G(x∗3 )−G(x∗2 ) R x∗3 , EJ2 (z; x) = x∗ 2 e J (z ;x) eL J2 (zL ;x) + He 2+eH eL +eH L H G(x∗3 )−G(x∗2 )
g(x)dx , and EJ3 (x) = D Appendix Figures and Tables Figure D1: High- and Low-educated Unemployment when AM Increases Unemployment Rate Low-Education High-Education 𝐴% Note: This figure illustrates comparative statics of unemployment rates in AM . The dashed line corresponds to the AM at which Panel (b) of Figure 4 begins. 100 Figure D2: Low-Education Share, λ, in Model and Data MLI MWI KHM Low Educational Share (Percent) 20 40 60 80 HTI IDN GHA ZMB BOL SLV IND MYS MAR NICDOM TUR BRA CRI URY MEX PRT FJI COL ECUARG CHLVEN NGA PAN ZAF ESP JAM GRC PSE FRA IRL PER CAN ROU USA 0 Model 6 7 8 9 ln(GDP per capita) 10 11 Note: This figure plots the values of λ used in the quantitative experiments of Section 5 (solid line), and the percent of the labor force that is low-educated in each of our countries (dots with identifiers). The data come from IPUMS. Low-educated individuals are defined to be those with less than a secondary school education. 54 Traditional Sector Size (Percent) for High Education 0 20 40 60 80 100 Figure D3: Traditional-Sector Share by Education Model KHM IND UGAMLI ZMB GHA BOL SLV Traditional Sector Size (Percent) for Low Education 0 20 40 60 80 100 6 7 PER ECU JAM TUR MEX DOM HUN PAN BRA ROU CHL VEN ZAF MYS PRT MAR 8 9 ln(GDP per capita) CHE IRL AUT USA ESPFRA CAN 10 11 MLI KHM UGA GHA ZMB BOL Model TUR PER ECU IND MAR SLV JAM ROU PAN MEX BRA MYS DOM CHL VEN PRT HUN ZAF 6 7 8 9 ln(GDP per capita) 10 ESP IRL FRA AUT USA CAN CHE 11 Note: This figure plots the size of the traditional sector against log GDP per capita in the data and model. Each dot represents the average in a country with at least two observations across all years of data, and the solid line is the prediction of the quantitative model. The top panel is for high-educated workers, and the bottom is for low-educated workers. Table D1: Home Production of the Unemployed Income Quartile Cooking Cleaning Childcare Shopping Collecting water and firewood Total Hours Bottom 6.8 Top 5.5 (3) (5) 2.1 7.1 (1) (5) 5.6 4.1 (2) (5) 1.7 3.5 (1) (5) 3.5 - (3) (0) 19.7 20.2 Note: This table reports average weekly hours by the unemployed using data from the American Time Use Survey and Multinational Time Use Surveys. Average weekly hours for each activity are computed for the bottom and top quartile countries in which data have been collected. The number of countries is in parentheses. Observations in the bottom quartile include Ghana in 1998, Malawi in 2005, Uganda in 2005, and the top quartile Canada in 2005, France in 1998, Spain in 2002, the UK in 2014, the US in 2005. Table D2: Definition of Traditional Sector Goods Item Details Shoe Repair - Women Street Shoes Replacement of 2 heels (glued and nailed); While-you-wait in shop service; Heel: Synthetic polyurethane, small heel. Shoe Repair - Men Classic Shoes Re-soling rubber soles (glued & nailed or stitched); Not “urgent” in shop service. Shoeshine Cleaning leather shoes with a brush and polishing; Manual work while keeping the shoes on; Exclude service in a shop. Taxi 7 km in the town center on working days at 3 p.m.; Includes: Possible fixed starting fee + price per km; Excludes: Taxi called by telephone. Men basic haircut Scissor cut of short hair for male adults; Type of establishment: Common men’s barber shop; No shampoo/washing nor styling/fixing products; Full price including tips if any. Ladies haircut - curlers Hair with curlers cut to medium (basic) for female adult; Shampoo/washing, blow drying, and styling/fixing products; Establishment: Common hairdresser (exclude hair stylist). Manicure Standard manicure on natural nails by nail technician; Establishment: Professional beautician; Full price including tips if any; Bath, filing, cuticles treatment, one-color varnishing. Ladies haircut - long hair Long hair cut to short for female adult; Shampoo/washing, blow drying, styling/fixing products; Establishment: Common hairdresser (exclude hair stylist). Note: The table reports the definitions of each ICP traditional service used in Table 6, and described in Section 5.1. The services come from the unpublished ICP 2011 Global Core list of goods and services. Table D3: Different Elasticity 1 1−σ but Matching Traditional Sector Slope Model Elasticity Slope Coefficients Data Lower Benchmark Aggregate traditional sector share -19.9 -19.7 -20.0 Traditional-sector share for low educated -21 -19.0 -19.3 Traditional-sector share for high educated -5.1 -10.0 -10.5 Aggregate unemployment rate 1.8 1.10 1.2 Unemployment rate for low-educated 3.4 2.4 2.4 Unemployment rate for high-educated .58 .58 .60 ul Ratio of unemployment rates uh .47 .37 .38 Relative price PT .60 .78 .63 Calibrated Elasticity of AT to AM , ψ1 0.09 0.19 1 1−σ Higher -19.7 -19.0 -10.3 1.14 2.3 .59 .37 .52 0.27 Note: This table reports the slope coefficients from regressions of the statistics in each row on log GDP per capita and a constant. The second column (Data) reports the slopes from our cross-country database, the third column (Lower) is for an elasticity of substitution between modern and traditional output of 3, the fourth column (Benchmark) is the benchmark model with an elasticity of 3.5, and the fifth column (Higher) is for an elasticity of 4.