1

Returns to Schooling in Egypt

ERF Project no. 2014-072

Murat G. KırdarBoğaziçi University, Istanbul, Turkey

Abdurrahman AydemirSabanci University, Istanbul, Turkey

Meltem DayıoğluMiddle East Technical University, Ankara, Turkey

Ragui AssaadUniversity of Minnesota, USA

2

AIM:

Estimate the causal effect of schooling on earnings in Egypt

Identification:

Reduction in the duration of primary school from 6 to 5 years in 1988

3

Motivation and Literature:

Developed countries:• Angrist and Krueger (1991): quarter-of-birth as an instrument for

completed years of schooling.

• Acemoglu and Angrist (2001), Lochner and Moretti (2004), Lleras-Muney (2005): Regional variation in schooling laws in the US

• Harmon and Walker (1995), Oreopoulos (2006), Devereux and Hart (2010) : Changes in compulsory schooling duration in the U.K. in 1947 and in 1973

4

Motivation and Literature (cont.):

Estimated returns vary widely:

• Devereux and Hart (2010): No evidence of any positive return for women and a return for men of 4–7% in the UK

• Pischke and von Wachter (2008): Zero return in Germany

• Oreopoulos (2006): 15% in the UK

5

Motivation and Literature (cont.):

A much smaller literature in developing countries:

• Duflo (2001): Massive school construction program in Indonesia in 70s.• Spohr (2003): Change in duration of compulsory schooling in Taiwan • Fang et al. (2012): Change in duration of compulsory schooling in China• Eble and Hu (2015) : Increase in duration of primary schooling by one

year in China (5 to 6yrs)• Aydemir and Kirdar (2014): Change in duration of compulsory schooling

in Turkey (1987)

Estimated returns to schooling vary widely: 0% (Turkey) – 20% (China)

6

Contribution:

Contribute to the returns to schooling literature by estimating the causal effect of schooling on earnings in Egypt based on an exogenous source of variation.

7

Education System in Egypt:

• 6+3+3 system: 6 years of primary schooling, 3 years of lower secondary and 3 years of upper secondary schooling

• Up until 1980: Compulsory schooling = 6 yrs• 1981: Compulsory schooling increased from 6 to 9 yrs (6+3)• 1988: Primary schooling reduced from 6 to 5 yrs, hence compulsory

schooling reduced from 9 t o 8 yrs.• 1999: Primary schooling increased from 5 to 6 yrs, hence compulsory

schooling increased from 8 to 9 yrs.

8

Figure 1. Proportion of adult population by school attainment

• a bimodal distribution

9

Data:

2012 Egypt Labor Market Panel Survey (ELMPS)• Provides detailed schooling and labor market information

• 49,186 individuals from 12,060 households

• Operational sample: Persons between the ages of 15/20 and 45

• Upper age limit dictated by data, detailed educ. questions in 2012 asked of individuals younger than 45.

• Also used 2006 ELMPS to add info to 2012 concerning schooling of individuals who were interviewed in 2006 and 2012.

10

Generating the years of school variable:

• Know the last schooling level completed

• Do not know who completed 5, who 6 yrs of primary schooling

• Use year started school and no. of repeated grades to find the year in which the person reached 5th grade in primary school.

• Based on the year the person reached 5th grade determine whether went through 5 or 6 yrs of primary schooling.

11

Methodology:

2SLS estimation:

where,s denotes yrs of schooling, w hourly wages, D dummy variable for policy, X region of birth, controls for age/birth cohort dummies.

– key parameter of interest - denotes the percent change in wages when years of schooling is raised by one.

12

Methodology cont:

• Policy dummy: 1 for those born in 1978 or later, 0 for others

• Dropped those born in 1977 – partially treated

• Interesting feature of Egyptian policy: reduction rather than an increase in compulsory schooling.

• No complier problem! Cannot choose between 5 and 6 years of primary school. Though could drop out of schooling.

• No sheep-skin effect: the same diploma – say high school diploma - granted though some has 11 yrs some 12 yrs of schooling.

13

6

8

10

12

14

16

Ave

rag

e Y

ea

rs o

f S

ch

oo

ling

70 72 74 76 78 80 82 84

Year of Birth

Figure 1: Average years of schooling by birth cohort – wage earners ( incl. 1977 birth cohort)

14

6

8

10

12

14

16A

ve

rag

e Y

ea

rs o

f S

ch

oo

ling

70 72 74 76 78 80 82 84

Year of Birth

Figure 2: Average years of schooling by birth cohort - wage earners ( excludes 1977 birth cohort)

15

First Stage Results: Wage workers

(1) (2) (3) (4) (5) (6) VARIABLES All age

[20 45] All age [25 45]

Male age [20 45]

Male age [25 45]

Female age [20 45]

Female age [25 45]

treated -1.918*** -1.936*** -2.079*** -2.221*** -0.716 0.030 [0.350] [0.453] [0.341] [0.436] [0.820] [1.036] birthday -19.429*** -21.508* -14.679*** -20.835* -30.851** -10.901 [4.903] [11.896] [5.150] [12.542] [12.433] [24.656] birthday2 0.255*** 0.283* 0.194*** 0.275* 0.401** 0.143 [0.062] [0.154] [0.065] [0.163] [0.160] [0.322] birthday3 -0.001*** -0.001* -0.001*** -0.001* -0.002** -0.001 [0.000] [0.001] [0.000] [0.001] [0.001] [0.001] 2.region -0.367 -0.322 -0.571** -0.471 0.523 0.301 [0.247] [0.267] [0.288] [0.311] [0.357] [0.384] 3.region -0.612** -0.563** -0.850*** -0.788*** 0.170 0.140 [0.238] [0.259] [0.268] [0.293] [0.447] [0.454] 4.region -0.236 -0.243 -0.610** -0.632** 0.983*** 0.735* [0.215] [0.241] [0.245] [0.274] [0.375] [0.410] 5.region -1.759*** -1.921*** -1.758*** -1.880*** -1.043** -1.443*** [0.202] [0.226] [0.223] [0.248] [0.405] [0.459] 6.region -3.021*** -3.035*** -2.948*** -2.978*** -1.114* -1.189* [0.245] [0.276] [0.264] [0.295] [0.593] [0.630] Observations 7,189 5,976 6,017 4,973 1,172 1,003 R-squared 0.071 0.067 0.065 0.063 0.049 0.043

16

Variables All

Ages 20-45

All Ages 25-45

Treated -0.634** -1.081*** -1.918*** -0.833*** -1.308*** -1.317*** -1.936*** -1.302***

[0.280] [0.317] [0.350] [0.295] [0.316] [0.318] [0.453] [0.312]

Birthday 0.015 0.073*** 0.208*** 0.108** 0.091*** 0.094*** 0.215*** 0.108**

[0.020] [0.027] [0.038] [0.044] [0.026] [0.027] [0.070] [0.044]

Birthday squared -0.006*** 0.000 -0.002 0.000

[0.002] [0.003] [0.002] [0.003]

Birthday cubed -0.001*** -0.001*

[0.000] [0.001]

Birthday*treated -0.124** -0.031

[0.049] [0.054]

Obs 7,189 7,189 7,189 7,189 5,976 5,976 5,976 5,976

R-squared 0.065 0.068 0.071 0.067 0.065 0.066 0.067 0.066

First Stage Results: Wage workers - ALL

17

Variables Men

Ages 20-45

Men Ages 25-45

Treated -0.989*** -1.442*** -2.079*** -1.207*** -1.604*** -1.622*** -2.221*** -1.597***

[0.273] [0.312] [0.341] [0.287] [0.314] [0.316] [0.436] [0.311]

Birthday 0.035* 0.094*** 0.197*** 0.134*** 0.105*** 0.110*** 0.228*** 0.133***

[0.020] [0.029] [0.039] [0.047] [0.028] [0.029] [0.071] [0.047]

Birthday squared -0.006*** -0.001 -0.003 -0.001 -0.05

[0.002] [0.003] [0.003] [0.003] [0.057]

Birthday cubed -0.001*** -0.001*

[0.000] [0.001]

Birthday*treated -0.130** -0.05

[0.051] [0.057]

Obs 6,017 6,017 6,017 6,017 4,973 4,973 4,973 4,973

R-squared 0.06 0.063 0.065 0.062 0.062 0.062 0.063 0.062

First Stage Results: Wage workers - MEN

18

Variables Women

Ages 20-45

Women Ages 25-45

Treated 0.985 0.626 -0.716 0.861 0.359 0.360 0.030 0.372

[0.603] [0.665] [0.820] [0.636] [0.665] [0.662] [1.036] [0.641]

Birthday -0.063 -0.013 0.202** 0.015 0.002 0.003 0.067 0.015

[0.045] [0.055] [0.099] [0.092] [0.054] [0.056] [0.156] [0.091]

Birthday squared -0.006 0.003 -0.001 -0.001

[0.004] [0.005] [0.005] [0.005]

Birthday cubed -0.002** -0.001

[0.001] [0.001]

Birthday*treated -0.109 -0.025

[0.105] [0.109]

Obs 1,172 1,172 1,172 1,172 1,003 1,003 1,003 1,003

R-squared 0.037 0.041 0.049 0.039 0.042 0.043 0.043 0.042

First Stage Results: Wage workers - WOMEN

19

(1) (2) (3) (4) (5) (6) VARIABLES All age

[20 45] All age [25 45]

Male age [20 45]

Male age [25 45]

Female age [20 45]

Female age [25 45]

treated -1.957*** -1.854*** -2.018*** -2.071*** -1.390 -0.375 [0.363] [0.459] [0.343] [0.430] [0.870] [1.088] birthday -19.557*** -16.941 -14.446*** -17.291 -35.718*** -5.135 [4.994] [11.883] [5.212] [12.751] [12.600] [23.581] birthday2 0.256*** 0.223 0.191*** 0.228 0.463*** 0.066 [0.063] [0.154] [0.066] [0.165] [0.162] [0.307] birthday3 -0.001*** -0.001 -0.001*** -0.001 -0.002*** -0.000 [0.000] [0.001] [0.000] [0.001] [0.001] [0.001] 2.region -0.581** -0.516* -0.783*** -0.661** 0.439 0.204 [0.250] [0.271] [0.290] [0.314] [0.375] [0.408] 3.region -0.770*** -0.706*** -1.004*** -0.910*** 0.057 -0.031 [0.242] [0.264] [0.274] [0.299] [0.466] [0.467] 4.region -0.338 -0.328 -0.757*** -0.758*** 1.021*** 0.750* [0.222] [0.246] [0.254] [0.281] [0.383] [0.415] 5.region -1.822*** -1.955*** -1.876*** -1.954*** -0.851** -1.288*** [0.206] [0.229] [0.226] [0.250] [0.420] [0.463] 6.region -3.081*** -3.091*** -3.037*** -3.055*** -1.127* -1.217* [0.252] [0.284] [0.272] [0.305] [0.624] [0.660] Observations 6,698 5,575 5,619 4,653 1,079 922 R-squared 0.072 0.067 0.067 0.064 0.047 0.037

First Stage Results: Full-time wage workers (30hrs+ per week)

20

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)

VARIABLES Birthyear >=1973

Birthyear >=1974

Birthyear >=1975

Birthyear >=1976

Birthyear >=1977

Birthyear >=1978

Birthyear >=1979

Birthyear >=1980

Birthyear >=1981

Birthyear >=1982

Birthyear >=1983

Treated 0.41 0.31 1.10* -0.65 -2.22*** -2.22*** -0.19 0.60 0.78** 0.57 0.72**

[0.77] [0.43] [1.88] [-0.58] [-5.69] [-5.69] [-0.19] [1.25] [2.36] [1.53] [2.71]

Observations 4,973 4,973 4,973 4,973 4,973 4,973 4,973 4,973 4,973 4,973 4,973

R-squared 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06

First Stage Results: Alternative definitions of policy (i.e. treatment dummy), males 25-45Falsification test: = 1 if birth year>=1978+i , i= -5...+5

21

(1) (2) (3) (4) (5) (6) VARIABLES All education

groups Less than post secondary

Less than high school

Less than middle school

Less than elementary

Less than literate

Treated -2.21*** -1.41*** -1.53*** -1.19*** -0.48 -0.52 [-5.64] [-5.54] [-3.27] [-2.99] [-1.25] [-0.95] Observations 4,971 3,647 1,619 1,349 843 654 R-squared 0.06 0.03 0.12 0.11 0.09 0.08

First Stage Results: Males 25-45, Samples restricted to different education groups

22

Second Stage Results: Wage outcome-male wage earners

(1) (2) (3) (4) VARIABLES OLS-20_45 IV2-20_45 OLS-25_45 IV2-25_45 Yrs of sch. 0.025*** 0.038* 0.028*** 0.066*** [0.004] [0.023] [0.004] [0.022] First stage F-value

39.54 32.40

Observations 6,286 6,017 5,242 4,973 R-squared 0.081 0.077 0.081 0.023 Full time workers

tot_schyrs3 0.026*** 0.035 0.029*** 0.064*** [0.004] [0.023] [0.004] [0.023] First stage F-value 32.63 25.53

Observations 5,875 5,619 4,909 4,653 R-squared 0.090 0.091 0.089 0.043

23

Second Stage Results: Wage outcome- private sector wage earners - males

(1) (2) (3) (4) VARIABLES OLS-20_45 IV2-20_45 OLS-25_45 IV2-25_45 Panel A- Private Sector Employees

Yrs of sch. 0.014*** 0.013 0.017*** 0.050 [0.003] [0.033] [0.004] [0.036] First stage F values

35.59 28.43

Observations 4,654 4,467 3,688 3,501 R-squared 0.051 0.053 0.047 0.000

24

Second Stage Results: Wage outcome-public sector wage earners - males

(1) (2) (3) (4) VARIABLES OLS-20_45 IV2-20_45 OLS-25_45 IV2-25_45 Panel B- Public Sector Employees

Yrs of sch. 0.069*** 0.125** 0.070*** 0.133** [0.004] [0.053] [0.004] [0.062] Observations 1,632 1,550 1,554 1,472 R-squared 0.196 0.118 0.191 0.089 First satge F values

12.03 6.42

25

Conclusion

• This paper estimates the causal impact of schooling on wages in Egypt using an IV framework.

• First stage results: Drop in number of schooling years completed among the affected cohorts of men. No first stage for women.

• Robustness checks to show observed changes are due to the change in the schooling law.

• Second stage results: Returns to schooling for men ages 25-45 is estimated at 6.6 percent.