Remittance Flows under Asymmetric Information: A Contracting Approach

Transcription

1 Remittance Flows under Asymmetric Information: A Contracting Approach Ganesh Seshan 1 Robertas Zubrickas 2 1 Georgetown University School of Foreign Service in Qatar 2 University of Zurich 3 June 2014 R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

2 Motivation R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

3 Motivation Asymmetry in information about migrant earnings matters for remittances. R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

4 Motivation Asymmetry in information about migrant earnings matters for remittances. McKenzie et al. (2013), Ambler (2013), Ashraf et al. (2014) R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

5 Motivation Asymmetry in information about migrant earnings matters for remittances. McKenzie et al. (2013), Ambler (2013), Ashraf et al. (2014) We explore this observation further, with interest in R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

6 Motivation Asymmetry in information about migrant earnings matters for remittances. McKenzie et al. (2013), Ambler (2013), Ashraf et al. (2014) We explore this observation further, with interest in magnitude of asymmetry R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

7 Motivation Asymmetry in information about migrant earnings matters for remittances. McKenzie et al. (2013), Ambler (2013), Ashraf et al. (2014) We explore this observation further, with interest in magnitude of asymmetry its effects on remittance behavior. R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

8 Motivation Asymmetry in information about migrant earnings matters for remittances. McKenzie et al. (2013), Ambler (2013), Ashraf et al. (2014) We explore this observation further, with interest in magnitude of asymmetry its effects on remittance behavior. Analysis is empirical and theoretical. R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

9 Asymmetry in Information R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

10 Asymmetry in Information Asymmetry is assessed by contrasting migrants earning reports R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

11 Asymmetry in Information Asymmetry is assessed by contrasting migrants earning reports... with reports about their earnings made by their wives. R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

12 Asymmetry in Information Asymmetry is assessed by contrasting migrants earning reports... with reports about their earnings made by their wives. Sample: 108 households from Kerala, with husbands working in Qatar. R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

13 Asymmetry in Information Asymmetry is assessed by contrasting migrants earning reports... with reports about their earnings made by their wives. Sample: 108 households from Kerala, with husbands working in Qatar. Our findings are R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

14 Asymmetry in Information Asymmetry is assessed by contrasting migrants earning reports... with reports about their earnings made by their wives. Sample: 108 households from Kerala, with husbands working in Qatar. Our findings are Wives report only 79% of their husbands income R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

15 Asymmetry in Information Asymmetry is assessed by contrasting migrants earning reports... with reports about their earnings made by their wives. Sample: 108 households from Kerala, with husbands working in Qatar. Our findings are Wives report only 79% of their husbands income Discrepancy increases with migrant earnings. R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

16 Discrepancy in Reports Reported earnings gap Migrant's annual earnings report R. Zubrickas The regression (U of Zurich) line ( solid red Asymmetry ) is shown and with Remittance its 95 % confidence Flows band ( green dashed 3 June lines 2014 ). 4 / 13

17 Asymmetry and Remittances R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

18 Asymmetry and Remittances Income share of remittances Migrant's annual earnings report The regression line ( solid red ) is shown with its 95 % confidence band ( green dashed lines ). R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

19 Reported Remittance Shares R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

20 Reported Remittance Shares Income share of remittances Migrant's annual earnings report non-parametric plot of migrant's reported income share of remittance non-parametric plot of wife's reported income share of remittance R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

21 Theoretical Approach R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

22 Theoretical Approach Our empirical findings R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

23 Theoretical Approach Our empirical findings demonstrate a significant degree of asymmetry R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

24 Theoretical Approach Our empirical findings demonstrate a significant degree of asymmetry call for the Costly State Verification approach (Townsend, 1979). R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

25 Theoretical Approach Our empirical findings demonstrate a significant degree of asymmetry call for the Costly State Verification approach (Townsend, 1979). Justification: R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

26 Theoretical Approach Our empirical findings demonstrate a significant degree of asymmetry call for the Costly State Verification approach (Townsend, 1979). Justification: verifications are feasible R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

27 Theoretical Approach Our empirical findings demonstrate a significant degree of asymmetry call for the Costly State Verification approach (Townsend, 1979). Justification: verifications are feasible existence of credible punishments R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

28 Theoretical Approach Our empirical findings demonstrate a significant degree of asymmetry call for the Costly State Verification approach (Townsend, 1979). Justification: verifications are feasible existence of credible punishments implicit remittance contracts R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

29 Theoretical Approach Our empirical findings demonstrate a significant degree of asymmetry call for the Costly State Verification approach (Townsend, 1979). Justification: verifications are feasible existence of credible punishments implicit remittance contracts We extend the exchange remittance model with asymmetric information and CSV. R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

30 Optimal Remittance Contract R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

31 Optimal Remittance Contract The recipient offers the migrant a contract (r(ŷ), S, p(y, ŷ)) R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

32 Optimal Remittance Contract The recipient offers the migrant a contract (r(ŷ), S, p(y, ŷ)) y is income, r remittance, S verification region, p penalty. R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

33 Optimal Remittance Contract The recipient offers the migrant a contract (r(ŷ), S, p(y, ŷ)) y is income, r remittance, S verification region, p penalty. Theorem S = [y, r), r(y) = min(y, r). R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

34 Optimal Remittance Contract The recipient offers the migrant a contract (r(ŷ), S, p(y, ŷ)) y is income, r remittance, S verification region, p penalty. Theorem S = [y, r), r(y) = min(y, r). We also show that the threshold r is lower when R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

35 Optimal Remittance Contract The recipient offers the migrant a contract (r(ŷ), S, p(y, ŷ)) y is income, r remittance, S verification region, p penalty. Theorem S = [y, r), r(y) = min(y, r). We also show that the threshold r is lower when the cost of verification is higher, R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

36 Optimal Remittance Contract The recipient offers the migrant a contract (r(ŷ), S, p(y, ŷ)) y is income, r remittance, S verification region, p penalty. Theorem S = [y, r), r(y) = min(y, r). We also show that the threshold r is lower when the cost of verification is higher, the recipient holds more optimistic beliefs. R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

37 Theoretical Predictions R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

38 Theoretical Predictions We define the reported earnings ratio by ω = y w y h ( = ŷ y ) R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

39 Theoretical Predictions We define the reported earnings ratio by ω = y w y h ( = ŷ y ) The model predicts R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

40 Theoretical Predictions We define the reported earnings ratio by ω = y w y h ( = ŷ y ) The model predicts 1 ω decreases with income. R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

41 Theoretical Predictions We define the reported earnings ratio by ω = y w y h ( = ŷ y ) The model predicts 1 ω decreases with income. 2 ω is lower when (i) there is more asymmetry in information, (ii) the recipient holds more pessimistic beliefs. R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

42 Theoretical Predictions We define the reported earnings ratio by ω = y w y h ( = ŷ y ) The model predicts 1 ω decreases with income. 2 ω is lower when (i) there is more asymmetry in information, (ii) the recipient holds more pessimistic beliefs. 3 The lower the value of ω, the lower the remittances, and vice-versa. R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

43 Theoretical Predictions We define the reported earnings ratio by ω = y w y h ( = ŷ y ) The model predicts 1 ω decreases with income. 2 ω is lower when (i) there is more asymmetry in information, (ii) the recipient holds more pessimistic beliefs. 3 The lower the value of ω, the lower the remittances, and vice-versa. 4 The remittance schedule is steeper at lower levels of income. R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

44 Determinants of the Reported Earnings Ratio R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

45 Determinants of the Reported Earnings Ratio (1) (2) Log(migrant's(annual(earnings((US$) *** 7 (0.06) Ratio(of(wife(to(husband's(account(of(annual(remittance( (0.03) (0.03) Log(median(montly(income(of(Gulf(migrants((US$) 0.138* 0.213** (0.07) (0.08) Wife(is(employed((Indicator) * 0.137** (0.05) (0.06) Only(migrant(decides(on(money(matters((Indicator) 0.139** 0.101* (0.05) (0.06) Wife(is(unaware(of(husband's(expenses((Indicator) ** * (0.05) (0.06) Disagreement(over(remittance(use((Indicator) ** ** (0.04) (0.04) Migrant(spends(on(temptation(goods((indicator) *** *** (0.05) (0.04) Constant R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

46 Determinants of Log Annual Remittance R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

47 Determinants of Log Annual Remittance Migrant's,Report (1) (2) Reported,earnings,ratio *** 2 (0.21) Log,migrant's,annual,earnings,(US$) 0.532** 0.737*** (0.24) (0.23) Hhold,has,other,members,,working,abroad,(indicator) *** *** (0.12) (0.10) Hhold's,annual,medical,expenditure,as,a,share,of,migrant's,income 0.679** 0.664** (0.26) (0.26) Log,household's,monthly,income,in,India,(US$) ** * (0.02) (0.02) Log,median,montly,income,of,Gulf,migrants,(US$) 0.780** 0.679** (0.32) (0.28) Wife,is,employed,(indicator) Only,migrant,decides,on,money,matters,(Indicator) R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / (0 0.7 ( (0 0.6 (0 0 (0 0. (0 0 (0 0 (0

48 Piece-Wise Regression for Annual Remittance R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

49 Piece-Wise Regression for Annual Remittance (1) (2) Migrant's0annual0earnings0(US$) 0.255** 0.547*** 0 (0.11) (0.13) Migrant's0annual0earnings0beyond0income0threshold0(US$) C0.455** C (0.17) Log0median0montly0income0of0Gulf0migrants0(US$) 1 Wife0is0employed0(indicator) Only0migrant0decides0on0money0matters0(Indicator) Wife0is0unaware0of0husband's0expenses0(Indicator) R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

50 Conclusions R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

51 Conclusions Asymmetry in information within the transnational household is substantial. R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

52 Conclusions Asymmetry in information within the transnational household is substantial. Asymmetry is a major determinant of remittance behavior R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

53 Conclusions Asymmetry in information within the transnational household is substantial. Asymmetry is a major determinant of remittance behavior... migrants strategically exploit their private information. R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

54 Conclusions Asymmetry in information within the transnational household is substantial. Asymmetry is a major determinant of remittance behavior... migrants strategically exploit their private information. The CSV approach is useful in explaining empirical patterns. R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

55 Conclusions Asymmetry in information within the transnational household is substantial. Asymmetry is a major determinant of remittance behavior... migrants strategically exploit their private information. The CSV approach is useful in explaining empirical patterns. Future work should R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

56 Conclusions Asymmetry in information within the transnational household is substantial. Asymmetry is a major determinant of remittance behavior... migrants strategically exploit their private information. The CSV approach is useful in explaining empirical patterns. Future work should revisit remittance models with the symmetry assumption R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13

57 Conclusions Asymmetry in information within the transnational household is substantial. Asymmetry is a major determinant of remittance behavior... migrants strategically exploit their private information. The CSV approach is useful in explaining empirical patterns. Future work should revisit remittance models with the symmetry assumption exploit panel/experimental data to examine how varying information conditions influences remittance flows. R. Zubrickas (U of Zurich) Asymmetry and Remittance Flows 3 June / 13