What Can We Learn about Financial Access from U.S. Immigrants?
Una Okonkwo Osili Indiana University Purdue University Indianapolis Anna Paulson Federal Reserve Bank of Chicago
*These are the views of the authors and not necessarily those of the Federal Reserve Bank of Chicago or the Federal Reserve System.
Why is Financial Access Important? There is substantial variation both across and within countries in the fraction of individuals who use formal financial services. Financial development accelerates growth, decreases poverty and inequality (Rajan and Zingales (1998), Levine (2005), Beck, Demirgüç-Kunt and Levine (2004), Clarke, Xu and Zou (2003)). Greater financial development allows firms and individuals to realize growth opportunities, to take advantage of new technologies (Beck, Demirgüç-Kunt and Levine, forthcoming).
Why study U.S. immigrants? Important, and growing, segment of the U.S. population Isolates some confounding factors in cross-county studies: • Same competitive environment for everyone, same regulatory environment, same legal structure, same infrastructure -- Supply factors are (to a first approximation) controlled for. Focuses attention on how these supply factors influence preferences and beliefs: • Can observe how time spent in a particular institutional environment influences preferences and beliefs and hence usage.
Empirical issues • Immigrants are not random representatives of their countries of origin. They typically choose to migrate. • Unobservable characteristics may be correlated with country of origin characteristics
Measuring Financial Access • Breadth (Bisj): any relationship with a bank (savings or checking account) • Depth (Disj): how many dimensions of financial markets are used: o Risk-free savings (savings, CD) o Payment services (checking, money market account) o Investment vehicles (stock, IRA, Keogh)
Data Individual data: 1996 – 2000 U.S. Survey on Income and Program Participation (SIPP), 36,000 individuals, 11% are immigrants from 75 countries. Includes information on wealth, income, education, race, labor force status, age, family structure, location, occupation, as well as financial market participation. Country data: institutional quality, gdp per capita, measures of financial development, legal origin, infrastructure, religion from a variety of sources.
Table 2A: Characteristics of Immigrants and the Native Born in the SIPP Data
Characteristic Individual Characteristics Age % Male % Married % non-white % unemployed or out of the labor force # of children < 18 in household Average monthly per capita hh income Median monthly per capita hh income Average household wealth 25th percentile of household wealth Median household wealth 75th percentile of household wealth Educational Attainment (%) Less than High School High School Graduate Some College Bachelor Degree Advanced Degree Financial Market Participation % who own stock % with a checking account % with a savings account % with banking relationship average financial depth median financial depth Number of Individuals Number of Observations Native Born 46.47 (17.52) 45.6% 57.4% 16.4% 33.8% 0.720 (1.090) $2,224.44 (2,832.45) $1,578 $185,754 (1,398,146) $14,660 $71,123 $186,512 15.0% 30.4% 30.6% 15.9% 8.1% 20.0% 63.8% 54.8% 76.3% 1.71 (1.02) 2.00 31,046 100,839 Immigrant 45.22 (16.51) 46.2% 65.6% 32.2% 36.7% 1.118 (1.347) $1,639.53 (2,375.44) $1,050 $122,685 (978,910) $3,017 $29,001 $117,917 35.8% 24.5% 20.1% 12.5% 7.1% 8.6% 47.0% 40.1% 61.0% 1.22 (1.01) 1.00 5,020 15,043
Table 3A: Summary of Country Variables
Characteristic Measures of Institutional Quality Protection from Expropriation British Legal Origin Latitude Other Country Characteristics English Speaking GDP Per Capita Catholic Protestant Muslim Internet Usage Banking Freedom Bank Concentration N 79 79 79 79 74 74 74 74 77 48 48 Mean 7.50 0.29 0.33 0.139 8,704 39.29 10.43 16.82 4.46 3.34 0.52 Std Dev. 1.74 0.46 0.19 0.35 10,376 40.01 19.39 33.77 8.40 0.73 0.18 Min 1.81 0.00 0.01 0.00 106 0.00 0.00 0.00 0.00 1.60 0.18 Med 7.51 0.00 0.33 0.00 3,208 27.2 0.35 0.00 .43 3.30 0.51 Max 10.00 1.00 0.71 1.00 42,873 97.00 87.00 100.00 41.77 5.00 0.97 U.S. value 10.00 1.00 0.42 1.00 24,831 24.00 52.00 1.00 33.96 4.00 0.18
Table 3B: Correlation between Country Variables
Characteristic Protection from Expropriation British Legal Origin Latitude Internet Usage Banking Freedom Bank Concentration Protection From Exp ---0.119 0.572*** 0.607*** 0.334** -0.296** British Legal Origin ----0.200* 0.114 0.046 0.006 Latitude Internet Access Banking Freedom Bank Concentration
---0.530*** 0.283** -0.069
---0.312** -0.173
----0.335**
----
Empirical Specification Bisj or Disj = α + β1Xi + β2Zj + δs + εisj Xi – individual characteristics (wealth, income, education, age, sex, marital status, kids, labor force status) Zj – country of origin characteristics (institutional quality, legal origin, geography, gdp, religion, continent, financial development) δs – MSA fixed effects Standard errors are clustered at the country of origin level and adjusted for heteroscedasticity when we estimate Bisj
Appendix Table 2: The Effect of Control Variables on the Probability of Having a Bank Relationship and Depth of Financial Market Participation
Probability of Having a Bank Relationship [1] Depth of Financial Market Participation [2]
Explanatory Variable
Age†
Age Squared† 2nd Wealth Quartile 3rd Wealth Quartile 4th Wealth Quartile Unemployed or Out of Labor Force Per Capita Income†† Per Capita Income Squared†† Male Married Number of Children Non-white High School Graduate Some College Bachelor Degree Advance Degree Protection from Expropriation Constant MSA Controls Adjusted R-Squared Number of Observations
0.704 (0.153) -0.004 (0.002) 0.183 (0.015) 0.173 (0.017) 0.168 (0.019) -0.086 (0.013) 37.900 (8.540) -0.001 (0.000) -0.050 (0.007) 0.163 (0.014) -0.022 (0.005) 0.019 (0.018) 0.126 (0.015) 0.187 (0.015) 0.200 (0.018) 0.189 (0.021) 0.027 (0.005) -0.115 (0.056) Yes 0.2666 14,232
*** *** *** *** *** *** *** *** *** *** ***
*** *** *** *** *** ***
1.818 (0.394) -0.015 (0.004) 0.331 (0.029) 0.444 (0.044) 0.689 (0.048) -0.073 (0.032) 115.200 (16.300) -0.003 (0.001) -0.120 (0.017) 0.302 (0.029) -0.046 (0.014) 0.012 (0.041) 0.235 (0.022) 0.404 (0.033) 0.449 (0.046) 0.595 (0.049) 0.071 (0.012) -0.731 (0.119) Yes 0.3969 14,232
*** *** *** *** *** ** *** *** *** *** ***
*** *** *** *** *** ***
Table 4: The Effect of Institutional Quality on Financial Access
Explanatory Variable Protection from Expropriation Latitude British Legal Origin Adjusted R-Squared Number of Observations Probability of Having a Bank Relationship [1] [2] [3] [4] 0.027*** 0.022*** (0.005) (0.007) 0.192*** 0.057 (0.070) (0.085) 0.044* 0.017 (0.023) (0.020) 0.2666 14,232 0.2644 14,232 0.2625 14,232 0.2668 14,232 Number of Financial Relationships [5] [6] [7] [8] 0.071*** 0.056*** (0.012) (0.016) 0.507*** 0.156 (0.165) (0.176) 0.137** 0.068 (0.059) (0.049) 0.3969 14,232 0.3932 14,232 0.3906 14,232 0.3976 14,232
Table 5A: The Effect of Institution Quality on the Probability of Having a Bank Relationship, Additional Country Controls
Explanatory Variable Protection from Expropriation Av. Per Capita GDP† English Speaking Bank Freedom Internet Usage Religion Controls Continent Controls Adjusted R-Squared Number of Observations No Yes 0.2697 14,232 Yes No 0.2706 13,250 No No 0.2687 13,336 No No 0.2688 13,336 No No 0.2877 10,799 [2] 0.023 *** (0.007) [3] 0.023 *** (0.005) [4] 0.020 *** (0.008) 2.490 ** (1.180) [5] 0.019 ** (0.008) 2.470 ** (1.150) 0.024 (0.018) [6] 0.029 ** (0.008) -0.301 (1.090) -0.020 (0.035) 0.037 ** (0.015) [7] 0.030 *** (0.008) -3.290 (1.370) -0.078 (0.035) 0.040 (0.014) 0.010 (0.002) No No 0.2887 10,799 ** ** *** ***
Table 5B: The Effect of Institution Quality on Depth of Financial Access, Additional Country Controls
Explanatory Variable Protection from Expropriation Av. Per Capita GDP† English Speaking Bank Freedom Internet Usage Religion Controls Continent Controls Adjusted R-Squared Number of Observations No Yes 0.3997 14,232 Yes No 0.3981 13,250 No No 0.4043 13,336 No No 0.4047 13,336 No No 0.4179 10,799 [2] 0.056 *** (0.015) [3] 0.057 *** (0.013) [4] 0.044 *** (0.016) 8.730 *** (2.770) [5] 0.039 ** (0.017) 8.660 *** (2.760) 0.080 * (0.049) [6] 0.055 *** (0.014) 7.170 *** (2.620) -0.0001 (0.068) 0.056 ** (0.026) [7] 0.056 *** (0.014) 2.30 (3.260) -0.095 (0.091) 0.061 ** (0.026) 0.016 (0.007) ** No No 0.4185 10,799
Dealing with Unobserved Heterogeneity
Bisj or Disj = α + β1Xi + β2ZjxECsj + β3ECsj + δs + δj + εisj, ECsj = # of immigrants from country j living in MSA s total population in MSA s δj – Country of origin fixed effects
Table 6: Does Enforcement Matter? Controlling for Unobserved Heterogeneity, Institution Quality and Ethnic Concentration
Probability of Having a Bank Relationship [1] [2] 1.118*** (0.300) -8.638*** (2.225) No 0.2626 13,867 0.827* (0.487) -5.734* (3.581) Yes 0.2754 13,867 Depth of Financial Access [3] 2.704*** (0.604) -20.989*** (4.516) No 0.3914 13,867 [4] 2.422** (1.075) -18.090** (7.995) Yes 0.4104 13,867
Explanatory Variable Protection from Expropriation * Ethnic Concentration Ethnic Concentration Country Controls Adjusted R-Squared Number of Observations
Table 7: Do Institutions Matter Differently for Different Types of Behavior?
Checking Account [3] 0.024 *** (0.006) 47.0% 0.2386 14,232 Visited a Doctor in Past 12 Months [6] 0.002 (0.003) 79.3% 0.0035 8,705
A. Baseline specification Protection from Expropriation Mean of Dependent Variable Adjusted R-Squared Number of Observations
Stock Ownership [1] 0.016 *** (0.005) 8.6% 0.2315 14,232
Savings Account [2] 0.022 *** (0.006) 40.1% 0.1835 14,232
SelfEmployment [4] 0.007 * (0.004) 8.7% 0.1090 14,231
Drive own Car to Work [5] -0.003 (0.006) 75.1% 0.0573 7,546
B. Controlling for Unobserved Heterogeneity Protection from Expropriation x Ethnic Concentration 0.696 * -0.282 (0.369) (0.598) Ethnic Concentration -5.142 * 2.533 (2.757) (4.417) Country Controls Yes Yes Adjusted R-squared 0.2599 0.1973 Number of Observations 13,867 13,867
1.625 *** (0.496) -12.004 *** (3.665) Yes 0.2492 13,867
0.037 (0.340) -0.400 (2.499) Yes 0.1230 13,866
0.892 (0.658) -6.702 (4.820) Yes 0.0682 7,340
-0.268 (0.492) 1.839 (3.619) Yes 0.0017 8,474
Table 8: Do Institutions Matter Differently for Different Types of Immigrants
Probability of Bank Relationship Protection from Expropriation Adjusted R-Squared Depth of Financial Access Protection from Expropriation Adjusted R-Squared Number of Observations High Educ. Immig. [2] 0.022*** (0.006) 0.1558 0.074*** (0.019) 0.3270 2,842 Low Educ. Immig. [3] 0.044*** (0.010) 0.2453 0.082*** (0.019) 0.2875 5,127 High Skill Workers [4] 0.006 (0.005) 0.1478 0.059*** (0.019) 0.3311 1,984 Low Skill Workers [5] 0.033*** (0.010) 0.2482 0.071*** (0.020) 0.3060 2,408 Citizens [6] 0.021*** (0.006) 0.1819 0.066*** (0.014) 0.3314 5,829 Exclude Mexico [7] 0.031*** (0.005) 0.2281 0.073*** (0.011) 0.3688 10,199
Table 9: The Persistence of Institutions The Effect Institution Quality on the Financial Access by Years of U.S. Experience
Probability of Bank Relationship 1–7 No Age at Arrival Controls Protection from Expropriation Adjusted R-Squared Age at Arrival Controls Protection from Expropriation Adjusted R-Squared Depth of Financial Access No Age at Arrival Controls Protection from Expropriation Adjusted R-Squared Age at Arrival Controls Protection from Expropriation Adjusted R-Squared Number of Observations 0.022** (0.011) 0.3653 0.022** (0.011) 0.3679 1–7 0.052** (0.023) 0.4293 0.052** (0.022) 0.4312 2,619 8 – 12 0.029** (0.011) 0.3168 0.027** (0.011) 0.3191 8 – 12 0.078*** (0.024) 0.4593 0.075*** (0.023) 0.4630 2,192 Years in the U.S. 13 – 17 18 – 27 0.023*** (0.008) 0.3033 0.022*** (0.008) 0.3046 Years in the U.S. 13 – 17 0.081*** (0.019) 0.4369 0.081*** (0.020) 0.4364 2,145 0.027*** (0.009) 0.2653 0.027*** (0.009) 0.2656 18 – 27 0.083*** (0.020) 0.4284 0.083*** (0.020) 0.4283 2,750 28+ 0.012 (0.011) 0.2369 0.010 (0.011) 0.2427 28+ 0.050* (0.029) 0.3967 0.039 (0.028) 0.4012 2,955
Table 10: Learning about Institutions The Effect Institution Quality on the Financial Access by Age at Migration
Probability of Bank Relationship 1 – 15 No Year of Arrival Controls Protection from Expropriation Adjusted R-Squared Year of Arrival Controls Protection from Expropriation Adjusted R-Squared Depth of Financial Access No Year of Arrival Controls Protection from Expropriation Adjusted R-Squared Year of Arrival Controls Protection from Expropriation Adjusted R-Squared Number of Observations 0.032*** (0.011) 0.3322 0.029** (0.013) 0.3350 1 – 15 0.061*** (0.020) 0.4727 0.054** (0.022) 0.4742 1,677 Age at Arrival in U.S. 16 – 20 21+ 0.026** (0.013) 0.3319 0.019*** (0.005) 0.2720
0.025* 0.019*** (0.013) (0.005) 0.3319 0.2769 Age at Arrival in U.S. 16 – 20 21+ 0.061** (0.029) 0.4344 0.062** (0.028) 0.4384 1,639 0.060*** (0.013) 0.4051 0.061*** (0.014) 0.4126 7,963
Table 11: Intergenerational Transmission of Institutional Lessons The Effect of Institution Quality on Financial Access Selected Natives and Immigrants
Probability of Bank Relationship Protection from Expropriation Adjusted R-Squared Depth of Financial Access Protection from Expropriation Adjusted R-Squared Number of Observations Native -0.0001 (0.012) 0.2226 Native 0.039 (0.038) 0.3666 44,181 Immigrant 0.041*** (0.010) 0.2964 Immigrant 0.127*** (0.029) 0.4300 7,040
The native-sample used in these estimates includes U.S. born individuals who identified their ancestral country as: Canada, France, the Netherlands, England, Germany, Hungary, Ireland, Italy, Poland, Russia, Cuba, Mexico, and the Dominican Republic. The immigrant sample includes foreign-born individuals who were born in these same countries.
Findings • Institutions have important indirect effects on the breadth and depth of financial market participation. • Institutional quality appears to be more important than geography and legal origin. • The effect of institutional quality is robust to adding additional country controls, and to dealing with unobserved heterogeneity. • Institutions matter for financial behavior but not for other behavior, they matter more for more “institutionally intense” behavior. • The effect of institutions is very persistent and impacts even those who migrate as young children.
Conclusions • The impact of institutions persists even when individuals voluntarily change countries. • Improving institutions can help to improve financial access. • The impact of reforms aimed at supply-side factors may take a long time to realize because beliefs and preferences must adjust as well • The indirect effect of changes in preferences/beliefs might (eventually) magnify the impact of institutional reforms.