Article Figures & Data
Figures
- Figure 1
Distribution of Willingness to Take Risks, by Migrant Status
Source: RUMiC-RHS Survey
Notes: The measure (wtRisk) varies between zero (lowest level of willingness to take risk) and ten (highest level of willingness to take risk).
- Figure 2
Individual Willingness to Take Risks and Household Average
Notes: The scatter plot shows residual willingness to take risks for each individual in our estimating sample (vertical axis) versus the average residual willingness to take risks of other members in the household (horizontal axis). Residuals are obtained by regressing individual willingness to take risks on basic demographic controls (gender, age, age-squared, and years of education) and a full set of county of residence dummies. The figure shows the regression fitted line (correlation = 0.59).
- Figure 3
Risk Attitudes and Individual Probability of Migrating, by Level of Willingness to Take Risks
Notes: In Panel A, individuals are defined as migrant if they migrated for work during the year before the survey, and in Panel B, if they ever migrated for work in the past. Individual probabilities of being a migrant are regressed on five dummy variables identifying different levels of willingness to take risks in which the excluded category corresponds to a willingness to take risks equal to zero. The graph plots the estimated coefficients on these dummies together with their 90 percent confidence intervals. Included in the regressions are individual controls (age, age-squared, a dummy for male, years of education, a dummy for married relation with household (HH) head dummies, order of birth, number of siblings, and number of children), household controls (number of family members under 16, in the labor force, and older than 60; per capita house value in logs), and 82 county dummies.
- Figure 4
Individual and Household Migration Decision Models
Notes: These figures are obtained from the simulation described in Section VI.
Tables
Variable Mean SD Min. Max. Obs. Individuals Male 0.50 0.50 0 1 6,332 Age 43.82 10.65 16 60 6,332 Married 0.92 0.27 0 1 6,332 Years of education 7.15 2.83 0 13 6,332 Birth order 2.24 1.33 0 10 6,123 Number of siblings 3.15 1.64 0 11 6,250 Number of children 1.68 0.99 0 7 6,332 Willingness to take risks (wtRisk) 2.57 2.36 0 10 6,332 Migrated last year 0.11 0.31 0 1 6,332 Ever migrated 0.23 0.42 0 1 6,280 Households Household size 4.08 1.32 2 11 2,961 HH members aged <16 0.57 0.73 0 5 2,961 HH members in the work force 2.89 1.09 1 8 2,961 HH members aged >60 0.34 0.61 0 4 2,961 HH head’s education (years) 7.25 2.58 0 12 2,961 Plot size (Mu, 15 Mu = 1 hectare) 4.12 4.08 0 75 2,961 House value per capita (yuan, in logs) 9.16 1.33 1.20 14.04 2,961 HH avg. willingness to take risks 2.46 2.03 0 10 2,961 At least one HH member migrated last year 0.16 0.36 0 1 2,961 Source: 2009 RUMiC-RHS Survey.
Notes: The sample includes all individuals in the labor force (that is, aged 16–60 and not currently in school or disabled) who live in a household (HH) in which more than one member in the labor force has reported risk attitudes.
Migrated Last Year Ever Migrated (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) wtRisk 0.014***
(0.0018)0.005***
(0.0019)0.005***
(0.0019)0.005***
(0.0019)0.030***
(0.0025)0.014***
(0.0027)0.014***
(0.0028)0.014***
(0.0028)wtRisk * male 0.006**
(0.002)0.014***
(0.003)wtRisk * female 0.005**
(0.002)0.013***
(0.003)Basic individual controls X X X X X X X X Additional individual controls X X X X X X Household controls X X X X County fixed effects X X X X X X X X X X Observations 6,332 6,332 6,103 5,992 5,992 6,280 6,280 6,052 5,946 5,946 R-squared 0.187 0.288 0.305 0.310 0.310 0.148 0.273 0.288 0.292 0.292 Notes: The table reports estimates from LPM regressions of a dummy for individual migration status on individual willingness to take risk (wtRisk) and other controls. The migration status dummy equals one if the individual migrated for work in the year before the interview (Columns 1–5) or had ever migrated for work (Columns 6–10). The wtRisk variable measures individual willingness to take risks (decreasing with risk aversion) and has a mean of 2.57 and a standard deviation of 2.36. In Columns 5 and 10, the variable wtRisk is interacted with a male and a female dummy. The basic individual controls are age, age-squared, a dummy for male, and years of education; the additional individual controls are a dummy for married, a dummy for relation to head of household, order of birth, number of siblings, and number of children. Household controls are household size and structure (number of family members under 16, in the labor force, and older than 60) and per capita house value (in logs). All regressions include 82 county fixed effects. The sample includes all individuals in the labor force (that is, aged 16–60 and not currently in school or disabled) who live in households in which more than one member in the labor force has reported risk attitudes. Robust standard errors are clustered at the household level and reported in brackets.
* p < 0.1,
↵** p < 0.05,
↵*** p < 0.01.
Migrated Last Year Ever Migrated (1) (2) (3) (4) (5) (6) wtRisk 0.006***
(0.0019)0.005***
(0.0019)0.006***
(0.0019)0.014***
(0.0028)0.014***
(0.0028)0.014***
(0.0028)BMI 0.004**
(0.0018)0.004**
(0.0018)0.004**
(0.0018)0.001
(0.0023)0.001
(0.0023)0.001
(0.0023)BMI-squared −0.000**
(0.0000)−0.000**
(0.0000)−0.000**
(0.0000)−0.000
(0.0000)−0.000
(0.0000)−0.000
(0.0000)Good health 0.032**
(0.0134)0.024*
(0.0142)0.059***
(0.0219)0.051**
(0.0224)Mental distress (GHQ-12) −0.002**
(0.0009)−0.002*
(0.0012)Individual and HH controls X X X X X X County fixed effects X X X X X X Observations 5,983 5,970 5,970 5,937 5,924 5,924 R-squared 0.311 0.312 0.312 0.292 0.293 0.293 Notes: The table tests the robustness of wtRisk coefficient to the inclusion of further individual health-related characteristics: body mass index (BMI = weight in kilograms over squared height in meters), BMI-squared, a dummy for good health (self-reported health status being average, good or very good, as opposed to poor or very poor), and a measure of mental distress (obtained from a 12-item General Health Questionnaire, GHQ-12, using a Likert scoring method; the variable ranges from 0 (no mental distress) to 36 (highest mental distress). The migration dummy equals one if the individual migrated for work in the year before the interview (Columns 1–3) or had ever migrated for work (Columns 4–6). Individual and household (HH) controls: age, age-squared, a dummy for male, years of education, a dummy for married, relation with HH head dummies, order of birth, number of siblings, and number of children; household size and structure (number of family members under 16, in the labor force, and older than 60); and per capita house value (in logs). All regressions include 82 county fixed effects. The sample includes all individuals in the labor force (that is, aged 16–60 and not currently in school or disabled) who live in households in which more than one member in the labor force has reported risk attitudes. Robust standard errors are clustered at the household level and reported in brackets.
↵* p < 0.1,
↵** p < 0.05,
↵*** p < 0.01.
First Migration in 2009 or Later (After Risk Measurement) Migration Last Year (2008) (Before Risk Measurement) 16–32
(1)16–34
(2)16–36
(3)16–60
(4)16–32
(5)16–34
(6)16–36
(7)16–60
(8)wtRisk 0.014**
(0.007)0.016***
(0.006)0.014***
(0.005)0.003**
(0.002)0.015**
(0.007)0.012**
(0.006)0.010**
(0.005)0.005***
(0.002)Individual & HH controls X X X X X X X X County fixed effects X X X X X X X X Observations 395 509 665 3,979 924 1,117 1,370 5,992 R-squared 0.58 0.50 0.43 0.15 0.48 0.46 0.44 0.31 Notes: This table tests whether self-reported risk attitudes measured at the beginning of year 2009 predict successive first-time migration decisions. In Columns 1–4 the dependent variable is an indicator for the individual having migrated for the first time in year 2009, 2010, or 2011, while in Columns 5–8 the dependent variable is the usual indicator for having migrated last year, that is, in year 2008. Individual controls are age, age-squared, a dummy for male, years of education, a dummy for married, relation with household (HH) head dummies, order of birth, number of siblings, and number of children. The household controls are household size and structure (number of family members under 16, in the labor force, and older than 60) and per capita house value (in logs). All regressions include 82 county fixed effects. The sample includes all individuals in our estimating sample who also reported information about the year of first migration, and whose age is within the indicated range. Robust standard errors are clustered at the household level and reported in brackets.
* p < 0.1,
↵** p < 0.05,
↵*** p < 0.01.
Change in wtRisk 2009–2011 wtRisk 2011 (1) (2) (3) (4) (5) (6) (7) (8) Panel A Migration in 2010 0.037 0.120 0.106 0.108 0.151 −0.059 −0.088 −0.093 (0.232) (0.239) (0.245) (0.244) (0.192) (0.196) (0.201) (0.201) Panel B Migration only in 2010 0.195 0.267 0.242 0.227 0.246 0.045 0.015 0.011 (0.319) (0.325) (0.333) (0.331) (0.257) (0.255) (0.261) (0.261) Migration in 2008 and 2010 −0.121 −0.027 −0.032 −0.013 0.056 −0.164 −0.192 −0.199 (0.303) (0.307) (0.316) (0.315) (0.261) (0.263) (0.270) (0.269) Observations 2,906 2,906 2,813 2,791 2,906 2,906 2,813 2,791 wtRisk 2009 X X X X Basic individual controls X X X X X X Additional individual controls X X X X Household controls X X County fixed effects X X X X X X X X Notes: This table tests the relationship between changes in self-reported risk attitudes between 2009 and 2011 and migration experience in 2010. In Columns 1–4, the dependent variable is the change in self-reported willingness to take risks between the 2009 and the 2011 waves, while in Columns 5–8 the dependent variable is self-reported willingness to take risks in 2011. In Panel A, the main regressor of interest is an indicator for the individual being recorded as migrant in year 2010. In Panel B, the main regressors of interest are an indicator for the individual having migrated only in 2010 and an indicator for having migrated in both 2008 and 2010. In Panel B, willingness to take risks reported in 2009 is always included in the controls. The basic individual controls are age, age-squared, a dummy for male, and years of education. The additional individual controls are a dummy for married, a dummy for relation to head of household, order of birth, number of siblings, and number of children. The household controls are household size and structure (number of family members under 16, in the labor force, and older than 60) and per capita house value (in logs). All regressions include 82 county fixed effects. The sample includes all individuals in our estimating sample who also reported risk attitudes in the 2011 wave. Robust standard errors are clustered at the household level and reported in brackets.
* p < 0.1,
** p < 0.05,
*** p < 0.01.
- Table 6
Within-Household Migration Decision: Relative Measures and Risk Preferences of Other Household Members
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) wtRisk 0.005*** 0.009*** 0.003* 0.005*** 0.009*** 0.003* 0.005*** (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) wtRisk_oth: avg wtRisk of other HH members −0.000 −0.006*** −0.006*** (0.002) (0.002) (0.002) wtRisk_rel: ranking in HH normalized 0.070*** 0.055*** 0.053*** (0.018) (0.020) (0.020) wtRisk_rel: dummy for highest wtRisk in HH 0.016** 0.014* 0.011 (0.008) (0.008) (0.008) Individual and HH controls X X X X X X X X X X HH wtRisk measures Only HH Members in Labor Force All HH Members Observations 5,992 5,992 5,992 5,992 5,992 5,992 5,992 5,992 5,992 5,992 R-squared 0.310 0.310 0.311 0.309 0.310 0.309 0.311 0.311 0.311 0.310 Notes: The table reports the estimates from LPM regressions of a dummy for individual migration status (in the previous year) on different measures of willingness to take risks (at both the individual and household level) and other controls. In Columns 2–3 and 8, we include the average risk preferences of the other household members (wtRisk_oth). In Columns 4–7 and 9–10, we include two alternative measures of the individual’s position in the household ranking of willingness to take risk (yvtRisk_rel) (i) individual ranking in risk attitudes within the household, obtained by ranking household members by their willingness to take risks, assigning a value of one to the most risk-averse person and progressively higher values to the other members, and then normalizing this measure by the number of members reporting risk preferences (Columns 4–5 and 9); (ii) an indicator for the individual having the highest willingness to take risks in the household (Columns 6–7 and 10). The measures of willingness to take risks relative to the household (wtRisk_oth and wtRisk_rel) are computed using only household members in the labor force (that is, aged between 16 and 60 and not currently in school or disabled) in Columns 1–7, and all household members in Columns 8–10. Individual and HH controls: age, age-squared, a dummy for male, years of education, a dummy for married, relation with HH head dummies, order of birth, number of siblings, and number of children; household size and structure (number of family members under 16, in the labor force, and older than 60); and per capita house value (in logs). All specifications include county fixed effects. The sample includes all individuals in the labor force (that is, aged 16–60 and not currently in school or disabled) who live in households in which more than one member in the labor force has reported risk attitudes. Robust standard errors are clustered at the household level and reported in brackets.
↵* p < 0.1,
↵** p < 0.05,
↵*** p < 0.01.
(1) (2) (3) wtRisk 0.040***
(0.005)0.016***
(0.005)0.016***
(0.005)Basic individual controls X X Additional individual controls X Household fixed effects X X X Observations 5,992 5,992 5,992 R-squared 0.602 0.672 0.679 Notes: The table reports the estimates from LPM regressions of a dummy for individual migration status (in the previous year) on willingness to take risks, other controls, and household (HH) fixed effects. Basic individual controls: age, age-squared, a dummy for male, and years of education. Additional individual controls: dummy for married, a dummy for relation to head of household, order of birth, number of siblings, and number of children. The sample includes all individuals in the labor force (that is, aged 16–60 and not currently in school or disabled) who live in households in which more than one member in the labor force has reported risk attitudes. Robust standard errors are clustered at the household level and reported in brackets.
* p < 0.1,
** p < 0.05,
↵*** p < 0.01.
(1) (2) (3) (4) (5) (6) (7) HH_avg_wtRisk 0.010*** 0.004 0.009*** 0.003 0.002 0.003 0.003 (0.003) (0.004) (0.003) (0.004) (0.004) (0.004) (0.004) HH_range_wtRisk 0.017*** 0.016*** 0.017*** 0.015*** 0.014*** (0.004) (0.004) (0.004) (0.004) (0.004) HH controls X X X X X Additional HH wealth controls X Excluding individuals aged above 70 X Excluding individuals aged above 60 X Observations 2,961 2,961 2,961 2,961 2,605 2,908 2,662 R-squared 0.31 0.31 0.31 0.32 0.32 0.32 0.33 Notes: The table reports estimates from LPM regressions of a dummy that equals one if the household (HH) has at least one migrant member in the labor force (that is, aged 16–60 and not currently in school or disabled) on different household-level measures of willingness to take risks and other controls. The variables HH_avg_wtRisk and HH_range_wtRisk measure the average and the range of willingness to take risks in the household, respectively. HH controls: household size and structure (number of family members under 16, in the labor force, and older than 60), per capita house value (in logs), size of the family plot, and the years of education and age of the head of household. Additional HH wealth controls: value of productive assets and household debt (in logs). All specifications include 82 county fixed effects. The sample includes all households in which at least two individuals have reported risk attitudes, and at least one of these is in the labor force (that is, aged 16–60 and not currently in school or disabled). In Columns 6 and 7, we exclude individuals older than 70 and 60 years, respectively, when computing household-level measures of risk preferences. Robust standard errors are reported in brackets.
* p < 0.1,
** p < 0.05,
↵*** p < 0.01.
Labor Force Age Range 16–60 16–50 (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) HH max wtRisk 0.012*** 0.017*** 0.011*** 0.015*** 0.016*** 0.015*** 0.013*** 0.017*** 0.017*** 0.017*** 0.018*** (0.003) (0.004) (0.003) (0.004) (0.004) (0.004) (0.004) (0.004) (0.005) (0.004) (0.004) HH oth wtRisk −0.009* −0.008 −0.009 −0.009* −0.007 −0.012** −0.014** −0.014** −0.015*** (0.005) (0.005) (0.006) (0.005) (0.006) (0.006) (0.006) (0.006) (0.006) HH controls X X X X X X X X X Additional HH wealth controls X X Excluding individuals aged above 70 X X Excluding individuals aged above 60 X X Observations 2,961 2,961 2,961 2,961 2,605 2,908 2,662 2,189 1,904 2,154 2,083 R-squared 0.31 0.31 0.32 0.32 0.32 0.32 0.33 0.36 0.35 0.36 0.36 Notes: The table reports estimates from LPM regressions of a dummy that equals one if the household (HH) has at least one migrant member in the labor force (that is, aged 16–60 and not currently in school or disabled) on the risk preferences of the individual with the highest willingness to take risks in the household among those in the labor force (HH_max_wtRiskhk, the average risk attitudes among all other household members (HH_oth_wtRiskhk, and other controls. HH controls: household size and structure (number of family members under 16, in the labor force, and older than 60), per capita house value (in logs), size of the family plot, and years of education and age of the head of household. Additional HH wealth controls: value of productive assets and household debt (in logs). All specifications include 82 county fixed effects. In Columns 1–7, the age bracket for workers to be considered part of the labor force is 16–60 years; in Columns 8–11 it is 16–50 years. The sample includes all households in which at least two individuals have reported risk attitudes, and at least one of these is in the labor force (that is, within the defined age bracket and not currently in school or disabled). In Columns 6 and 10, we exclude individuals older than 70 years, when computing the risk tolerance of other household members (HH_oth_wtRisk), while in Columns 7 and 11 we exclude those older than 60. Robust standard errors are reported in brackets.
↵* p < 0.1,
↵** p < 0.05,
↵*** p < 0.01.
Free alternate access to The Journal of Human Resources supplementary materials is available at https://uwpress.wisc.edu/journals/journals/jhr-supplementary.html
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.