SPSS ASSIGNMENT-10 PAGES

The Determinants of Income among Hispanics in the California Central Valley: A Survey Antonio Avalos, Ph.D.* Working Draft as of April 1, 2008 NOT FOR CIRCULATION Abstract This paper examines the determinants of income among Hispanics in the California Central Valley which co mprises Fresno, Kings, Madera and Tulare counties. An augmented standa rd human capital earnings function is applied to examine survey data fo r 768 Hispanics residents of the area.

Findings show a relatively high rate of return to schooling, no gains associated with English proficiency but a positive impact of income for Hispanics born in the United States, and that marriage is associated with higher income levels. In addition, resu lts indicate a significant positive association between Internet use and income, and that the mobility given by driving a car is a sign ificant determinant of income among Hispanics in the Central Valley. JEL Classification : D33, O12, R11, R23 Keywords : Income Determinants, Hispanics, Survey Data California Central Valley ________________________________________________________________________\ * Assistant Professor, Department of Economics, California State University, Fresno, 5245 North Backer Ave. M/S PB20, Fresno CA 93740. Phone (559) 278-8793. Fax (559) 278-7234. [email protected] 1. Introduction In 2006 there were 800,807 individuals of Hispanic origin living in Central California, which comprises Fresno, Kings, Madera and Tulare counties. 1 The rapid growth rate of this demographic segment, which already represents almost 50% of the total population in these four counties, undoubtedly has important implications for the regional economy. Surprisingly, however, lit tle is known about the economic status of this population group in Central California, in particular with rega rd to their income levels and its determinants. The literature investigating income determin ants of immigrants is extensive. The vast majority of studies have approached the issue focusing on human capital investment, arguing that after some period of time, immigr ants in the U.S. increase their education level, accumulate experience, and their gene ral knowledge of the U.S. economy expands.

For example, as Chiswick (1978), Carliner (1980) and Stewart and Hyclack (1982) have shown, these increased abilities and enhanced motivation lead to higher income levels.

More recently, using new data for all groups of immigrants, Espenshade and Fu (1997) and Park (1999) have focused on English- speaking ability as an important factor determining income. More specifically, but still within the human capital theory framework, a growing number of researchers ha ve explored the determinants of earnings among different segments of the immigrant population, particularly Hispanics. Using national level data, Tienda (1983) and Grenie r (1984) for example conclude that English language proficiency and length of U.S. resi dence significantly influence earnings among Hispanics, although these effects depend on nativity and national origin. Similarly, Dávila and Mora (2001) conclude that some Hispanic groups acquire English fluency at a faster pace than others, and that Englis h-skill investments distinctly influence the earnings distributions of different Hispanic groups. Using this research as a point of depart ure, this paper focuses on examining the determinants of income among Hispanics in Central California. The paper aims to contribute to the literature in at least two re spects. First, rather than examining national level survey data, it employs survey data for only four counties in Central California.

Second, it combines the traditional human capita l approach with variables that capture additional potential income-enhancing characteristics.

Overall, results of this investigation re veal substantial variation in income among Hispanics depending on assorted characteristics. More specifically, evidence shows 1) a positive and relatively high rate of return to schooling among Hispanics, 2) a positive correlation between years of la bor experience and income, 3) no gains associated with English proficiency but a positive impact of income for Hispanics born in the United States, 4) a positive association between marriage and income, 5) a strong positive association between Internet use and income and 6) that the mobility attained by driving a car is a contributor to enhanced income.

1 U.S. Census Bureau, Ameri can Community Survey (2006). II. Methodology The tests in this study are implemente d with an augmented standard human capital earnings function, which was orig inally introduced by Becker and Chiswick (1966) and was expanded by Mincer (1974). Fo llowing the literature, the estimating equation in a semi-log form can be written as:

ln Y = 0β + SCHOOL + 1β 2β EXP + 3β EXP 2 + iiX β +ε (1) The dependent variable (ln Y) is the natural logar ithm of annual income.

SCHOOL is the years of sc hooling completed. Coefficient 1β can be interpreted as the rate of return to schooling. EXP is the years of labor experience. This specification assumes that labor market experience begi ns after completing schooling and continues uninterrupted, so EXP is in fact measuring the years of potenti al labor experience. 2 EXP 2 (years of labor experience squared) is inte nded to capture the concavity of the income- experience profile. That is, experience should increase income but in a nonlinear fashion. iX is a vector of demographic and socio economic dummy variables. The first set of variables is included to control for empl oyment characteristics. FTIME takes the value of 1 if working full time and 0 otherwise. PT IME takes the value of 1 if working partial time and 0 otherwise. FARMING takes the valu e of 0 is person is employed in a farming related occupation. Central California is known nationally and inte rnationally for the breadth and productivity of its mostly agri cultural-based economy. Underlying this clear comparative advantage however, it has also been shown that agricultural based employment is associated with low earning levels. Thus, the coefficient for this v\ ariable is expected to be negative. The second set captures additional demographic and human capital variables. MARRIED takes the value of 1 if person is married and 0 otherwise.

This variable controls for possible di fferences in motivation due to family responsibilities, including ch ildren. The assumption is that marriage tends to promote employment stability and to increase the level of investment in human capital. Therefore, it is expected that marriage will be positiv ely associated with income. MALE takes the value of 1 if person is male and 0 otherwise. There is abundant literature that utilizes human capital theory to examine the existence of male-female earnings differences. The factors tested include native ability, formal education, voc ational education, and on-the- job training and experience. Furthermore, so me researchers such as Tellez and Murgia (1990) have argued that Hispanic women face a double form of discrimination because of their gender and their ethnic background. Without trying to disentangle the discrimination debate, this variable is in cluded to test for income differences. BORNUSA takes the value of 1 if person was born in the U.S. and 0 otherwise. This variable is included to assess the importance of place of birth for in come. BILINGUAL takes the value of 1 if person fluently speaks both English and Sp anish and 0 otherwise. This variable is included to capture the effects of English language proficiency on the process of economic assimilation. As McManus, Goul d and Welch (1983) showed, English language proficiency has significant positive effects on earnings for male Hispanics.

2 Years of experience are calculated following the standard formulation proposed by Chiswick (1978):

EXP=AGE-SCHOOL-5. Thus, the coefficient for this variable is expected to be positive. The third set is included to assess the importance of factors related to financial security and wealth, access to information and mobility. OWNHOME takes the value of 1 if person owns a home. As Kain and Quigley (1972) have argued, home ow nership is an important factor that could explain income and economic status achieveme nt since it allows for capital accumulation and enhanced financial security. Furtherm ore, home ownership provides favorable tax breaks and a hedge against inflation. Thus, it is expected that home ownership be positively related to income. Following the wo rk of Raphael and Rice (2002) who find significant earnings differences between those workers that own a car and those that do not, the variable DRIVECAR is included, whic h takes the value of 1 if person drives a car and 0 otherwise. The argument is that dr iving a car not only saves time enhancing productivity, but also lets workers reach more distant places in pursuit of higher sources of income. Thus, the coefficient for this variable is expected to be positive. Finally, the variable INTERNET, which takes the value of 1 if person has access to internet at home and 0 otherwise, is included to assess the imp act of the so called digital divide. Based on the findings by Sangmoon (2003) , who report that not only Internet access itself, but information-seeking activity, produces differenc es in earnings, it is expected that access to digital and information technolog y is positively related to income.

III. The Survey Data The data for this study are from a survey implemented from June through August, 2007 by the Social Research Laboratory at California State University, Fresno. Using Computer Assisted Telephone Interviewi ng (CATI) technology, the questionnaire was designed to elicit information on income , employment, demographic and other socioeconomic characteristics of the Hispan ic population in four counties in Central California: Fresno, Madera, Kings and Tulare. To enhance reliability, the interviewer was fully bilingual in English and Spanish, so th e survey instrument was administered in the language of preference of the respondents who were Hispanics 18 years and older. This explains the high mean age of the sample of 39.2 years old. 50.25% of the interviews were conducted in English and 49.75% in Span ish. The survey resulted in 768 complete interviews, 397 males and 371 females. There we re 189 refusals, for an interview refusal rate of 24.6%. However, the average rate of ‘refuse to respond’ items was only 6.25%.

Respondents were given a $30 gift card for co mpleting the questionnaire. Of the total 768 respondents interviewed, 57.2% (439) lived in Fresno County, 26.3% (202) in Tulare County, 7.3% (57) in Kings County, and 9.1% (70) in Madera County. 3 Descriptive statistics for the sample are presented in Table 1. The sample is broken down by several characteristics shown in the columns. The data reveal at least four salient facts. First, th ere is substantial variation in income among Hispanics. While the mean income for the whole sample is $25,627, it is 25% higher for those Hispanics born in the U.S. Among the factors that possibly explain this difference are better access to education and financial opportunities for U.S. natives. Notice that among those Hispanics born in the U.S., the mean of year s of schooling is higher by more than 2 years.

3 The distribution of the 800,807 Hispanics living in these four counties is: 53.1% in Fresno, 29.2% in Tulare, 8.6% in Kings and 8.9% in Madera. More dramatically, the mean income is 63% higher for those Hispanics with access to the internet. Similarly, although the mean age and mean years of experience are lower than for the whole sample, the mean of years of schooling is higher by more than 3 years.

Thus, data indicates significan t returns to human capital investment. Second, data suggest that English proficiency plays an important role in the process of economic assimilation.

The mean income for Hispanics that are bi lingual is almost 20% higher than the mean income for all Hispanics. In fact, the co rrelation between bilingual ability and farming occupations is negative (-0.24 for the whol e sample) which suggests that English proficiency can propel Hispanics away from agriculture related activities and towards better remunerated jobs. It also worth noticing that the correla tion between age and farming occupations is negative (-0.09 for the whole sample) which indicates that younger Hispanics are less involved with agricu lture related activities. Third, driving a car and owning a home are posi tively correlated not only with a higher mean income, but also with a higher mean age and a with a hi gher percent of people married. Thus, sample data is consistent with the expectation th at older adults have accumulated wealth over their lifespan, and have attained more social stability and financial security. Finally, simple data analysis indicates that forei gn born Hispanics who are not bilingual and that neither drive a car nor own a home, are less like ly to be employed full time. In fact, they are also more likely to be either employed part time or unemployed (n ot shown in table). Table 1: Descriptive Statistics ALL BORNUSA BILINGUAL DRIVECAR INTERNET OWNHOME Mean Mean Mean Mean Mean Mean (Std. Dev) (Std. Dev) (Std. Dev) (Std. Dev) (Std. Dev) (Std. Dev) INCOME $25,627.73 $33,094.56 $30,712.50 $30,134.87 $41,822.03 $33,757.86 ($23,307.22) ($26,808.59) ('$26,394.53) ($25,773.38) ($35,232.22) ($29,248.43) AGE 39.2 38.67 38.79 39.04 33.38 43.21 (14.9) (16.06) (15.53) (14.09) "(11.47) (14.84) SCHOOL 10.41 12.55 11.89 11.23 13.56 11.38 (3.75) (2.71) (3.11) (3.69 (2.65) (3.97) EXP 23.92 21.16 21.92 22.86 14.82 26.91 (15.99) (16.68) (16.23) (15.01) (11.49) (16.05) Percent Percent Percent Percent Percent Percent MARRIED 76.3% 65.7% 73.5% 79.2% 74.4% 79.9% OWNHOME 43.3% 50.3% 51.5% 52.9% 62.4% - INTERNET 16.8% 28.5% 23.0% 20.1% -24.7% DRIVECAR 70.9% 82.2% 82.4% -84.0% 86.8% BILINGUAL 55.3% 74.2% -64.2% 75.2% 65.9% BORNUSA 48.7%-65.3% 56.4% 80.1% 56.6% RETIRED 17.6% 19.4% 17.3% 13.7% 9.6% 18.9% FTIME 43.0% 51.9% 50.6% 48.2% 56.0% 48.5% PTIME 26.8% 23.9% 22.2% 24.1% 28.0% 21.9% FARMING 7.0% 0.8% 2.3% 6.2% 0.8% 2.9% OBS 768 376 427548 125334 Source: 2007 Survey by Social Research Labora tory, California State University, Fresno IV. Results Table 2 presents Ordinary Least Squares (OLS) regression results of estimating equation 1. In all models, heteroskedastic ity-consistent robust standard errors are computed. Significance tests for the county dummies were conducted and were found to be not significant. Because of missing values for several variables the analysis is for 697 observations. Table 2: Determinants of Income (Dependent variable: natural logarithm of income) Variable 1 2 3 4 CONSTANT 8.64 8.52 8.67 8.81 (8.02)*** (8.54)*** (7.29)*** (6.27)*** SCHOOLING 0.089 0.073 0.052 0.045 (13.34)*** (9.46)*** (6.43)*** (4.17)*** EXP 0.018 0.021 0.018 0.016 (4.02)*** (4.85)*** (4.11)*** (3.63)*** EXP2 / 100 -0.023 -0.028 -0.025 -0.022 (-3.28)*** (-4.11)*** (-3.68)*** (-3.19)*** FTIME 0.266 0.201 0.190 0.191 (7.79)*** (5.91)*** (5.79)*** (5.84)*** PTIME -0.210 -0.164 -0.159 -0.157 (-5.92)*** (-4.73)*** (-4.72)*** (-4.75)*** FARMING -0.102 -0.033 -0.063 -0.32 (-1.17) (-0.38) (-0.75) (-0.37) MALE 0.342 0.304 0.361 (6.84)*** (6.23)*** (3.52)*** MARRIED 0.086 0.051 0.045 (2.09)** (1.66) (1.12) BORNUSA 0.181 0.126 0.115 (3.49)*** (2.32)** (2.07)* BILINGUAL 0.128 0.072 0.108 (1.04) (1.54) (1.26) OWNHOME 0.167 0.101 (3.65)*** (2.19)*** INTERNET 0.185 0.172 (3.09)*** (2.86)*** DRIVECAR 0.190 0.186 (3.77)*** (3.67)*** SCHOOLING*BORNUSA 0.031 (2.16)* SCHOOLING*MALE 0.019 (1.51) HOMEOWNER*BORNUSA 0.124 (1.41) Prob > F 0.000 0.000 0.000 0.000 0.364 0.397 0.428 0.438 Observations 697 697 697 697 2 R Notes : * Significant at10%; ** significa nt at 5%; *** significant at 1% The numbers in parenthesis are t-statistics computed with robust standard errors. Consistent with standard human capital th eory, income increases with education and experience. An extra year of schooling raises income by 4.5-8.9%. 4 This estimate is slightly higher than Grenier ( 1984)’s rate of return of 4-5% and Tienda (1983)’s 4-7% for Hispanics. Since both of these studies examine national data, this result implies a slightly higher than average rate of return to educat ion to Hispanics living in Central California.

This result resonates with the findings of Ch iswick (1991) that schooling tends to have a bigger effect on income for immigrants from Latin America. Furthermore, the interaction variable SCHOOLING*BORNUSA in model 4 show s that the rate of return to schooling is 3% higher for those Hispan ics born in the United States. The interaction variable SCHOOLING*MALE suggests that th e rate of return to schooling is the same for males and females. Similarly, an extra year of labor experience raises income by 1.6-2.1%.

Also, as reported in the literature, the negative sign for EXP 2 is evidence that income increases with experience at a decreasi ng rate. With regard to the employment characteristics variables, evidence show s that full time employed Hispanics earn substantially more income than those that are not employed full time (between 16% and 26%) Similarly, retired Hispanics earn substa ntially less income. There is no evidence that income of Hispanics in farming occupa tions is less than those working in other industries. Although agriculture is generally characterized by low-pay occupations, anecdotal evidence suggests that Central Valle y workers in farming-related activities also make extensive use of welfare programs to supplement their incomes. This issue however is beyond the scope of this examination. The second set of variables tests for a dditional demographic and human capital variables. After controlling for schooling and experience, the analysis shows that income of Hispanic males is between 30% and 36% hi gher than their female counterparts, which suggests the presence of discrimination. Th e literature examining gender discrimination among Hispanics, however, is limited. The finding of this paper resonates with the results obtained by Espino and Franz (2002). Using th e Hauser Index of Occupational Prestige they conclude that Hispanic females have, on average, a rank one and a half units below that of Hispanics males. As expected the coefficient for MARRIED is positive and significant in model 2, which suggests that marriage tends to promote employment stability and to increase the level of i nvestment in human capital. However, the coefficient for this variable is not statis tically significant in models 3 and 4, which indicates that this finding is not robust. In contrast to findings by other authors such as Grenier (1984) who argue that the ability to speak English is necessary to acquire information about the labor market and thus to increase the potential of higher income, evidence for the Central Valley indicates that being bilingual is not associated with higher income (the coefficient for BILINGUAL is positive but not statistically significant in the three models). Alternative specificat ions showed similar results. For example, when testing for those that speak Spanish onl y or testing for those that speak English only, although the expected signs are obtained, none of the coefficients are statistically significant. According to Census data, among those Hispanics living in the Central Valley that speak Spanish at home, more than 80% al so speak English. This fact perhaps at least 4 Technically, the rate of return is calculated as . However, following common practice this rate is approximated with )%1(100 − βe %100β . partially explains why being bilingual does not make a signif icant difference in terms of income. A more revealing result however, is the positive coefficient for BORNUSA which suggests that place of birth is an important determinant of income among Hispanics in the Central Valley. Results indica te that those Hispanics born in the United States make on average more than $3,500 a nnually than those Hispanics born outside the country. This finding is similar to Tienda (1983)’s who, using national data for 1976, reports an earnings differential between $1,300 and $1,700 a year for the same groups, particularly for Mexicans and Puerto Ri cans. One possible explanation is that employment opportunities are not the same for both groups. For example, in order to be employed, those Hispanics born outside the Un ited States must demonstrate their legal status in the country which may entail ac quiring social security number, permanent residency or even citizenship. These processe s, which Hispanics born in the country do not have to go through, may also diminish the sets of jobs available to Hispanics born outside the United States or may relegate this group to lower paying activities. The third set tests for variables related to financial security and wealth, access to information and mobility. As expected, home ow nership is associated with higher income levels for Hispanics. Results indicate that income of home owners is between 10% and 16% higher than those that do not own their home. The interaction variable HOMEOWNER*BORNUSA in model 4 shows the effect of home ownership on income is the same for foreign born and United States native Hispanics. Interpretation of this result however must be taken with caution due to the possible reverse causality between home ownership and income. The positive sign of DRIVECAR indicates that mobility is a determinant of income among Hispanics in the Central Valley. Results indicate that incomes of those that drive a car are approxi mately 19% higher than those that do not drive a car. The basic argument is that driv ing a car may positively impact the number of hours worked and that relative to those who must rely on public transportation, workers with cars can be more flexible in searching for better paid jobs. Furthermore, as Raphael and Rice (2002) and Ong (1996) argue, such fl exibility is likely to reduce the duration of unemployment spells and, hen ce to increase average annual income. Finally, similar to the results reported by DiMaggio and Bonikowski (2006) fi ndings indicate significant positive associations between Internet use a nd income, which suggests that some skills and behaviors associated with Internet use were rewarded by the labor market. As shown in Table 1, the annual income of the 125 individuals surveyed is more than $16,000 higher than the whole sample average.

V. Summary and Conclusion This analysis has examined the factors affecting income levels among Hispanics in the California Central Valley. A regression model of income determinants was estimated using survey data to examine the impact of demographic, human capital and other variables related to fina ncial security and wealth. Sign ificant variation in income was found among Hispanics depending of thei r characteristics. Evidence shows that although there are no income gains associated with being bilingual, being a native of the United States is an important f actor that positively affects income. Results also show that overall, married Hispanics receive a higher in come than single people, and that female Hispanics receive a lower income than males. With regard to financial security and wealth, the finding that Internet use increases income suggests that better access to digital and information technology produces a large positive impact. Similarly, evidence indicates that the mobility attained by drivi ng a car is a contributor to enhance income.

Finally, the relatively high rate of return to sc hooling indicates that formal education is an important contributor to incr ease income among Hispanics. References Becker, Gary S. and Barry R. Chiswick, (1966), “Education and the D\ istribution of Earnings”, American Economic Review 56 (1/2): 358-369.

Carliner, Geoffrey, (1980), “Wages, Earnings and Hours of First, Second, and Third Generation American Males”, Economic Inquiry 18: 87-102.

Chiswick, Barry R., (1991), “Speaking, Reading, and Earnings among Low-Skilled Immigrants”, Journal of Labor Economics 9(2): 149-170.

Chiswick, Barry R., (1978), “The Effect of Am ericanization on the Earnings of Foreign Born Men”, Journal of Political Economy 86(5): 897-921.

Espenshade, Thomas J. and Haisan Fu, (1997),” An Analysis of English-Language Proficiency Among U.S. Immigrants”, American Sociological Review 62: 288- 305.

Dávila, Alberto and Marie T. Mora, ( 2001), “Hispanic Ethnicity, English-Skill Investments, and Earnings”, Industrial Relations 40(1): 83-88.

DiMaggio, Paul and Bart B onikowski, (2006), “Make Money Surfing the Web?: The Impact of Internet Use on the Earnings of U.S. Workers”, Princeton University, unpublished manuscript. Espino, Rodolfo and Michael Franz, (2002) , “Latino Phenotypic Discrimination Revisited: The Impact of Skin Color on Occupational Status”, Social Science Quarterly 83(2): 612-623.

Grenier, Gilles, (1984), “The Effects of Language Characteristics on the Wages of Hispanic-American Males”, The Journal of Human Resources 19(1): 35-52.

Kain, John F, and John M. Quigley, (1972), "Housing Market Discrimination, Home Ownership, and Savings Behavior", American Economic Review 62:263-77.

McManus, Walter, William Gould and Finis Welc h, (1983), “Earnings of Hispanic Men: The Role of English Language Proficiency”, Journal of Labor Economics 1(2):

101-130. Mincer, Jacob, (1974), Schooling, Experience and Earnings, New York: National Bureau of Economic Research.

Ong, Paul, (1996), "Work and Automobile Ownership Among Welfare Recipients," Social Work Research 20(4): 255-262.

Park, Jin Heum, (1999), “The Earnings of Immigr ants in the United States: The Effect of English-Speaking Ability”, American Journal of Economics and Sociology 58(1):

43-56.

Raphael, Steven and Lorien Rice, (2002), “C ar Ownership, Employment and Earnings”, Journal of Urban Economics 52(1): 109-130.

Sangmoon, Kim, (2003), “The Impact of Une qual Access to the Internet on Earnings: A Cross-Sectional Analysis”, Perspectives on Global Development and Technology 2(2): 215-236.

Stewart, James B. and Thomas Hyclack, (1982), “An Analysis of the Earnings Profiles of Immigrants”, The Review of Economics and Statistics 66(2): 292-296.

Telles, Edward E., and Edward Murgia, ( 1990), "Phenotypic Discrimination and Income Differences Among Mexican Americans", Social Science Quarterly 72:682-696.

Tienda, Marta, (1983), “Market Characteristics and Hispanic s Earnings: A Comparison of Natives and Immigrants”, Social Problems 31(1): 59-72.