American Economic Association - Wharton Faculty
American Economic AssociationAre All the Good Men Married? Uncovering the Sources of the Marital Wage PremiumAuthor(s): Kate Antonovics and Robert TownSource: The American Economic Review, Vol. 94, No. 2, Papers and Proceedings of the OneHundred Sixteenth Annual Meeting of the American Economic Association San Diego, CA,January 3-5, 2004 (May, 2004), pp. 317-321Published by: American Economic AssociationStable URL: http://www.jstor.org/stable/3592902Accessed: 20/09/2010 15:46Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available rms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unlessyou have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and youmay use content in the JSTOR archive only for your personal, non-commercial use.Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained herCode aea.Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printedpage of such transmission.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact support@jstor.org.American Economic Association is collaborating with JSTOR to digitize, preserve and extend access to TheAmerican Economic Review.http://www.jstor.org
Are All the Good Men Married?Uncovering the Sources of the MaritalWage PremiumTOWN*ANDROBERTBy KATEANTONOVICSA longstandingand yet unsettled question inlabor economics is: Does marriagecause men'swages to rise? Cross-sectional wage studiesconsistently find that marriedmen earn higherwages thando men who arenot currentlymarried.Even after controllingfor a broad set of covariates, this estimateddifferentialis large, rangingfrom 10 to 50 percent. Among the competingexplanations for the marital wage premium,three receive the most attention.The first is thatmarriagemakes men more productiveby allowing them to specialize in non-householdproduction. The second is that employers discriminatein favor of marriedmen, and the thirdis thattheunobservable characteristics that make menmore productivein the labor marketalso makethem more attractive in the marriage market.The primary difference between the first twoexplanations and the third is that the first twosuggest that the marriagehas a causal effect onmen's wages, while the third implies that theestimatedmaritalwage premiumis the result ofan omitted-variablebias. This paperattemptstoidentify the causal effect of marital status onearnings by using data on monozygotic (MZ)twins to control for unobservedheterogeneity.Data on monozygotic twins have most frequently been used to obtain estimates of thereturnsto schooling (e.g., Orley AshenfelterandAlan Kruger, 1994; Jere Behrmanet al., 1996).These studies control for differences in geneticendowments and family backgroundby examining the relationshipbetween within-twinvariation in schooling and wages. In a similarfashion, we use within-twinvariationin maritalstatus, to examine the effect of marriage onmen's wages.1 We find that, when the data aretreatedas a cross section, the estimated maritalwage premium is 19 percent. When we lookwithin MZ twins, the estimated premium doesnot fall. In fact, the point estimate increases toapproximately26 percent. These results are robust to alternative specifications of the wageequation and various attempts to control formeasurementerror. Thus, the findings indicatethatlittle, if any, of the maritalwage premiumisdue to the selection of more productivemen intomarriage.Previous studies of the marital wage premium, have attemptedto control for unobservable heterogeneity by using panel data todifference out individual-level fixed effects(e.g., Sanders Korenmanand David Neumark,Comwell1991; Eng Seng Loh, 1996; Christopherand Peter Rupert, 1997; Jeffrey Gray, 1997;Leslie Stratton, 2002). Estimates from thesestudies vary considerably.While some authorsreportthatthe maritalwage premiumdisappearsonce individual-level fixed effects have beencontrolled for, others report that the maritalwage premiumremainspositive and significant.There are numerouspotential problems withthese fixed-effects estimates. First, these estimates are likely to be biased if past earningsshocks affect currentmarital status. For example, if men are more likely to get marriedafterreceiving a positive wage shock, then fixedeffects estimatesof the causal effect of marriage1 A common criticism of twin-studies estimates of thereturnsto schooling is that they may exacerbatethe biasescaused by unobservedheterogeneitysince there are likely tobe unobservabledifferences even between identical twins,and it is difficult to imagine what, besides those unobservable differences, would lead twins to choose differentlevelsof education. Our study may be less open to this criticismsince there is arguablya largerrandomcomponent to marital status. See John Bound and Gary Solon (1999) for acomplete discussion of the biases associated with twinbased estimation.* Antonovics:Departmentof Economics, University ofCalifornia-San Diego, 9500 Gilman Drive, La Jolla, CA92103; Town: School of Public Health, Health ServicesResearch and Policy, University of Minnesota, 516 Delaware St., S.E., 15-200 PWB, Minneapolis, MN 55455. Wethank Jere Behrman and Mark Rosenzweig for our data,Andrea Beller, Eli Berman, and Julie Hotchkiss for theirhelpful comments, and Jennifer Poole for her excellentresearchassistance.317
on wages are likely to be biased downwarddueto regression to the mean.2 In addition, fixedeffects estimates will also be biased if unobserved productivity is time-varying. Forexample, fixed-effects estimates of the maritalwage premium will be biased upward if menpostpone marriageuntil increases in their unobserved productivitylead to higher wages.Only one other paper, Harry Krashinsky(2004), uses twin data to study the impact ofmarriage on wages. As in Ashenfelter andKrueger (1994), his data were collected fromthe Twinsburg Twins Festival. Krashinsky'scross-sectional results imply that marriedmaletwins earn 23 percent more than unmarriedtwins. However, the within-twinestimates dropthe returns to marriage to 6 percent, but thestandarderrorsare large (7.7 percent), and thusit is difficult to infer much about the causalrelationshipbetween wages and marriagefromhis study.I. EmpiricalFrameworkWe assume that wij, the logarithmof wagesfor individual i E { 1, 2} from family j is givenby(1)MAY2004AEA PAPERSAND PROCEEDINGS318w,i PiM, yXi, Lij fj uijinterpretedas the causal effect of marriageonwages.For an MZ twin pair, equation (1) can berewrittenas(2)w,j PM, yXlj lj- fj (3)w2j 3M2j yX2j Ulj/2j fj u2j.The principal identifying assumption in ouranalysis is that, for MZ twins, ljj u2j.Thatis, we assume that the genetically determined,individual-specificearningsendowmentis identical across twins. Given this assumption it ispossible to difference equations (2) and (3) sothat(4)wj-w2j 3(M lj- Mv) '/(X, - Xy) (,- uy).Differencingequations(2) and (3) sweeps outindividual-specificand family-specific earningsendowments.As a result, the least-squaresestimate of equation (4) produces an unbiased estimate of 3. If the estimates of 13from equations(1) and (4) are similar, then this suggests thatmarital status is unrelated to unobservedproductivity.where Mj takes on the value of 1 if the man ismarried and 0 otherwise, Xij is a vector ofcontrol variables including age, experience andyears of schooling, /Lijis an individual-specific,genetically determinedearnings endowment,fjis a family-specificearningsendowment,and uiis a mean-zero independently and identicallydistributederrorterm. It is assumed that Aij,f,and u0iare unobservableto the econometrician.The parameterof interest in this study is 3,the marginal impact of marriageon wages. Ifmore-productivemen select into marriage,thenMj will be positively correlatedwith either ijor fj (or both) and the ordinary least-squares(OLS) estimate of 3 will be biased upward. Amajorgoal of this and other studies of the marital wage premiumis to eliminate this selectionbias so that the resulting estimate of 3 can beOur data come from the Socioeconomic Survey of Twins.3 This survey was sent to a subsetof twins from the Minnesota Twins Registry(MTR). The MTR is the largest birth-recordbased twin registry in the United States andcomprises about 80 percent of the approximately 10,400 surviving intact twin pairs bornin Minnesotafrom 1936 through1955. Between1983 and 1990, the MTR staff was able tolocate both membersof about 80 percentof thesurviving pairs and sent them a four-page bibliographic questionnaire (BQ). Then, betweenMay and November of 1994, the Socioeconomic Survey of Twins was sent to the members of the pairs who had filled out the BQ andfor whom the MTR still had a currentaddress.2 See JoshuaAngrist and Krueger (1999) for a fulldiscussion.3 See Behrmanet al. (1996) for furtherdiscussion of thedata.II. Data Description
VOL.94 NO. 2319GENDERIN POLICYAND THE LABORMARKETIn total,dataare availablefrom both membersof487 male twin pairs,of which 280 pairsare MZ.Our analysis focuses solely on these MZ pairs.Our indicator of marital status is currentmarital status.4 It takes on a value of 1 if theindividual is currently married and 0 otherwise. Our measure of schooling is constructedusing the respondents' report of their highestcompleted degree. From these reports we construct four indicator variables for whether anindividual has less than a high-school degree,a high-school degree but no college degree, acollege degree but no postgraduate degree, ora postgraduate degree. The other right-handside variables include tenure at current joband region-of-the-country dummy variables.For the cross-sectional analysis we also include age and age-squared as additional control variables.We restrictour sample in a numberof ways.First, we consider only individualswho work atleast 26 weeks per year and at least 20 hoursperweek. In addition, we drop observations inwhich individuals earn above 60/hour (lessthan 6 percent of the sample) or below 4.25/hour (the Federal minimumwage in 1994). Wealso drop a small number of observations inwhich individuals report working more than100 hours per week. For two individuals whoindicate that they worked more than 52 weeksper year, we code them as having worked 52weeks. Observations with missing data aredropped.We lose 116 twin pairs due to missingvalues and an additional28 twin pairsdue to oursample-selection criteria. Cleaning the dataleaves us with 136 MZ twin pairs. The twins in31 (23 percent) of these pairs differ in theirmaritalstatus.In order to determinewhether our sample isrepresentative of the U.S. population, Table1 compares the means of various demographicand job-tenure variables for the twins in oursample to those of a similarlyselected cohort ofmen in the 1995 March supplementof the Current PopulationSurvey. The CPS sample is similar to our sample of twins with regard toaverage age, weeks worked per year, hours4 We alsoexplored including an indicator for divorcedand widowed, and the results are not qualitativelydifferentfrom those we reporthere.TABLE 1-COMPARISON OF MINNESOTA TWIN SAMPLEAND THE CURRENT POPULATION SURVEYCPSTwins sampleVariableHourly wageAge in yearsWeeks workedper yearHours workedper weekLess than highschoolHigh schoolCollegeMore ied Married Unmarried ed per week, and percentage married. Inaddition, consistent with previous studies, wefind that unmarriedmen earn less, are younger,are less educated, and have lower job tenurethan their marriedcounterparts.III. ResultsThe first column of Table 2 presents thecross-sectional regression results of the logarithm of wages on the marriageindicator andour other explanatoryvariables.The coefficienton marital status is 0.19 (t statistic 1.98).Thus, in the cross section, marriedmen earn a19-percent higher wage than unmarriedmen,controllingfor othercharacteristics.In line withother cross-sectional work on the returns toschooling, the parameterestimates also indicatewages increase with education (e.g., Ashenfelter and Krueger, 1994).The second column of Table 2 reports thewithin-twincoefficient estimatesof the returnto
320AEA PAPERSAND PROCEEDINGSTABLE 2-REGRESSIONOF LOGARITHM OF WAGESON MARITAL STATUSVariableCurrentlymarriedHigh schoolCollegeMore than uthWestCrosssectionWithinMZ 7)1360.10* Statistically significant at the 5-percent level.** Statistically significant at the 1-percentlevel.marriage. The coefficients indicate that menwho are married earn 26 percent more thanunmarried men (t statistic 2.69). Furthermore, under the assumption that within-twindifferences in marital status are exogenous, thenthe 26-percent increase in wages associatedwith marriage has a causal interpretation. Theestimated returns to education are positive butsubstantially smaller than the OLS estimates.Since, these education coefficients are imprecisely estimated, we cannot infer much aboutthe returns to education.A well-known problem with first-differencing equations (2) and (3) is that doing so tendsto exacerbate the biases caused by measurement error, especially if the right-hand-sidevariables are highly correlated within twins (ZviGriliches, 1979). Fortunately, marital status canbe inferred from two separate questions in thesurvey. In only two cases does the respondentgive conflicting answers, and our results do notchange when we drop these individuals fromMAY2004our analysis. In addition, we have estimated 3both in the cross section and within twin pairsusing each twin's report of the other's schooling as an instrument (here education is treatedas a continuous variable) using a strategy suggested by Ashenfelter and Krueger (1994). Theresults are very similar to the non-instrumentalvariables estimates.It is noteworthy that the implied marital wagepremium from the within-twin-pairs regressionis similar in magnitude to the cross-sectionalestimate, suggesting that men are not selectinginto marriage based on unobserved heterogeneity in earnings capacity. Thus, we find no evidence that the observed marital wage premiumarises due to the selection of more productivemen into marriage. In addition, the estimatedcoefficient on marital status remains above 0.21when wage at first full-time job, wife's full-timework experience, or number of children is included in our analysis.IV. ConclusionIn this paper, we examine why married menearn more than men who are not currentlymarried. We use data on monozygotic twins todistinguish between the selection hypothesis(that more productive men are more likely tomarry) and the hypothesis that marriage causesmen's wages to rise. Our results provide littlesupport for the selection hypothesis. Evenwithin MZ twins, the marital wage premiumremains large, and the point estimate is on parwith that from cross-sectional regressions.Thus, the answer to the question posed in thetitle of our paper, appears to be "no." Not all thegood men are married. Rather, our results suggest that marriage causes men's wages to rise.REFERENCESAngrist, Joshua and Krueger, Alan. "EmpiricalStrategies in Labor Economics," in OrleyAshenfelter and David Card, eds., Handbookof Labor Economics. New York: Elsevier,1999, pp. 1277-366.Ashenfelter, Orley and Krueger, Alan. "Estimates of the Economic Return to Schoolingfrom a New Sample of Twins." AmericanEconomic Review, December 1994, 84(5),pp. 1157-73.
VOL.94 NO. 2GENDERIN POLICYAND THE LABORMARKETBehrman,Jere; Rosenzweig,Markand Taubman,Paul. "College Choice and Wages: EstimatesUsing Data on Female Twins." Review ofEconomics and Statistics, November 1996,78(4), pp. 672-85.Bound,John and Solon,Gary. "DoubleTrouble:On the Value of Twins-Based Estimationofthe Returnto Schooling." Economics of Education Review, October 1999, 18(4), pp.169-82.Cornwell,Christopherand Rupert,Peter."Unobservable Individual Effects, Marriage, andthe Earnings of Young Men." Economic Inquiry, April 1997, 21(2), pp. 285-94.Gray, Jeffrey. "The Fall in Men's Return toMarriage:Declining Productivity Effects orChanging Selection?" Journal of HumanResources, Summer 1997, 32(3), pp. 481504.321Griliches,Zvi. "SiblingModels and Data in Economics: Beginnings of a Survey."Journal ofPolitical Economy,October 1979, 87(5), part2, pp. S37-S64.Korenman,Sandersand Neumark,David."DoesMarriage Really Make Men More Productive?" Journal of Human Resources, Spring1991, 26(2), pp. 282-307.Krashinsky, Harry. "Do Marital Status andComputer Usage Really Change the WageStructure?"Journal of Human Resources,2004 (forthcoming).Loh, Eng Seng. "ProductivityDifferences andthe Marriage Wage Premium for WhiteMales." Journal of Human Resources, Summer 1996, 31(3), pp. 568-89.Stratton,Leslie."Examiningthe WageDifferentialfor Marriedand CohabitatingMen."EconomicInquiry,April 2002, 40(2), pp. 199-212.
American Economic Association Are All the Good Men Married? Uncovering the Sources of the Marital Wage Premium Author(s): Kate Antonovics and Robert Town Source: The American Economic Review, Vol. 94, No. 2, Papers and Proceedings of the One Hundred Sixteenth Annual Meeting of the American Economic Association San Diego, CA,
Wharton Chamber of Commerce members. For additional information, contact the Wharton Chamber of Commerce at 979-532-1862. Wharton Chamber of Commerce & Agriculture 225 N. Richmond Road Wharton, TX 77488 Phone: 979-532-1862 Fax: 979-532-0102 Email: admin@whartonchamber.com whartonchamber.com Wharton Journal-Spectator 115 W. Burleson St. Wharton .
Wharton scholarship. Though we were unable to run in-person studies during much of 2021, demand for online studies remained high. In 2021, The Wharton Behavioral Lab supported the research of 53 faculty across 8 Wharton departments. Reflecting this broad engagement, the Wharton Behavioral Lab supported 40 published articles and 112 working .
volunteering with the Penn Rotaract Club At the Wharton Sports Business Initiative where my Wharton GRC friends and I participated case competition held by Deloitte Right: info session by a Wharton Club called Wharton Leadership Ventures. 12 School Hurrah Official Orientation talk held
A Wharton Executive MBA pays off in more ways . than you can imagine, right from the beginning . Chicago, IL . Previously: Portfolio Manager, GE Ventures; San Francisco, CA . University of Pennsylvania. Wharton San Francisco 415.777.1000 . mbaexecwest-admissions@wharton.upenn.edu 2 Harrison St., Sixth Floor
American Economic Association U.S. Economic Growth since 1870: One Big Wave? Author(s): Robert J. Gordon Reviewed work(s): Source: The American Economic Review, Vol. 89, No. 2, Papers and Proceedings of the One Hundred Eleventh Annual Meeting of the American Economic Association (May, 1999), pp. 123-128 Published by: American Economic Association
You can also contact one of the Wharton MBA Consulting Club Officers: President: Neel Bhatia neel.bhatia.wg04@wharton.upenn.edu Executive Vice President: Creighton Schenkel creighton.schenkel.wg04@wharton.upenn.edu Executive Vice President: Ry
10 Wharton Leadership Ventures 40 Wharton clubs 300 Penn clubs Wharton creates an environment where students are not afraid to take risks in the pursuit of their entrepreneurial goals. I have been able to find mentors, conduct research, secure funding for my own business, and work with
Relying on pen and ink, which is the conventional medium used for archaeological illustration (Dillon, B. 1987), the artist was able to use semiotic principles to turn traditional scientific illustrations into allegories of the subjective experience (figure 4). Figure 4 Illustrations have been used to draw the viewers’ attention to the skill of making and the focal points of the fledgling .
