LABOR MARKET COMPETITION AND INDIVIDUAL PREFERENCES OVER IMMIGRATION POLICY 141 The actual coefficient estimates in table 2 identify the TABLE 3.-ESTIMATED EFFECT OF INCREASING SKILL LEVELS ON THE qualitative effect on Immigration Opinion of skills and our PROBABILITY OF SUPPORTING IMMIGRATION RESTRICTIONS other regressors.However,these coefficients do not answer Increase Skill Measure Change in Probability of From Mean to Supporting Immigration our key substantive question of how changes in skill levels Maximum Year Restrictions affect the probability that an individual supports immigra- Occupation Wage 1992 -0.086 tion restrictions.To answer this question,we used the (0.031) estimates of model 1 and 2 to conduct simulations that [-0.138,-0.036] Education Years -0.126 calculate the effect on immigration preferences of changing (0.029) skills,while holding the other variables constant at their [-0.174,-0.0811 sample means. Occupation Wage 1994 -0.337 (0.050) Our simulation procedure works as follows.Recognizing [-0.416,-0.252 that the parameters are estimated with uncertainty,we drew Education Years -0.112 1,000 simulated sets of parameters from their sampling (0.019) [-0.143,-0.081] distribution defined as a multivariate normal distribution Occupation Wage 1996 -0.201 with mean equal to the maximum-likelihood parameter (0.047) [-0.274,-0.120] estimates and variance equal to the variance-covariance Education Years -0.085 matrix of these estimates.For each of the 1,000 simulated (0.017) sets of coefficients,we then calculated two probabilities. [-0.113,-0.057 Setting all variables equal to their sample means,we first Using the estimates from model 1 and 2.we simulated the consequences of changing each skill measure from its mean to its maximum on the probability of supporting immigration restrictions.The calculated the estimated probability of supporting immigra- effect is 0% with the eoof this estimate in parentheses followed by tion restrictions (that is,the probability of supporting a reduction in immigration by either“alot”or“a little").We then calculated the estimated probability of supporting im- porting immigration restrictions.(Table A2 gives simulation migration restrictions when our skills measure is increased results for all variables in model 1 and 2).15 One possible objection to our analysis is the claim that to its sample maximum,while holding fixed all other re- Occupation Wage and Education Years measure labor-mar- gressors at their means.The difference between these two ket skills.For example,Education Years might indicate estimated probabilities is the estimated difference in the greater tolerance or civic awareness.To test this possibility, probability of supporting immigration restrictions between we split our sample between those in the labor force and an individual with average skills and someone with"max- those not in the labor force and then reestimated model 1 imum"skills.We calculated this difference 1,000 times,and and 2 on each subsample.We defined the subset of labor- then-to show the distribution of this difference-we cal-force participants as those individuals reporting that they culated its mean,its standard error,and a 90%-confidence were either employed or unemployed but seeking work.In interval around the mean. every year,the not-in-labor-force subsample was dispropor- Table 3 reports the results of this simulation for our two tionately female:approximately two females for every male, skills regressors.For 1992,increasing Occupation Wage versus a majority of males in the labor-force group.In every from its mean to its maximum($512 per week to $1138 per year,the not-in-labor-force subsample was also much older: week),holding fixed all other regressors at their means, an average age of approximately sixty versus forty for those reduces the probability of supporting immigration restric- in the labor force.It is well known that females and older tions by 0.086 on average.This estimated change has a people have much lower labor-force participation rates than standard error of 0.031 and a 90%-confidence interval of the overall population (-0.138,-0.036).The 1992 results for Education Years are If Occupation Wage and Education Years measure labor- similar:increasing Education Years from its mean to its market skills,then the correlation between these regressors maximum (approximately 12.9 years to 17 years),holding and Immigration Opinion should hold among only labor- fixed all other regressors at their means,reduces the prob- force participants.If these regressors measure non-labor- ability of supporting immigration restrictions by 0.126 on market considerations,then their explanatory power should average.This estimated change has a standard error of 0.029 not vary across the two subsamples.Table 4 reports the and a 90%-confidence interval of (-0.174,-0.081).All results.For the labor force subsample,both Occupation three years give the same result:higher skills are strongly Wage and Education Years are strongly significant,with larger coefficient estimates than the full-sample estimates and significantly correlated with lower probabilities of sup- from table 2.For the not-in-labor-force subsample,the coefficient estimates are much smaller than the full-sample 15 For our simulation procedures,we used the Stata program CLARIFY significance of the effects of several control variables differs across the (Tomz,Wittenberg,&King,1998).These procedures are discussed in two methodologies. King et al.(2000). This content downloaded from 202.120.14.193 on Mon,15 Feb 2016 10:04:26 UTC All use subject to JSTOR Terms and ConditionsLABOR MARKET COMPETITION AND INDIVIDUAL PREFERENCES OVER IMMIGRATION POLICY 141 The actual coefficient estimates in table 2 identify the qualitative effect on Immigration Opinion of skills and our other regressors. However, these coefficients do not answer our key substantive question of how changes in skill levels affect the probability that an individual supports immigra￾tion restrictions. To answer this question, we used the estimates of model 1 and 2 to conduct simulations that calculate the effect on immigration preferences of changing skills, while holding the other variables constant at their sample means. Our simulation procedure works as follows. Recognizing that the parameters are estimated with uncertainty, we drew 1,000 simulated sets of parameters from their sampling distribution defined as a multivariate normal distribution with mean equal to the maximum-likelihood parameter estimates and variance equal to the variance-covariance matrix of these estimates. For each of the 1,000 simulated sets of coefficients, we then calculated two probabilities. Setting all variables equal to their sample means, we first calculated the estimated probability of supporting immigra￾tion restrictions (that is, the probability of supporting a reduction in immigration by either "a lot" or "a little"). We then calculated the estimated probability of supporting im￾migration restrictions when our skills measure is increased to its sample maximum, while holding fixed all other re￾gressors at their means. The difference between these two estimated probabilities is the estimated difference in the probability of supporting immigration restrictions between an individual with average skills and someone with "max￾imum" skills. We calculated this difference 1,000 times, and then-to show the distribution of this difference-we cal￾culated its mean, its standard error, and a 90%-confidence interval around the mean. Table 3 reports the results of this simulation for our two skills regressors. For 1992, increasing Occupation Wage from its mean to its maximum ($512 per week to $1138 per week), holding fixed all other regressors at their means, reduces the probability of supporting immigration restric￾tions by 0.086 on average. This estimated change has a standard error of 0.031 and a 90%-confidence interval of (-0.138, -0.036). The 1992 results for Education Years are similar: increasing Education Years from its mean to its maximum (approximately 12.9 years to 17 years), holding fixed all other regressors at their means, reduces the prob￾ability of supporting immigration restrictions by 0.126 on average. This estimated change has a standard error of 0.029 and a 90%-confidence interval of (-0.174, -0.081). All three years give the same result: higher skills are strongly and significantly correlated with lower probabilities of sup￾porting immigration restrictions. (Table A2 gives simulation results for all variables in model 1 and 2).15 One possible objection to our analysis is the claim that Occupation Wage and Education Years measure labor-mar￾ket skills. For example, Education Years might indicate greater tolerance or civic awareness. To test this possibility, we split our sample between those in the labor force and those not in the labor force and then reestimated model 1 and 2 on each subsample. We defined the subset of labor￾force participants as those individuals reporting that they were either employed or unemployed but seeking work. In every year, the not-in-labor-force subsample was dispropor￾tionately female: approximately two females for every male, versus a majority of males in the labor-force group. In every year, the not-in-labor-force subsample was also much older: an average age of approximately sixty versus forty for those in the labor force. It is well known that females and older people have much lower labor-force participation rates than the overall population. If Occupation Wage and Education Years measure labor￾market skills, then the correlation between these regressors and Immigration Opinion should hold among only labor￾force participants. If these regressors measure non-labor￾market considerations, then their explanatory power should not vary across the two subsamples. Table 4 reports the results. For the labor force subsample, both Occupation Wage and Education Years are strongly significant, with larger coefficient estimates than the full-sample estimates from table 2. For the not-in-labor-force subsample, the coefficient estimates are much smaller than the full-sample TABLE 3.-ESTIMATED EFFECT OF INCREASING SKILL LEVELS ON THE PROBABILITY OF SUPPORTING IMMIGRATION RESTRICTIONS Increase Skill Measure Change in Probability of From Mean to Supporting Immigration Maximum Year Restrictions Occupation Wage 1992 -0.086 (0.031) [-0.138, -0.036] Education Years -0.126 (0.029) [-0.174, -0.081] Occupation Wage 1994 -0.337 (0.050) [-0.416, -0.252] Education Years -0.112 (0.019) [-0.143, -0.0811 Occupation Wage 1996 -0.201 (0.047) [-0.274, -0.120] Education Years -0.085 (0.017) [-0.113, -0.057] Using the estimates from model I and 2, we simulated the consequences of changing each skill measure from its mean to its maximum on the probability of supporting immigration restrictions. The mean effect is reported first, with the standard error of this estimate in parentheses followed by a 90%-confidence interval. significance of the effects of several control variables differs across the two methodologies. 15 For our simulation procedures, we used the Stata program CLARIFY (Tomz, Wittenberg, & Kirng, 1998). These procedures are discussed in King et al. (2000). This content downloaded from 202.120.14.193 on Mon, 15 Feb 2016 10:04:26 UTC All use subject to JSTOR Terms and Conditions