American Political Science Review Vol.96,No.3 TABLE 2.Probit Estimations for Senate Votes on Trade Bills-Industry Group Model Estimation Result Effect of Individual Variables (Dependent Variable =Vote for Protection)a on Probability of Vote for Protection 1824-601875-19131922-371945621970-941824-601875-19131922-371945621970-94 Exporting industries -2.30** -1.09* -3.55*-2.80* -4.79* -0.73 -0.48 -0.50 -0.26 -0.54 (0.31) (0.25) (0.43) (1.06) (1.41) (0.05) (0.07 (0.04) (0.05) (0.10) mport-competing 1.27 1.27* 1.70 1.24 3.45* 0.65 0.52 0.47 0.46 0.73 industries (1.03) (0.56) (1.06) (0.92) (0.79) (0.04) (0.06) (0.09) (0.26) (0.06) N 372 532 367 280 382 log-likelihood -199.80-347.00 -226.24-129.36-229.52 Pseudo-R2 .2249 .0578 .1041.0768.0750 Estimations include constant and dummy variables for individual bills(not shown).Standard errors in parentheses.p<05;*p<.01. PEffects estimated for change in each variable from minimum(0)to maximum(1)values for equations including only that variable and bill dummies using Clarify(King,Tomz,and Wittenberg 2000). TABLE 3.Probit Estimations for House Votes on Trade Bills-Class Model Estimation Result Effect of Individual Variables on (Dependent Variable Vote for Protection)a Probability of Vote for Protection 1824601875-19131922-371945-621970-941824601875-19131922-371945621970-94 Value of farm -1.36* -0.53* -0.032.69* -1.72* -0.40 -0.68 -0.52 0.28 -0.49 production (0.26) (0.11) (0.31) (0.43) (0.43) (0.03) (0.03) (0.04) (0.12) (0.04) Employment in 6.47* 8.46* 15.57* 8.95* 4.04* 0.64 0.73 0.81 0.81 0.69 manufacturing (1.07 (1.17 (2.06) (1.53) (1.73) (0.02) (0.02) (0.02) (0.04 (0.03) Profits in -2.25* -1.32 -2.46* -0.31 0.19 0.45 0.80 0.71 0.68 0.45 manufacturing (0.99) (1.00) (0.72) (0.87) (0.68) (0.10) (0.02) (0.05) (0.12) (0.11) N 1,584 2.656 1.5651.2622.480 log--likelihood-985.28-1,658.12-909.46-754.86-1,605.73 Pseudo-R2 .1001 .0992.1552 .0504 .0638 Estimations include constant and dummy variables for individual bills(not shown).Standard errors in parentheses.*p<.05:"p<.01 bEffects estimated for change in each variable from minimum(0)to maximum(1)values for equations including only that variable and bill dummies using Clarify(King,Tomz,and Wittenberg 2000). exporting industries should be negatively related to Overall,the results of the analysis of the Senate votes for protection,since individuals employed or votes are quite consistent with expectations based upon invested in those industries benefit from trade liber- changes in factor mobility over time.Voting decisions alization,while the importance of import-competing more closely reflect Senator's consideration of the in- industries should be positively related to votes for terests of broad factor classes when levels of mobil- protection ity were higher(in the years between 1875 and 1937) As expected,in each period the estimated coeffi- than when mobility levels were lower (in the periods cients for the exporting industries variable are negative, between 1824 and 1860 and from 1945 to the 1990s). and the coefficients for import-competing industries The pattern works just the other way when we exam- are positive.Again,we must exercise care here in inter- ine the responsiveness of Senate voting to demands preting the size and significance of the separate effects, from free-trade and protectionist industries within each since these two variables appear quite collinear across state. states in early periods.Again(on the right),Isimply cal- Tables 3 and 4 report the results of the analysis of culated the first difference effects on the probability of House votes for each model.These must be treated voting for protection for a change in each variable from with a little more caution since the measures of the its theoretical minimum to its theoretical maximum importance of classes and industries are available only (Oto 1)when excluding the other industry variable from at the state level,rather than the district level,and so we the estimation.Here the pattern in the size of effects are relying on an assumption that the class and industry over time is the reverse of that for the class variables: composition of districts within states are similar. Both industry variables have larger effects on voting in The results are very similar to those obtained from the first and last period and smaller effects on votes in the analysis of Senate votes.The estimated coefficients between. are comparable for each class and industry variable 601American Political Science Review Vol. 96, No. 3 TABLE 2. Probit Estimations for Senate Votes on Trade Bills-Industry Group Model Estimation Result Effect of Individual Variables (Dependent Variable = Vote for Pr~tection)~ on Probability of Vote for Protectionb 1824-60 1875-1913 1922-37 1945-62 1970-94 1824-60 1875-1913 1922-37 1945-62 1970-94 Exporting industries -2.30** -1.09** -3.55** -2.80** -4.79** -0.73 -0.48 -0.50 -0.26 -0.54 (0.31) (0.25) (0.43) (1.06) (1.41) (0.05) (0.07) (0.04) (0.05) (0.10) Import-competing 1.27 1.27 1.70 1.24 3.45* 0.65 0.52 0.47 0.46 0.73 industries (1.03) (0.56) (1.06) (0.92) (0.79) (0.04) (0.06) (0.09) (0.26) (0.06) N 372 532 367 280 382 log-likelihood -1 99.80 -347.00 -226.24 -1 29.36 -229.52 Pseudo- R2 .2249 ,0578 ,1041 .0768 ,0750 aEstimations include constant and dummy variables for individual bills (not shown). Standard errors in parentheses. *p< .05; **p< .01. b~ffectsestimated for change in each variable from minimum (0) to maximum (1) values for equations including only that variable and bill dummies using Clarify (King, Tomz, and Wittenberg 2000). - TABLE 3. Probit Estimations for House Votes on Trade Bills--Class Model Estimation Result Effect of Individual Variables on (Dependent Variable = Vote for Pr~tection)~ Probability of Vote for Protectionb 1824-60 1875-1913 1922-37 1945-62 1970-94 1824-60 1875-1913 1922-37 1945-62 1970-94 Value of farm -1.36** -0.53** -0.03 2.69** -1.72** -0.40 -0.68 -0.52 0.28 -0.49 production (0.26) (0.11) (0.31) (0.43) (0.43) (0.03) (0.03) (0.04) (0.1 2) (0.04) Employment in 6.47** 8.46** 15.57** 8.95** 4.04** 0.64 0.73 0.81 0.81 0.69 manufacturing (1.07) (1.1 7) (2.06) (1.53) (1.73) (0.02) (0.02) (0.02) (0.04) (0.03) Profits in -2.25* -1.32 -2.46" -0.31 0.19 0.45 0.80 0.71 0.68 0.45 manufacturing (0.99) (1.00) (0.72) (0.87) (0.68) (0.1 0) (0.02) (0.05) (0.1 2) (0.11) N 1,584 2,656 1,565 1,262 2,480 log-likelihood -985.28 -1,658.1 2 -909.46 -754.86 -1,605.73 Pseudo-R2 .I001 .0992 .I552 ,0504 ,0638 aEstimations include constant and dummy variables for individual bills (not shown). Standard errors in parentheses. *pi .05; **p< .01. bEffects estimated for change in each variable from minimum (0) to maximum (1) values for equations including only that variable and bill dummies using Clarify (King, Tomz, and Wittenberg 2000). exporting industries should be negatively related to Overall, the results of the analysis of the Senate votes for protection, since individuals employed or votes are quite consistent with expectations based upon invested in those industries benefit from trade liber- changes in factor mobility over time. Voting decisions alization, while the importance of import-competing more closely reflect Senator's consideration of the in￾industries should be positively related to votes for terests of broad factor classes when levels of mobil￾protection. ity were higher (in the years between 1875 and 1937) As expected, in each period the estimated coeffi- than when mobility levels were lower (in the periods cients for the exporting industries variable are negative, between 1824 and 1860 and from 1945 to the 1990s). and the coefficients for import-competing industries The pattern works just the other way when we exam￾are positive. Again, we must exercise care here in inter- ine the responsiveness of Senate voting to demands preting the size and significance of the separate effects, from free-trade and protectionist industries within each since these two variables appear quite collinear across state. states in early periods. Again (on the right), I simply cal- Tables 3 and 4 report the results of the analysis of culated the first difference effects on the probability of House votes for each model. These must be treated voting for protection for a change in each variable from with a little more caution since the measures of the its theoretical minimum to its theoretical maximum importance of classes and industries are available only (0 to 1) when excluding the other industry variable from at the state level, rather than the district level, and so we the estimation. Here the pattern in the size of effects are relying on an assumption that the class and industry over time is the reverse of that for the class variables: composition of districts within states are similar. Both industry variables have larger effects on voting in The results are very similar to those obtained from the first and last period and smaller effects on votes in the analysis of Senate votes. The estimated coefficients between. are comparable for each class and industry variable