Table 19 shows linear IV regressions and tests for endogeneity in the two suspected variables:basic and conspicuous consumption. Table 20 presents the corresponding first-stage regressions. Basic consumption is insignificant. Conspicuous consumption is slightly significant (10 % level) for LS, highly significant for H, and insignificant for D. The size of the coefficients does not significantly differ from the non-instrumented linear estimates in table 21, considering the 95 percent confidence interval. Instrument validity is not rejected, considering the Hansen-J test. Also, the instruments are not rejected as relevant: Shea’s partial R-squared of the first-stage regression is 0.1763 for basic consumption and 0.068 for conspicuous consumption. Under-identification is rejected via the Kleibergen-Paap LM test.