Diversified conglomerates are valued less than matched portfolios of pure-play firms. Recent studies find that this diversification discount results from conglomerates' inefficient allocation of capital expenditures across divisions. Much of this work uses Tobin's q as a proxy for investment opportunities, therefore hypothesizing that q is a good proxy. This paper treats measurement error in q. Using a measurement-error consistent estimator on the sorts regressions in the literature, I find no evidence of inefficient allocation of investment. The results in the literature appear to be artifacts of measurement error and of the correlation between investment opportunities and liquidity.