We explore the relation between the index of a core industry and the index of a bundle of industries to detect stock market bubbles by analyzing the movement of regression residuals in these two indexes and examining the causality in the movement of these two indexes for two major stock market bubbles during the last century. Results show that when the correlation between the conditional volatility of a core industry such as the Dow Transports and the conditional volatility of a bundle of industries such as the Dow Jones Industrials is decreasing, a stock market bubble is likely being brewing. Results also imply that while an overheated core industry such as Dow Transports in 1920s could explain (and thus be used to detect) a stock market bubble, this same industry bundle no longer has the same power in 1990s. This is likely due to the fact that, while Transports is still a core industry in 1990s, its relative economic significance has been overtaken by other industries such as Technology industry during this period.