Capital Heterogeneity, Time-To-Build, and Return Predictability

I study how the two major types of business investment, equipment investment and structures investment, are differently linked to stock returns. I empirically show that equipment investment has a significantly stronger predictive power for stock returns than structures investment, both in-sample and out-of-sample, using US aggregate-, US asset-, US industry-, and UK aggregate-level data. To explain this empirical finding, I build a general equilibrium production model in which it takes a shorter time-to-build for equipment investment than for structures investment to transform into productive capital. In the model, equipment investment reacts to productivity shocks in a more timely manner, and thus it reflects more of the information contained in stock prices. In addition, the model provides theoretical support for previous empirical findings of return predictability uncovered from planned investment.