Working Papers

Technological Progress and Capital Obsolescence: Implications for Productvity Growth (Job Market Paper)

This paper examines how endogenous capital depreciation due to the replacement of obsolete capital-embodied technologies can lead to mismeasurement of productivity growth. In an environment where technology is capital-augmenting, the optimal decision to retire old, less-productive capital increases in the rate of economic growth. Failure to allow de- preciation to vary with the state of the economy leads to mismeasurement of the capital stock and implies that Total Factor Productivity (TFP) growth is a biased metric of true productivity growth. I show that approximately 15% of the productivity slowdown in late-1970s to mid-1990s, is due to a higher bias following a rapid acceleration in the rate of capital-embodied technological innovation. In turn, the true slowdown in productivity since mid-2000s is underestimated by around 17%.

“Estimating the Putty-Clay Investment Model with Likelihood-Based Methods”

My goal is to estimate the growth rate of capital-embedded technological progress within the Putty-Clay model using likelihood-based methods. The heterogeneous firm structure of the putty-clay model does not allow for a recursive formulation of the investment problem, which complicates the solution. For the purpose of estimation, I propose a recursive approximation, which I implement in Dynare, whose parameters incorporate restrictions imposed by the true model.

Capital Search and the Business Cycle

This paper presents a Real Business Cycle model with separate markets for new and used capital, where the reallocation of used capital is subject to search frictions. Households supply new capital to firms and used liquid capital comes from the separation of existing firm-capital relationships. In the general model, used capital separation is an endogenous choice of the firm. We compare this to the case where we fix the separation rate exogenously. Both cases, with and without endogenous separation show significant output amplification as compared to the baseline RBC model. The exogenous separation specification performs well in replicating the procyclical used capital reallocation that we see in the data. Reallocation becomes countercyclical when separation decision is endogenous, with the benefits to reallocation being counterfactually procyclical.