with Fergal Hanks

Models of lumpy capital adjustment are too responsive to interest rates relative to empirical evidence. We argue that allowing for small convex adjustment costs in labour can help these models better match the data. Convex costs cause labour to increase slowly in response to a shock thus smoothing out the impact on the marginal product of capital. Due to both depreciation and uncertainty over future productivity, this delay in the benefits of additional capital can have a large impact on the responsiveness of capital investment.

with George-Marios Angeletos

We study optimal taxation when agents are rationally inattentive. Using a flexible model of inattention, we show that there is no corrective role for taxes when attention costs satisfy an invariance condition. At the same time, some forms of inattention generate inter-dependence across states– tax decisions in one state are no longer independent of those in other states. This coupling arises because an inattentive demand function depends on the price of goods in multiple states of the world. In contrast to the traditional view, inter-dependence may call for lower tax rates. In addition, it breaks Slutsky matrix symmetry and can justify state-dependent taxation. Finally, we clarify how traditional estimates of elasticities and welfare costs of taxation may be biased when agents are inattentive.