The standard new Keynesian model predicts that economies behave differently at the zero lower bound: completely wasteful government spending or forward guidance is very stimulative, and capital destruction or oil supply shocks are expansionary. I provide empirical evidence on this prediction and find it wanting: The Great East Japan Earthquake and oil supply shocks are contractionary at the zero lower bound. Modifications of the model that are consistent with this evidence also overturn other unusual policy predictions, such as large fiscal multipliers. My results suggest that many of the usual rules of economics continue to hold at zero nominal interest rates.

The prevailing neo‐Wicksellian view holds that the central bank's objective is to track the natural rate of interest ( r * ), which itself is largely exogenous to monetary policy. We challenge this view using a fixed‐cost model of durable consumption demand, in which expansionary monetary policy prompts households to accelerate purchases of durable goods. This yields an intertemporal trade‐off in aggregate demand as encouraging households to increase durable holdings today leaves fewer households acquiring durables going forward. Interest rates must be kept low to support demand going forward, so accommodative monetary policy today reduces r * in the future. We show that this mechanism is quantitatively important in explaining the persistently low level of real interest rates and r * after the Great Recession.

We present evidence that the high estimated marginal propensities to consume (MPCs) from the leading household studies result in implausible macroeconomic counterfactuals. Using the 2008 tax rebate as a case study, we calibrate a standard macro model with the estimated micro MPCs to construct counterfactual macroeconomic consumption paths in the absence of a rebate. The counterfactual paths imply that consumption expenditures would have plummeted in spring and summer 2008 and mostly recovered in September 2008. We use narratives and forecasts to argue that these paths are implausible. We show that standard two-way fixed effect estimates of the micro MPCs are upward biased. When we correct for the biases, we estimate smaller micro MPCs using the CEX data than the previous literature. We show that realistic modifications of the model result in general equilibrium forces that dampen rather than amplify micro MPCs. The combination of smaller micro MPCs and dampening general equilibrium forces implies general equilibrium consumption multipliers that are below 0.2.

From March to July 1933, US industrial production rose 57 percent. We show that an important source of recovery was the effect of dollar devaluation on farm prices, incomes, and consumption. Devaluation immediately raised traded crop prices, and auto sales grew more rapidly in states and counties most exposed to these price increases. The response was amplified in counties with more severe farm debt burdens. For plausible assumptions about farmers’ relative MPC, the incidence of higher farm prices, and the aggregate multiplier, this redistribution to farmers accounted for a substantial portion of spring 1933 growth. This farm channel thus provides an example of how the distributional consequences of macroeconomic policies can have large aggregate effects. That recovery in 1933 benefited from redistribution to farmers suggests an important limitation to the use of 1933 as a guide to the effects of monetary regime changes in other circumstances.