Household behavior below the zero bound

The dramatic drop in interest rates is one of the most salient macroeconomic developments in recent decades. In this project, we study empirically how negative interest rates affect household decisions about consumption and investment.

We use customer data from a major bank in Denmark to measure spending, investment and exposure to negative interest rates for each of its roughly 1 million customers and employ both bunching techniques and panel data techniques for empirical identification.

We explore the behavioral mechanisms by allowing the effect of negative interest rates to vary by preferences (e.g. loss aversion) and cognitive biases (e.g. nominal illusion), which we measure for a subset of bank customers through surveys. The results will provide guidance to today’s monetary policy decisions by producing the first evidence on transmission of negative interest rates to the real economy through household consumption and investment.


One of the most salient macroeconomic developments in recent decades is the dramatic drop in interest rates. While nominal interest rates have historically been well within positive territory and conventional wisdom held that they could not fall below zero – the “zero lower bound” – central banks in Europe and Japan have recently resorted to negative policy rates to stimulate the economy.

In Denmark, where policy rates dropped below zero before anywhere else, banks have chosen to pass on the cost to households in the form of negative interest rates on deposits. In other countries, banks have largely refrained from doing so, but negative deposit rates are bound to become much more common if policy rates remain negative.

In this project, we ask: how do negative interest rates affect household consumption and investment? While the interplay between interest rates and household choices is a classical theme in economics, behavioral theory highlights several reasons why interest rate changes within negative territory (e.g. 0% to -0.5%) may induce very different responses than analogous changes within positive territory (e.g. +0.5% to 0%).

Most importantly, loss aversion suggests a higher sensitivity to negative rates (losses) than to positive rates (gains). This is reinforced by nominal illusion where gains and losses are perceived in nominal rather than real terms. By implication, past evidence on the transmission of monetary policy in positive territory may be less relevant in today’s negative territory.

Our empirical design uses data from a major Danish bank to accurately measure consumption, investment and deposit balances for around 1 million individuals and exploits sharp variation in exposure to negative rates created by the bank’s time-varying interest rate schedules.






Name Title Job responsibilities Image
Asger Lau Andersen Associate Professor Household Economics and Finance; Political Economics; Applied Microeconometrics Billede af Asger Lau Andersen
Niels Johannesen Professor Tax Policy; Tax Havens; Tax Evasion; International Taxation; Financial Crisis Billede af Niels Johannesen

Funded by:

Logo: Independent Research Fund Denmark

Household behavior below the zero bound has received a 3½ year funding from Independent Research Fund Denmark

Period:  July 2023 - December 2026


Principal investigator Niels Johannesen