We develop a method to measure ambiguity—uncertainty about the distribution of out-comes—in asset markets, using the volatility of the empirical distribution of unpredictable components in transaction prices. For comparison, we measure risk as the volatility of the unpredictable price component itself, following the conventional practice of using the cross-sectional standard deviation. Applying this framework to 22 million secured lending transactions in the EU, we estimate ambiguity and risk perceived by major money market lenders. Unexpected monetary policy tightening raises both measures. Higher ambiguity reduces repo market liquidity by lowering loan volumes and increasing repo rates, thereby amplifying contractionary effects. Higher risk lowers loan volumes but also repo rates, partly dampening contractionary effects. Our results suggest that ambiguity plays a distinct and quantitatively important role in monetary policy transmission that is overlooked when focusing on risk alone.

This paper investigates the effects of uncertainty on the macro economy by replicating the micro effects of uncertainty on individual subjective beliefs. In our model, the representative household has smooth ambiguity preferences and is uncertain about which scenario the economy will be in the next period: normal growth or recession. We anchor the ratio of expected utilities between the two scenarios through the empirical macroeconomic uncertainty index. The higher the macroeconomic uncertainty rises, the deeper the recession that the household is expecting. Our estimations demonstrate that the smooth ambiguity model with an appropriate level of ambiguity aversion outperforms the benchmark model with no uncertainty in fitting output growth rate, especially during recessions. This holds true even when tested with out-of-sample forecasts. Finally, our analysis reveals that the Global Financial Crisis was associated with an increase in both risk aversion and ambiguity aversion, while the Dot-com Crisis only affected risk aversion.

We study how individual income expectations adjust in times of high macroeconomic uncertainty, using survey data of professional forecasters and households. We find that mean expectations decrease disproportionally with rising macroeconomic uncertainty, as predicted by most economics and finance models. However, the relationship between macroeconomic uncertainty and individual subjective uncertainty is more nuanced. While there is a positive correlation for professional forecasters, this is not the case for households. In fact, for households the correlation is often negative, contrary to what is typically predicted. In addition, subjective uncertainty shows significant persistence, even after controlling for macroeconomic and individual factors. We demonstrate that our empirical results can be reconciled with Max-Min expected utility and smooth ambiguity preferences. This has implications for economics and finance research aiming to study periods of high macroeconomic uncertainty.