Interest Rate Surprises: A Tale of Two Shocks 
Studies that use high-frequency interest rate changes measured around FOMC announcements to identify the effects of monetary policy shocks may not isolate central bank information shocks from pure monetary policy shocks. Instead, they may capture both surprises due to the central bank deviating from its usual stance and surprises due to the central bank's reactions to its private assessment of the economic outlook.
In the former case, a surprise involving a negative interest rate can be expansionary, whereas in the latter case the negative-interest-rate surprise can be contractionary, because the central bank communicates a negative economic outlook by lowering the interest rate presumably to combat economic weakness.
This paper proposes a new technique to isolate central bank information shocks using information on changes in interest rate expectations around macroeconomic news data releases in addition to changes around FOMC announcements. By isolating the shocks, the authors can (1) obtain pure monetary policy shocks that are clean of central bank information and that can be directly compared with monetary policy shocks in standard models, and (2) analyze the impact of new information about the macroeconomy on the overall economy.
Key Findings
- An interest rate surprise from an FOMC announcement may have both a pure monetary policy shock and an information shock confounded within it.
- The estimated effects of a pure monetary policy shock (measured after being separated from an information shock) are more pronounced than the estimated response to an overall interest rate surprise from an FOMC announcement.
- A properly identified contractionary monetary policy shock leads to lower inflation, reduced economic activity, lower stock prices, and higher bond risk premia.
- An information shock that manifests as a positive interest rate surprise leads to higher prices, higher activity, higher stock prices, and lower bond risk premia.
Implications
This paper introduces a novel method for separating the pure monetary policy shocks and central bank information shocks that jointly enter into interest rate surprises commonly used to identify the effect of monetary policy shocks. The key to this method is the use of high-frequency instruments that enable information shocks to be cleanly identified in isolation, which thereby allows these shocks to be parsed out of high-frequency interest rate surprises.
Unlike previously introduced methods, this paper’s method does not require assumptions about the signs of responses to either policy or information shocks, nor does it require assumptions about a small set of central bank forecasts to be a summary of the bank’s private information.
Results based on this new method show that pure monetary shocks indeed have theory-consistent effects that are quantitatively larger than those estimated using overall interest rate surprises. Information shocks are also found to have theory-consistent effects. Importantly, stock prices are found to increase slightly in response to information that would result in higher interest rates even when the sign of the stock market response is not imposed in the estimation method. This indicates that the discount rate effect is slightly outweighed by the cash flow effect of such information shocks.
Abstract
Interest rate surprises around FOMC announcements reveal both the surprise in the monetary policy stance (the pure policy shock) and interest rate movements driven by exogenous information about the economy from the central bank (the information shock). In order to disentangle the effects of these two shocks, we use interest rate changes on days of macroeconomic data releases. On these release dates, there are no pure policy shocks, which allows us to identify the impact of information shocks and thereby distill pure policy shocks from interest rate surprises around FOMC announcements. Our results show that there is a prominent central bank information component in the widely used high-frequency policy rate surprise measure that needs to be parsed out. When we remove this central bank information component, the estimated effects of monetary policy shocks are more pronounced relative to those estimated using the entire policy rate surprise.