Editors’ note: This column is based on CEPR Discussion Paper No. 21241 “The Overdelivery Premium: When Monetary Policy Decisions Exceed Market Expectations”. 

Central banks undertake considerable efforts to make monetary policy more predictable. Yet financial markets are still often surprised by policy decisions. 

A large body of research uses such surprises to identify the effects of monetary policy on asset prices or the real economy (Ragusa et al. 2020). It is common practice in that literature to analyse the effect of a surprise with a given sign and magnitude in the same way, regardless of how it arose. But is that warranted?

A recent debate asks whether monetary policy announcements reveal information beyond the policy decision itself. When central banks surprise markets, investors may learn something about policymakers’ assessment of the economy (Nakamura and Steinsson 2018, Jarociński and Karadi 2018) or they may update their perceptions of the central bank reaction function (Bauer et al. 2024).

In new research (Ehrmann and Hubert 2026), we approach this debate from a different angle. We show that two monetary surprises of identical size can have dramatically different effects, depending on whether the central bank exceeds market expectations (what we call ‘over-delivery’) or falls short of them (which we label as ‘under-delivery’).

The distinction turns out to be crucial. Short-term interest rates react up to ten times more strongly when central banks over-deliver than when they under-deliver. Moreover, the evidence suggests that information effects emerge primarily in over-delivery episodes.

Two surprises, one measured shock

Consider two scenarios, both illustrated in Figure 1.

Figure 1 An example of over-delivery versus under-delivery monetary policy surprises

In the first, markets expect a 25-basis point rate increase and the central bank raises rates by 50 basis points. In the second, markets expect a 50-basis point rate cut and the central bank cuts rates by only 25 basis points. In both cases, financial markets experience the same measured surprise: policy is 25 basis points tighter than expected.

Standard measures therefore treat the two events as equivalent.

But they are fundamentally different cases. In the first case, the central bank has exceeded expectations; in the second, the central bank has fallen short of expectations. The distinction matters because monetary surprises affect markets only when investors revise their beliefs. The question is therefore not merely how large the surprise is, but how informative it is.

An over-delivery decision sends a strong and clear signal that is relatively straightforward to interpret. For the central bank to pursue such an aggressive move, two conditions must hold simultaneously. The central bank must feel the need for more aggressive action (which can arise due to more pronounced economic conditions, or because the central bank wants to be more responsive to inflation or output), and policymakers must overcome their well-known preferences for gradualism (Bernanke 2004).

Under-delivery is harder to interpret. It could arise if one of the above two conditions does not hold. A smaller-than-expected policy move can reflect that the central bank is satisfied with a weaker response, or a preference for gradualism. Because investors cannot easily distinguish among these possibilities, under-delivery creates a signal extraction problem that makes its information content more ambiguous and limits the belief revision it triggers.

If this interpretation is correct, over-delivery should trigger stronger belief revisions than under-delivery.

Evidence from 14 advanced economies

To test this idea, we analyse monetary policy decisions in 14 advanced economies between 1997 and 2023, using Bloomberg surveys of policy-rate expectations. These surveys allow us to identify whether a central bank policy decision exceeded or fell short of market expectations. The sample contains 2,381 policy announcements. Around 18% qualify as over-delivery decisions and 20% as under-delivery decisions, with the remainder fully anticipated.

We then examine how financial markets react. The results reveal a striking asymmetry. For short-term interest rates, over-delivery surprises generate responses that are up to ten times larger than under-delivery surprises of comparable size. A 100-basis point over-delivery surprise changes one-month interest rates by roughly 100 basis points. An equivalent under-delivery surprise moves them by less than ten basis points. Figure 2 illustrates the result.

Figure 2 Market responses to over-delivery and under-delivery surprises

Notes: Based on Table 3 in Ehrmann and Hubert (2026). The figure plots regression coefficients with 68% and 90% confidence intervals for one-month, three-month, six-month, and 12-month rates, in response to over-delivery and under-delivery monetary policy surprises.

The effect, which we label the ‘over-delivery premium’, is concentrated at short maturities and gradually disappears further out the yield curve. This suggests that over-delivery primarily changes beliefs about near-term policy and economic conditions.

Importantly, the result is not driven by whether policy is being tightened or loosened. The same pattern appears for both tightening and easing surprises. Nor is it explained by the magnitude of the surprise, by particular countries, crisis episodes, or the way surprises are measured. The asymmetry also appears when we use high-frequency market-based measures for the US or the euro area.

What do investors learn?

This raises the question why markets respond so differently.

One possibility is behavioural. Investors may react more strongly when outcomes exceed expectations than when they fall short, consistent with reference-dependent preferences (Kahneman and Tversky 1979). Another possibility is that markets revise their beliefs about the central bank’s reaction function (Bauer et al. 2024). A third possibility is that over-delivery reveals information about economic conditions (Nakamura and Steinsson 2018).

To distinguish among these explanations, we conduct several tests. First, we examine how professional forecasters revise their expectations for inflation, unemployment, and growth after monetary policy decisions. The results strongly support the information channel. Following an over-delivery tightening, forecasters revise inflation expectations upward and unemployment expectations downward. In other words, they infer that policymakers see stronger inflationary pressures and a stronger economy than previously believed. When central banks under-deliver, the opposite occurs. In response to an under-delivery tightening surprise, inflation forecasts decline and unemployment forecasts increase, particularly at longer horizons, consistent with standard monetary transmission. The differences are substantial. This suggests that markets interpret over-delivery as a signal that economic conditions are more pronounced than previously thought. Figure 3 summarises the contrast.

Figure 3 Forecast revisions after over-delivery and under-delivery tightening surprises

Notes: Based on Table 7 in Ehrmann and Hubert (2026). The figures plot revisions to one-year-ahead inflation (left panel) and unemployment (right panel) forecasts in response to over-delivery and under-delivery monetary policy surprises, along with 68% and 90% confidence intervals.

Second, we estimate perceived central bank reaction functions at the country-month level and examine whether monetary policy surprises lead to revisions in the perceived intercept, smoothing parameter, or responsiveness to inflation or output. We find that over-delivery surprises lead to revisions in the estimated intercept but affect none of the other coefficients. This suggests that over-delivery conveys information about the appropriate level of policy rates, consistent with the central bank’s assessment of economic conditions.

Third, to assess whether the over-delivery premium reflects a general behavioural phenomenon of reference-dependent reactions to news, we examine asset price responses to the surprise component contained in macroeconomic data releases, and test whether over-delivery (i.e. data releases that exceed expectations) also generates stronger market reactions. We find no such effect, meaning that the over-delivery premium is specific to monetary policy.

Reconciling the information-effects debate

These findings help explain why the literature has reached differing conclusions regarding central bank information effects. Pooling over-delivery and under-delivery decisions obscures an important source of heterogeneity. Over-delivery surprises appear to generate information effects. Under-delivery surprises, by contrast, appear to behave more like conventional monetary policy shocks. 

The implication is not that every monetary surprise contains information about the economy. Rather, information effects emerge under specific circumstances — particularly when central banks do more than markets expect.

Implications for monetary policy and research

The findings carry two broader lessons.

For policymakers, they suggest that markets pay attention not only to the size of policy surprises but also to what those surprises reveal about policymakers’ confidence and assessment of economic conditions. Decisions that exceed expectations may communicate much more than the policy move itself.

For researchers, the results imply that monetary policy shocks of identical measured size are not necessarily equivalent. Standard identification approaches typically focus on the magnitude of surprises while abstracting from how those surprises arise. Yet the distinction between over-delivery and under-delivery appears to be economically important.

In an era in which central banks increasingly rely on communication and expectation management, understanding how markets interpret different kinds of surprises may be as important as measuring the surprises themselves.

Authors’ note: The views expressed in this column present the authors’ opinions and do not necessarily reflect the views of the Banque de France or the Eurosystem.

References

Bauer, M, C Pflueger and A Sunderam (2024), “Perceptions about Monetary Policy”, The Quarterly Journal of Economics 139: 2227–2278.

Bernanke, B (2004), “Gradualism”, remarks at an economics luncheon co-sponsored by the Federal Reserve Bank of San Francisco and the University of Washington.

Ehrmann, M and P Hubert (2026), “The Overdelivery Premium: When Monetary Policy Decisions Exceed Market Expectations”, CEPR Discussion Paper No. 21241.

Jarociński, M and P Karadi (2018), “The Transmission of Policy and Economic News in the Announcements of the US Federal Reserve”, VoxEU.org, 3 October.

Kahneman, D and A Tversky (1979), “Prospect Theory: An Analysis of Decision under Risk”, Econometrica 47: 263–91.

Nakamura, E and J Steinsson (2018), “High-Frequency Identification of Monetary Non-Neutrality: The Information Effect”, Quarterly Journal of Economics 133: 1283–1330.

Ragusa, G, R Gürkaynak, R Motto and C Altavilla (2020), “Financial Market Reactions to Monetary Policy Signals”, VoxEU.org, 3 August.