Why Expectations Matter: Diagnostic Overreaction and Its Effects on Open-Economy Stability

The paper shows that when people overreact to economic news—a behavioral pattern called Diagnostic Expectations—open economies like Canada experience sharper exchange-rate swings, stronger shock amplification, and worse policy trade-offs than traditional models predict. It finds that this behavioral model fits the data better, forecasts crises more accurately, and fundamentally reshapes how monetary and fiscal shocks transmit through inflation, output, and the exchange rate.

A new working paper from the International Monetary Fund’s Institute of Capacity Development, with academic input from Université Laval, challenges one of macroeconomics’ most enduring assumptions: that people form expectations rationally. In the study, Selim Elekdag, Mananirina Razafitsiory, and Luis-Felipe Zanna argue that a behavioral mechanism known as Diagnostic Expectations (DE), where economic agents systematically overreact to new information, offers a more compelling account of real-world fluctuations, especially in an economy deeply exposed to external shocks like Canada’s. Their study builds the first estimated small open-economy DSGE model featuring DE and finds that it dramatically improves data fit, reshapes shock transmission, and complicates the job of policymakers.

When News Hits, People Overreact, And the Economy Follows

DE stems from the psychological principle of representativeness: when something surprising occurs, say, inflation suddenly rises, people overweight the new information and extrapolate it too strongly. The authors embed this behavioral twist into a standard New Keynesian open-economy model equipped with sticky prices, habit formation, incomplete foreign asset markets, and deviations from the law of one price. DE introduces “surprise terms’’ into core equations like the Euler equation, Phillips curves, and the uncovered interest parity (UIP) condition. These surprise terms cause households and firms to react more aggressively to shifts in expected inflation, interest rates, and exchange rates, altering both the magnitude and timing of economic adjustments.

Canada as a Testing Ground for Behavioral Expectations

Using quarterly Canadian and U.S. data from 1992 to 2023, the authors estimate the model and find striking support for the DE framework. The diagnosticity parameter, the measure of how strongly people overreact, is estimated at 1.96, essentially matching values from earlier closed-economy studies. Bayesian model comparisons show DE overwhelmingly outperforming Rational Expectations (RE), and robustness checks across different priors deliver the same verdict. The estimated structure of the economy also changes: domestic prices become more sensitive to marginal costs, imported goods inflation becomes less responsive, and exchange rate expectations start playing a larger and more volatile role.

A More Turbulent Exchange-Rate Channel

Where DE makes its presence felt most strongly is the exchange-rate channel. Across monetary, fiscal, external, cost-push, and productivity shocks, the DE model produces sharper swings in the real exchange rate. A monetary tightening, for instance, triggers a much stronger currency appreciation under DE than under RE, which further depresses exports and deepens the output contraction. Government spending, normally expansionary, delivers a smaller boost because the appreciation crowds out foreign demand. Risk-premium shocks send the currency into deeper depreciations, boosting exports but increasing volatility.

These amplified currency movements reshape Canada’s macroeconomic story. Under RE, output volatility is driven mainly by demand shocks. Under DE, supply shocks, particularly productivity changes and cost-push disturbances, become the dominant forces. CPI inflation becomes heavily influenced by cost-push shocks, domestic inflation becomes more sensitive to interest-rate policy, and the real exchange rate becomes highly dependent on risk-premium shifts. This reallocation of shock importance echoes Canada’s commodity-linked structure, where global price swings play a defining role.

Sharper Forecasts, but Harder Policy Trade-Offs

The authors also test the model’s forecasting ability. DE improves near-term predictions of output, core inflation, and the real exchange rate, particularly around turning points. In 2016, the DE model accurately captures an output dip that RE misses. During the Global Financial Crisis, it more closely replicates the timing and depth of Canada’s downturn without relying on financial-sector frictions. Because DE reacts strongly to surprising information, it provides earlier warning signals than traditional models.

Yet the policy implications are sobering. Inflation–output volatility frontiers shift outward under DE, meaning policymakers face a worse trade-off: stabilizing inflation requires accepting bigger swings in output, and stabilizing output demands tolerating more inflation volatility. Higher diagnosticity intensifies this constraint, making the economy appear inherently more fragile.

The IMF study argues that diagnostic expectations offer a more realistic and empirically grounded way to model open economies. By embracing the reality that people overreact to news, the DE framework explains the volatility, asymmetry, and exchange-rate sensitivity that define modern macroeconomic environments. The authors encourage future work to extend the approach to emerging markets and to broaden behavioral structures, but their central message is already clear: expectations matter, and they are far less rational than traditional theory assumes.