Stavrakeva and Tang showed, using a present-value decomposition of exchange-rate changes disciplined by professional survey forecasts, that macroeconomic announcement surprises explain roughly half of the variation of the currency risk premium and the large majority of quarterly exchange-rate variation, with amplification during U.S. recessions and periods of financial uncertainty, and that exchange rates respond to past news — inconsistent with uncovered interest parity under full-information rational expectations. The paper discusses candidate theory classes but builds and tests none of them. The open problem is to construct and empirically test a microfounded structural mechanism through which macroeconomic announcement surprises drive revisions in expected excess currency returns — the currency risk premium broadly defined — so as to quantitatively generate the observed magnitudes and state dependence of the macro-driven currency-risk-premium revisions.

Background

The exchange-rate disconnect puzzle is the long-standing finding in international macroeconomics that nominal exchange-rate movements appear largely unrelated to macroeconomic fundamentals (output, inflation, money, interest rates) at business-cycle frequencies, making exchange rates look more like financial asset prices than macroeconomic quantities. A benchmark theory here is uncovered interest parity (UIP), the condition that the expected change of an exchange rate equals the interest-rate differential between the two currencies, so that expected returns are equalized; under the additional assumption of full-information rational expectations (FIRE) — all investors know the true data-generating process and share the same information — exchange rates should respond only to genuinely new information. Deviations from UIP define the currency risk premium: the expected excess return from holding a higher-yield foreign currency uncovered, which in equilibrium reflects compensation for risk or frictions rather than expected depreciation alone. A macroeconomic surprise (or news shock) is the difference between a released macroeconomic statistic (e.g., GDP, employment, inflation) and the consensus forecast of professional forecasters collected in surveys such as Consensus Economics and Blue Chip shortly before the release.

Stavrakeva and Tang (working paper 2020; Review of Economics and Statistics 108 (5), 2026) revisit the disconnect with these ingredients. They estimate a survey-augmented vector autoregression (a VAR — a linear system of regressions of each variable on lagged values of all variables — estimated so that the model-implied expectations closely match survey consensus forecasts, which disciplines the unobservable expectations of investors and mitigates small-sample biases), and use it to decompose exchange-rate changes into revisions to expected future fundamentals and revisions to expected future excess currency returns (the risk-premium component). They find macroeconomic surprises explain about half of the variation of the currency risk premium and most quarterly exchange-rate variation — 65% over all periods on average and up to 91% during U.S. recessions in the published version — with the leading driver being the response to past, already-released news, a pattern inconsistent with UIP under FIRE. This is an empirical 'reconnect' of fundamentals to exchange rates, but the mechanism behind it is unexplained: the paper only discusses candidate theory classes — currency risk-premium models, regulatory or institutional frictions, and deviations from FIRE — without constructing or testing a structural model. Related structural work sharpens the tension rather than resolving it: the wedge-accounting survey of Itskhoki and Mukhin (2025) argues that standard macroeconomic shocks are inconsistent with exchange-rate behavior while financial (risk-sharing) shocks fit it, and that news shocks violate the finance-side disconnect; how macroeconomic announcements translate into precisely the observed risk-premium revisions therefore remains an open quantitative gap. The decomposition itself carries the source's own acknowledged caveats — survey consensus as a proxy for the marginal trader's beliefs (with measurement error minimized by the joint estimation objective), zeros inserted for missing surprises in the early sample, and the news-index aggregation as a dimension-reduction choice — but these are limitations of the authors' own estimates, not open tasks posed to successors.

Problem Statement

Explain the Stavrakeva–Tang exchange-rate reconnect structurally. The single scientific objective is to identify the mechanism through which macroeconomic announcement surprises revise currency risk premia: construct a microfounded model — an explicitly specified economic environment with named frictions — in which announcements about macroeconomic indicators (releases relative to survey consensus forecasts) drive revisions in expected excess currency returns (the currency risk premium broadly defined), and in which this transmission generates, quantitatively, the documented facts: (i) macroeconomic surprises account for approximately half of the variation of the currency risk-premium component of the survey-disciplined decomposition, and for the large majority (about 65% on average, up to 91% in U.S. recessions in the published version) of quarterly exchange-rate variation; (ii) the dominant response is to past rather than newly announced news, as the source documents; and (iii) the explanatory shares are state-dependent, amplified during recessions and periods of high financial uncertainty. The mechanism may be drawn from any of the theory classes the source discusses: models featuring currency risk premia (such as time-varying risk-bearing capacity or slow-moving capital), regulatory or institutional frictions, or deviations from full-information rational expectations (such as information or belief frictions that slow the incorporation of news). The target shares are panel-level adjusted R-squared statistics from quarterly regressions of exchange-rate changes and of their decomposition components on the macro news indices constructed in the source, for its panel of nine advanced-economy currencies against the U.S. dollar (the survey-augmented VAR is estimated for 10 advanced economies over quarterly data for 1990–2015, and the news regressions start in 2001:Q4); the headline figures above are those of the published Review of Economics and Statistics version, while the working paper reports the corresponding panel statistics for its own sample (70% of exchange-rate variation, 51% of the risk-premium component, and 84% in U.S. recessions versus 65% in normal times). The answer must be tested empirically against the decomposition, distinguishing revisions in the expected-excess-return (risk-premium) component from revisions in expected paths of relative short-term interest rates and relative inflation (the expected-fundamentals component). Completion is judged by expert review of the mechanism's microfoundations and of the quantitative and empirical agreement with the documented magnitudes and state dependence, not by qualitative consistency alone; quantitative agreement must constitute a genuine test of the mechanism, so the model's frictions and structural parameters must be motivated and disciplined by evidence independent of the targeted shares in (i)–(iii) — for example micro-level evidence, non-targeted macroeconomic moments, or out-of-sample periods — rather than reverse-calibrated to reproduce them.

References

  1. Stavrakeva, Vania, and Jenny Tang. 2026. 'A Fundamental Connection: Exchange Rates and Macroeconomic Expectations.' Review of Economics and Statistics 108 (5): 1332–1348. DOI: 10.1162/rest_a_01520. https://doi.org/10.1162/rest_a_01520
  2. Stavrakeva, Vania, and Jenny Tang. 2020. 'A Fundamental Connection: Exchange Rates and Macroeconomic Expectations.' Federal Reserve Bank of Boston Working Paper No. 20-20, version December 2020. DOI: 10.29412/res.wp.2020.20. https://doi.org/10.29412/res.wp.2020.20
  3. Rey, Hélène, Vania Stavrakeva, and Jenny Tang. 2024. 'Currency Centrality in Equity Markets, Exchange Rates and Global Financial Cycles.' NBER Working Paper 33003. DOI: 10.3386/w33003. https://doi.org/10.3386/w33003
  4. Stavrakeva, Vania, and Jenny Tang. 2025. 'Explaining the Great Moderation Exchange Rate Volatility Puzzle.' IMF Economic Review 73 (1): 196–238. First published online September 18, 2024. DOI: 10.1057/s41308-024-00264-9. https://doi.org/10.1057/s41308-024-00264-9
  5. Itskhoki, Oleg, and Dmitry Mukhin. 2025. 'What Drives the Exchange Rate?' IMF Economic Review 73 (1): 86–117. First published online June 25, 2024. DOI: 10.1057/s41308-024-00251-0. https://doi.org/10.1057/s41308-024-00251-0
  6. Miyamoto, Wataru, Thuy Lan Nguyen, and Hyunseung Oh. 2025. 'In Search of Dominant Drivers of the Real Exchange Rate.' Review of Economics and Statistics 107 (4): 1059–1073. DOI: 10.1162/rest_a_01342. https://doi.org/10.1162/rest_a_01342
  7. Stein, Hillary. 2025. 'Got Milk? The Effect of Export Price Shocks on Exchange Rates.' Journal of International Economics 155, article 104080. DOI: 10.1016/j.jinteco.2025.104080. https://doi.org/10.1016/j.jinteco.2025.104080
  8. Valente, Joao Paulo, Kaushik Vasudevan, and Tianhao Wu. 2021. 'The Role of Beliefs in Asset Prices: Evidence from Exchange Rates.' SSRN working paper 3872077. DOI: 10.2139/ssrn.3872077. https://doi.org/10.2139/ssrn.3872077
  9. Bippus, Balduin, Simon Lloyd, and Daniel Ostry. 2024. 'Granular Banking Flows and Exchange-Rate Dynamics.' SSRN working paper 4578889. DOI: 10.2139/ssrn.4578889. https://doi.org/10.2139/ssrn.4578889
  10. Ostry, Daniel, Simon Lloyd, and Giancarlo Corsetti. 2025. 'Trading Blows: The Exchange-Rate Response to Tariffs and Retaliations.' SSRN working paper 5349043. DOI: 10.2139/ssrn.5349043. https://doi.org/10.2139/ssrn.5349043
  11. Beckmann, Joscha, and Joseph Agyapong. 2024. 'Media Belief in Foreign Exchange Rate Models.' SSRN working paper. DOI: 10.2139/ssrn.5000832. https://doi.org/10.2139/ssrn.5000832
  12. Shugliashvili, Teona, and Erekle Pirveli. 2025. 'Unveiling News Impact on Exchange Rates: A Hybrid Model Using NLP and LDA Techniques.' Applied Economic Analysis 33 (99): 184–205. DOI: 10.1108/aea-04-2024-0183. https://doi.org/10.1108/aea-04-2024-0183
  13. Huang, Naijing, Xiaofen Tan, Jie Xia, and Zixuan Zhao. 2025. 'China's Exchange Rate Forecasting in a Data-Rich Environment: A High-Dimensional Macroeconomic Framework.' Applied Economics Letters, published online December 23, 2025; issue 33 (8): 1224–1232. DOI: 10.1080/13504851.2025.2605188. https://doi.org/10.1080/13504851.2025.2605188
  14. Ma, Yu, Wenxia Zhao, and Zijun Ding. 2025. 'Does Climate Risk Influence Exchange Rates?' Global Finance Journal 68, article 101203. DOI: 10.1016/j.gfj.2025.101203. https://doi.org/10.1016/j.gfj.2025.101203
  15. Modugno, Michele, and Berardino Palazzo. 2025. 'Decoding Equity Market Reactions to Macroeconomic News.' Federal Reserve Board FEDS Working Paper 2025-007. DOI: 10.17016/feds.2025.007. https://doi.org/10.17016/feds.2025.007