For a complete list of publications, please see my Curriculum Vitae.
Publication Indicators
Google Scholar: h=12 (All time)
IDEAS: Top 5% of authors worldwide in one citation category.
Social Media
“Hitting the (Bench)mark when Forecasting the Price of Oil” Ergs and Equilibrium Podcast – Energy, environment, economics, January 26, 2023. Ivey Business School, University of Western Ontario.
Working Papers
Bottom-Up Mixed-Frequency Data Sampling (BUMIDAS) (with Quinlan Lee)
– Presentations: 2025 CEAs, XIX New York Camp Econometrics, 2025 Midwest Econometrics Group (MEG), 2025 Canadian Econometrics Study Group
– NEW! 2025 Presentation on YouTube
– SSRN Working Paper. Revise and resubmit, Journal of Applied Econometrics.
We derive the bottom-up functional form of mixed-frequency data sampling (BUMIDAS) models for temporally aggregated VARMA processes. The BUMIDAS specification depends solely on the structure of the disaggregated data-generating process and is invariant to the degree of time aggregation. Compared to existing unrestricted and restricted MIDAS approaches, this reduces the number of estimated parameters by up to a factor equal to the aggregation frequency, simplifies implementation, and yields more efficient and robust estimation. We show when high-frequency information on both explanatory and response variables is required for efficiency and quantify the value of observing high-frequency data. Simulation results demonstrate that BUMIDAS is the only MIDAS-type method that rivals recursive bottom-up methods and often outperforms them in real-time empirical applications.
Can Futures Prices Predict the Real Price of Primary Commodities? (with Markos Farag and Gregory Upton)
– Presentations: USAEE 2023, Louisiana State University; 2025 NCCC-134
– LCERPA Working Paper 2024-3, Revise and resubmit, Canadian Journal of Economics.
We evaluate futures-based forecasts of period-average commodity prices, tested against the traditional random walk benchmark. Using high-frequency, end-of-month futures data, we examine seventeen commodities across energy, agriculture, biofuels, and base metals. Forecasts significantly outperform the end-of-month no-change at short horizons for several commodities and for most commodities at medium and long horizons. Mean-squared forecast errors are reduced by up to 45 percent relative to forecasts based on monthly averages. Additional gains arise from methods that align contract timing with forecast targets. These results suggest that futures markets contain useful information for predicting average commodity prices in real time across markets.
Don’t Ruin the Surprise: Temporal Aggregation Bias in Structural Innovations
– Presentations: 2023 CEAs, 2023 IAAE, Bank of Mexico, 2023 CESG, 2023 USAEE, University of Saskatchewan, 2025 AEAs, New York Camp Econometrics XIX
– LCERPA Working Paper 2024-7, submitted.
Structural innovations estimated from temporally aggregated data (sums or averages) are shown to be mistimed. Selective sampling rectifies this bias only under specific conditions. We propose a method to test for temporal aggregation bias that reveals more than 70 percent of structural innovations occurred in previous months. Applying this test, we find significant mistiming in shocks to the global crude oil market and to monetary policy in SVARs and DSGE models. Critically, financial markets are shown to have already responded to the predictable component. This predictability and mistiming challenge the reliability of structural economic conclusions drawn from monthly or quarterly data.
Forecasts of Period-Average Exchange Rates: New Insights from Real-Time Daily Data (with Martin McCarthy)
– Presentations: 2023 ISF, 2024 CEAs, 2024 SEM; 2025 RTE Data Conference; 3rd Vienna Workshop on Economic Forecasting
– NEW! LCERPA Working Paper 2024-6, submitted.
– NEW! Real-time daily and monthly effective exchange rate data
Forecasting period-average exchange rates requires using high-frequency data to efficiently construct forecasts and to test the accuracy of these forecasts against the traditional random walk hypothesis. To achieve this, we construct the first real-time dataset of daily effective exchange rates for all available countries, both nominal and real. Our findings indicate that forecasts constructed with daily data can significantly improve accuracy, up to 40 percent compared to using monthly averages. We also find that unlike bilateral exchange rates, daily effective exchange rates exhibit properties distinct from random walk processes. When applying efficient estimation and testing methods made possible for the first time by the daily data, we find new evidence of real-time predictability for effective exchange rates in up to fifty percent of countries.
Seize the Last Day: Period-End-Point Sampling for Forecasts of Temporally Aggregated Data* (with L. Arango-Castillo and R. Ellwanger)
– Presentations: 2023 ISF, 2023 CEAs, 2025 James MacKinnon Conference
– LCERPA Working Paper 2023-6 (Updated Dec 2024), submitted.
– NEW! 2025 MacKinnon Conference Presentation on YouTube
Forecasts of temporal aggregates, such as monthly or quarterly averages of daily data, are often constructed using the aggregated data due to difficulties in disaggregating estimation. However, we show that when the daily data is persistent, forecasts constructed with aggregated data are inefficient and can nearly double forecast error. We propose a new forecasting technique, Period-End-Point Sampling (PEPS), that corrects for the information loss and allows models to maintain the lower frequency of the forecast target. Predicated on the intermediate value theorem, PEPS uses end-of-period data to construct point-in-time forecasts and ex-post modifies the end-of-period forecasts to be equal to the period average forecast. Applications to real-time forecasts of 10-year bond yields and the price of copper show that PEPS rivals the accuracy of bottom-up approaches.
Leverage and Rate of Return Heterogeneity among U.S. Households
– NEW! LCERPA Working paper 2024-8
– Presentations: 2023 AEA, CEA, EEA, and IPDC conferences.
– 2023 AEA presentation on YouTube
This paper proposes measures of panel-data for returns to U.S. households’ wealth and documents new facts on the heterogeneity in returns to wealth between households and over the wealth distribution. First, leverage exhibits permanent heterogeneity and explains most of the permanent heterogeneity in the returns to wealth. As such, the permanent heterogeneity in returns to assets understates the permanent heterogeneity in returns to wealth, with a standard deviation of 3.8 and 9.2 percentage points, respectively. Second, returns to wealth decline as households become wealthier and exhibit declining returns to scale and specialization. Third, household-specific returns to wealth and assets are correlated with the persistent component of labor earnings. Fourth, housing and the regressivity of after-tax mortgage rates are critical to explaining the permanent heterogeneity in returns. These findings inform the direction of scale dependence and degree of type dependence in returns heterogeneity for the study of portfolio allocation, wealth inequality, social mobility, and corresponding policies.
Doctoral dissertation
Snudden, S. (2019). Household Return Heterogeneity in the United States.
– Recipient of the 2019-2020 C.A. Curtis Prize in Economics for best PhD thesis.
This thesis provides empirical evidence on heterogeneity in returns on wealth among U.S. households.
Refereed Publications
Benmoussa, A., R. Ellwanger, and S. Snudden. 2026. Carpe Diem: Can daily oil prices improve model-based forecasts of the real price of crude oil? International Journal of Forecasting, vol. 42(1), 281–295
– Real-time data, updated monthly
– Complete replication package.
We propose high-frequency techniques that use the underlying nominal daily data in model-based forecasts of average real series. The high-frequency forecasts reduce errors by nearly half compared to current practices. Contrary to models estimated with monthly average data, we demonstrate that model-based real-time forecasts of the real price of crude oil can outperform the end-of-month no-change forecast, the traditional random walk forecast.
Arain, R., and S. Snudden. 2025. When Are Statistical Forecast Gains Economically Relevant? Evidence from Bitcoin Returns, Journal of Forecasting, 1–16.
Puzzle: 50+ studies claim Bitcoin is forecastable, yet almost always they’d lose you money in practice. We show that standard metrics (MSFE, hit rate) collapse during extreme moves, propose Extreme Directional Accuracy as the profit-critical measure, and find that only the USD Index and Shanghai stocks generate robust excess returns. We find significant evidence of excess profitability for our proposed threshold-based trading strategies, questioning the Efficient Market Hypothesis for Bitcoin.
McCarthy, M., and S. Snudden. 2025. Predictable by Construction: Assessing Forecast Directional Accuracy of Temporal Aggregates, Applied Economics, 1–16.
In macroeconomics, forecasted data are often constructed as an aggregate over some time interval, such as monthly average prices or quarterly total output. We demonstrate that assessing the null of no directional accuracy requires computing the sign of change relative to the latest disaggregated observation rather than the latest aggregated observation. Changes are predictable by construction when assessed relative to the latest aggregated observation, resulting in inflated type I error and loss of power under the alternative. In contrast, assessing the change in the temporal aggregate relative to the latest disaggregated observation results in expected success ratios of 0.5, hypothesis tests with the intended type I error rate, and high power under the alternative.
Benyo, E., R. Ellwanger, and S. Snudden. 2025. A Reappraisal of Real-Time Forecasts of the Real Price of Oil. Economic Inquiry, 1–10.
– Complete replication package
We replicate Baumeister and Kilian (2012) to reappraise real-time forecasts of the real price of crude oil against the end-of-month no-change forecast, the equivalent naive benchmark used for asset prices. We find no consistently significant improvements in the predictive accuracy of model-based forecasts over this naive benchmark at short horizons. Only futures-based forecasts consistently outperform the end-of-month no-change forecast, and only at longer horizons. These results challenge the consensus on the predictability of the real price of crude oil and the merits of alternative forecast approaches. Our findings motivate broader reassessment and replication of forecasting models of temporally aggregated series.
Ellwanger, R., and S. Snudden. 2025. Putting VAR forecasts of the real price of crude oil to the test, Finance Research Letters, vol. 77(106940): 1–6.
This study reassesses crude oil price forecasts from state-of-the-art SV-BVAR models using the traditional end-of-month no-change forecast, consistent with the random walk hypothesis. The forecasts do not significantly outperform the random walk at horizons under one year, and inference changes for some models up to 18 months. The differences stem from a systematic bias in test statistics induced by benchmark choice, with implications for evaluating long-horizon forecasts of averaged series.
Snudden, S. 2025. Idiosyncratic Asset Return and Wage Risk of US Households, Economic Inquiry, vol. 63(2): 636–657.
-10-min. 2022 ASSA presentation on YouTube
–NEW! Replication code and data
This paper documents the degree of idiosyncratic asset return risk and its correlation with wage risk for US households. Novel panel-data measures for returns on household assets are proposed. Sizable idiosyncratic return risk is documented to exist concurrently with permanent heterogeneity in household-specific returns, and exhibits negative serial correlation. On average, idiosyncratic permanent risk to wages and transitory risk to total asset returns are correlated. This arises primarily from correlated wage and return risk to primary housing assets, and is age-dependent. The estimates inform the covariance structure of idiosyncratic asset returns and wage risk.
Ellwanger, R. and S. Snudden, 2023. Forecasts of the Real Price of Oil Revisited: Do they Beat the Random Walk? Journal of Banking & Finance, vol. 154(106962), 1–8.
– Short presentation on YouTube
– Complete replication package and real-time data, updated monthly
We demonstrate that only the end-of-period no-change forecast reflects the traditional random-walk forecast for forecasts of period averages. This end-of-period no-change is significantly more accurate and can halve mean-squared forecast error at the one-period ahead. When applied to the real price of crude oil, the end-of-period no-change outperforms the period average no-change up to one-year ahead. All existing one-month ahead forecasts in the real price of crude oil literature are shown to perform worse than the end-of-month no-change forecast. The result calls into question the usefulness of existing forecasting approaches relative to naive forecasts in literatures examining forecasts of temporally aggregated series.
Ellwanger, R. and S. Snudden, 2023. “Futures Prices are Useful Predictors of the Spot Price of Crude Oil” The Energy Journal, vol. 44(4): 65–82.
– Short presentation on Youtube
– Complete replication package and real-time data, updated monthly
Contrary to the established view, crude oil futures prices significantly improve upon the accuracy of monthly average no-change forecasts. This results from two innovations. First, we document that independent of the construction of futures-based forecasts, longer-horizon futures prices have become better predictors of crude oil spot prices since the mid-2000s. Second, we show that futures curves constructed using end-of-month prices instead of average prices have consistently been able to generate large accuracy-improvements for short-horizon forecasts of average prices. These findings are remarkably robust and apply to all major crude oil benchmarks.
Coletti, D., R. Lalonde, P. Masson, D. Muir and S. Snudden, 2021. “Commodities and Monetary Policy: Implications for Inflation and Price Level Targeting” Journal of Policy Modelling, vol. 43(5): 982–999.
This paper examines the relative merits of price level versus inflation targeting in response to international shocks to primary commodity markets. Persistent crude oil price movements result in significant deterioration of the inflation-output gap trade-off available to central banks. When such terms-of-trade shocks are prevalent, price level targeting is inferior to inflation targeting.
Snudden S., 2019. “Labor and Behavior Determinants of Remittances in Saudi Arabia.” Economic Notes, Special Issue, vol. 48(3): 1–16. (working paper)
This is the first study to structurally deconstruct remittance dynamics into the behavioral and labor market outcomes of migrants. For Saudi Arabia, the estimates suggest that migrant labor supply is highly elastic. The important determinates of remittance dynamics are the marginal propensity to remit, migrant wages, and the extensive margin of migrant labor supply. The marginal propensity to remit is found to respond countercyclical to foreign gross domestic product.
Snudden S., 2018. “International Remittances, Migration, and Primary Commodities.” The World Economy, vol. 41(11): 2934–2953. (working paper, appendix)
This paper documents the global crude oil market as a major driver of international migration and remittances. Large oil exporters who are labor importers transmit international financial and labor spillovers to oil-importing labor exporters. Despite large remittance flows induced by global crude oil shocks, the remittees’ economic conditions are dominated by primary commodity terms of trade channels.
Snudden S., 2018. “Targeted Growth Rates for Long-Horizon Crude Oil Price Forecasts.” International Journal of Forecasting, vol. 34: 1–16.
Ever wonder why forecasters take the H/2 growth rate of the data when targeting forecast horizon H? That comes from this paper! That relationship is derived from spectral analysis and is proposed as a technique to improve medium- to long-horizon forecast accuracy. When applied to crude oil forecasts, the method can generate similar precision at 1-5 year horizons that had previously only been found at horizons of less than one year.
Snudden S., 2016. “Cyclical Fiscal Rules for Oil-Exporting Countries.” Economic Modelling, vol. 59: 473–483.
This paper examines optimal fiscal and monetary policy responses to temporary shocks to the international market for crude oil. Budget-balance tax-gap rules and inflation targeting are the preferred regimes to stabilize macroeconomic volatility and welfare in oil-exporting countries. The output-inflation trade-off is of particular concern for oil exporters due to the pass-through of oil prices into headline inflation.
Andersen, D., B. Hunt, and S. Snudden. 2014. “Fiscal Consolidation in the Euro Area: How Much Pain can Structural Reforms Ease?” Journal of Policy Modeling, vol. 36(5): 785–799.
This paper examines the scope for structural reforms in the euro area to offset the negative effects of fiscal consolidation. The results suggest that structural reforms in core countries could be expected to offset the near-term negative impact on activity arising from the required fiscal consolidation. However, for the periphery, the results suggest that it would take several years before structural reforms could return the level of output back to its pre-consolidation path.
Beaton, K., R. Lalonde, and S. Snudden. 2014. “The Propagation of U.S. Shocks to Canada: Understanding the Role of Real-Financial Linkages.” Canadian Journal of Economics, vol. 47(2): 466-497.
This paper introduces a financial accelerator, inter-bank lending markets, and international bank lending into an international DSGE model. We find that the U.S. banking and interbank markets are an important source of variability. The presence of both the demand and the real supply sides of credit in the model capture the positive comovement of consumption and investment in both domestic and international business cycles.
Coenen, G., C. de Resende, C. Erceg, C. Freedman, D. Furceri, J. in’t Veld, M. Kumhof, R. Lalonde, D. Laxton, J. Linde, A. Mourougane, D. Muir, S. Mursula, J. Roberts, W. Roeger, S. Snudden, and M. Trabant, 2012. “Effects of Fiscal Stimulus in Structural Models.” American Economic Journal: Macroeconomics, vol. 4(1): 22–68.
This paper compares discretionary fiscal stimulus using seven policy and two prominent academic DSGE models. Considerable agreement across models is found on both the absolute and relative sizes of different types of fiscal multipliers. Fiscal policy is most effective if it has moderate persistence and if monetary policy is accommodative. Permanently higher deficits imply significantly lower initial multipliers.
Klyuev, V. and S. Snudden, 2011. “Effects of Fiscal Consolidation in the Czech Republic.” The Czech Journal of Economics and Finance, vol. 61(4): 306-326.
This paper assesses dynamic fiscal multipliers for a variety of fiscal instruments, consolidation durations, assumptions about credibility, and monetary policy responses in the Czech Republic. The article evaluates proposed and alternative “growth-friendly” fiscal consolidations to achieve the government’s balanced budget target.
Chapters
Duttagupta, R., J, Bluedorn, A. Pescatori, and S. Snudden . “Commodity Price Cycles and Commodity Exporters,” Chapter 4 in the World Economic Outlook, April 2012. International Monetary Fund: Washington D.C.
Anderson, D., M. Badia, E. Ruiz, S. Snudden and F. Vitek, 2015. “Fiscal Consolidation under the SGP: Some Illustrative Simulations,” Chapter 7 in Mechanics of a Strong Euro Area, by M. Pradhan, and P. K. Brooks (eds), International Monetary Fund: Washington D.C.
Technical Reports and Policy Papers
de Resende, C., R. Lalonde, and S. Snudden, “The Power of Many: Assessing the Economic Impact of the Global Fiscal Stimulus,” Bank of Canada Discussion Paper 2010-1.
de Resende, C., K. Beaton, R. Lalonde, and S. Snudden, 2010. “Prospects for Global Current Account Rebalancing,” Bank of Canada Discussion Paper 2010-4.
Andersen, D., B. Hunt, M. Kortelainen, M. Kumhof, D. Laxton, D. Muir, S. Mursula, and S. Snudden, 2013. “Getting to Know GIMF: The Simulation Properties of the Global Integrated Monetary and Fiscal Model,” IMF Working Paper No. 2013-55.
Andrle, M., P. Blagrave, P. Espaillat, K. Honjo, B. Hunt, M. Kortelainen, R. Lalonde, D. Laxton, E. Mavroeidi, D. Muir, S. Mursula, and S. Snudden, 2015. “The Flexible System of Global Models – FSGM,” IMF Working Paper No. 2015-64.
Blog Posts
“Oil Exporters Should NOT Price Level Target,” Economics and Policy Blog, The John Deutsch Institute for the Study of Economic Policy, Queen’s University, March 31, 2017.
“How we decided on 2% fiscal stimulus during the Great Recession,” Economics and Policy Blog, The John Deutsch Institute for the Study of Economic Policy, Queen’s University, June 21, 2017.
In the News:
“This Means (Trade) War: How Will U.S. Tariffs Impact Canadians?” Inspiring Conversations, Wilfrid Laurier University. April 4, 2025.
“Tariffs threaten Canadian aluminum critical to U.S. energy transition” March 10, 2025, Trellis, By Andrew Kaminsky.
Stephen Snudden 