American Political Science Review

Research Article

Benchmarking across Borders: Electoral Accountability and the Necessity of Comparison

MARK ANDREAS KAYSERa1 c1 and MICHAEL PERESSa2 c2

a1 Hertie School of Governance

a2 University of Rochester

Abstract

When the economy in a single country contracts, voters often punish the government. When many economies contract, voters turn against their governments much less frequently. This suggests that the international context matters for the domestic vote, yet most research on electoral accountability assumes that voters treat their national economies as autarkic. We decompose two key economic aggregates—growth in real gross domestic product and unemployment—into their international and domestic components and demonstrate that voters hold incumbents more electorally accountable for the domestic than for the international component of growth. Voters in a wide variety of democracies benchmark national economic growth against that abroad, punishing (rewarding) incumbents for national outcomes that underperform (outperform) an international comparison. Tests suggest that this effect arises not from highly informed voters making direct comparisons but from “pre-benchmarking” by the media when reporting on the economy. The effect of benchmarked growth exceeds that of aggregate national growth by up to a factor of two and outstrips the international component of growth by an even larger margin, implying that previous research may have underestimated the strength of the economy on the vote.

Correspondence:

c1 Mark Andreas Kayser is Professor of Applied Methods and Comparative Politics, Hertie School of Governance, Friedrichstraße 180, 10117 Berlin, Germany (kayser@hertie-school.org).

c2 Michael Peress is Assistant Professor of Political Science, University of Rochester, Harkness Hall 326, Rochester, NY 14627 (mperess@mail.rochester.edu).

Footnotes

We thank Jeff Arnold, Evelyn Sauniere, and CJ Yetman for excellent research assistance; Ignacio Urquizu Sancho and Stuart Soroka for kindly sharing their data with us; Thomas Gschwendt for pointing out in a previous paper that we had reinvented the statistical wheel; Christopher Anderson, Larry Bartels, Pablo Beramendi, Sarah Birch, Raymond Duch, Timothy Hellwig, Greg Huber, Marko Klasnja, Gabe Lenz, Rene Lindstaedt, David Rueda, Joshua Tucker, Jack Vowles, and Chris Wlezien for helpful comments; and Tourun Dewan and Pepe Fernandez for an early conversation.

Metrics