Institutional scholars often presume that the selection of a particular structural design necessarily shapes resulting policy outputs and outcomes. This phenomenon is an inherently political choice, yet we know little about its consequences regarding bureaucratic performance in a democracy. In this study, we examine whether the quality of agency outputs varies in relation to its institutional design. We empirically investigate this issue by examining the quality of current-year macroeconomic and fiscal projections generated by both presidential (Council of Economic Advisers and Office of Management and Budget) and congressional (Congressional Budget Office) support agencies, as well as an independent regulatory commission (Federal Reserve). Our statistical results uncover trivial cross-agency differences regarding bias and accuracy contained in this type of policy information. While varying degrees of political insulation might affect the durability of administrative agencies ex post, our empirical evidence rejects the politics of structural design thesis for short-termoriented policy analysis. Specifically, bureaucratic task outputs are generally unaffected in an ex ante fashion by the extent to which an agency’s institutional structure is insulated from political influence. Our empirical evidence suggests that these particular agencies’ concern with reputational considerations are fairly homogeneous and thus outweigh the varying political pressures that they confront attributable to the institutional structure that they operate under. Is a public bureau’s performance ‘‘hardwired’’ by the structural design choice made by politicians? The answer to this question is of utmost importance to those interested in the nexus between democratic accountability and the administrative state because it tells us not only whether politicians are able to exert control over specialized experts in the bureaucracy A previous version of this article was prepared for delivery at the 2003 Public Choice Society Meetings, Nashville, Tenn., March 21–23, 2003. We thank David Lewis, Florenz Plassman, and the anonymous JPART reviewers for helpful comments on earlier versions of this article. Address correspondence to George A. Krause at george.krause@sc.edu. doi:10.1093/jopart/mui038 Advance Access publication on March 2, 2005 a The Author 2005. Published by Oxford University Press on behalf of the Journal of Public Administration Research and Theory, Inc. All rights reserved. For permissions, please e-mail: journals.permissions@oupjournals.org. JPART 15:281–306