Zoé Hamstead
The recent debacle over Reinhart and Rogoff’s (2010) National Bureau for Economic Research working paper raises a number of ethical issues for policy analysis generally. Their study, based on 200 years of data on GDP growth and debt in 44 countries finds that average growth rates fall dramatically above a 90% debt/GDP ratio. With the recent release of a PERI working paper that replicated the Reinhart Rogoff study (Herndon et al., 2013), but found that the authors made a number of serious methodological errors, other, less technical issues surrounding the study have emerged. In particular, the extent to which the original findings permeated policy discussions around austerity measures raises the issue of to what extent researchers should be cautious about raising policy implications of their findings, and the role of media dissemination.
Reinhart and Rogoff’s study makes a claim about the direction of causality which is arguably linked to their normative stance on economic policy, but does not adequately allow for alternative interpretations or freely admit to a political bias. Their statistical analysis does not directly examine the direction of causality, yet they speculate in their concluding remarks that countries with high debt will need to “tighten fiscal policy in order to appear credible to investors…” Moreover, they argue that countries with high debt are vulnerable to unexpected financial crises. Yet, they do not balance this speculation with alternative possibilities, such as that raised by Paul Krugman and others. Krugman argues that historical examination can lead one to see that the extreme cases of high debt and slow growth, such as occurred in Japan and Italy, are instances in which growth slowdown has led to higher debt (Krugman, 2013). Warwick & Pettigrew (1983) argue that researchers should be cautious of using expansive interpretations and clearly separate interpretations from the actual research findings, which in Reinhart and Rogoff’s study, were mere associations accompanied by arguably weak causal hypotheses. Allowing the subjects of bias, personal values and normative considerations to enter into the discussion of policy implications would draw the reader to question assumptions made by the researchers, and enrich the dialogue about policy implications (Amy, 1984).
As Herdon et al. point out, the Reinhart Rogoff study was used in testimony before the Senate Budget Committee, their tables were reprinted in publications and presentations of the Centre for Economic Policy Research and Peter G. Peterson Institute for International Economics, was cited in numerous popular media outlets such as The Economist, Wall Street Journal, New York Times, National Public Radio, and cited in more than 500 academic publications (Herndon et al., 2013). What responsibility do the authors have of controlling the message their research communicates to these audiences and specifying the extent to which the findings can be interpreted for policy guidance? Warwick and Pettigrew (1983) argue that social scientists should not engage in publicity campaigns for their policy preferences as part of disseminating research results, should emphasize limitations, and again, separate empirical and normative issues. Audiences of mass media publications will be unable to distinguish between normative and empirical issues unless they are made clear in reporting; these audiences cannot be expected to review research methodology and design of every reported finding in order to ascertain for themselves the research design limitations. This problem relates to the social science research profession’s roots in technocratic optimism, and a desire to see the social scientist as someone who is value-free and therefore capable of advising policy without being influenced by values (Amy, 1984). This false perception of the researcher degrades the ability of policymakers and public stakeholders to engage in informed debate about public policy issues.
References
Amy, D. J. (1984). Why Policy Analysis and Ethics Are Incompatible. Journal of Policy Analysis and Management, 3(4), 573–591.
Herndon, T., Ash, M., & Pollin, R. (2013). Does High Public Debt Consistently Stifle Economic Growth?: A Critique of Reinhart and Rogoff. Political Economy Research Institute. Amherst, MA.
Krugman, P. (2013). Grasping at Straw Men. The New York Times: The Opinion Pages. Retrieved April 29, 2013, from http://krugman.blogs.nytimes.com/2013/04/26/grasping-at-straw-men/
Reinhart, C. M., & Rogoff, K. S. (2010). Growth in a Time of Debt.
Warwick, D. P., & Pettigrew, T. F. (1983). Toward Ethical Guideliens for Social Science Research in Public Policy. In D. Callahan & B. Jennings (Eds.), Ethics, The Social Sciences and Policy Analysis (pp. 335–369). Hastings-on-Hudson, New York: The Hastings Center.
The recent debacle over Reinhart and Rogoff’s (2010) National Bureau for Economic Research working paper raises a number of ethical issues for policy analysis generally. Their study, based on 200 years of data on GDP growth and debt in 44 countries finds that average growth rates fall dramatically above a 90% debt/GDP ratio. With the recent release of a PERI working paper that replicated the Reinhart Rogoff study (Herndon et al., 2013), but found that the authors made a number of serious methodological errors, other, less technical issues surrounding the study have emerged. In particular, the extent to which the original findings permeated policy discussions around austerity measures raises the issue of to what extent researchers should be cautious about raising policy implications of their findings, and the role of media dissemination.
Reinhart and Rogoff’s study makes a claim about the direction of causality which is arguably linked to their normative stance on economic policy, but does not adequately allow for alternative interpretations or freely admit to a political bias. Their statistical analysis does not directly examine the direction of causality, yet they speculate in their concluding remarks that countries with high debt will need to “tighten fiscal policy in order to appear credible to investors…” Moreover, they argue that countries with high debt are vulnerable to unexpected financial crises. Yet, they do not balance this speculation with alternative possibilities, such as that raised by Paul Krugman and others. Krugman argues that historical examination can lead one to see that the extreme cases of high debt and slow growth, such as occurred in Japan and Italy, are instances in which growth slowdown has led to higher debt (Krugman, 2013). Warwick & Pettigrew (1983) argue that researchers should be cautious of using expansive interpretations and clearly separate interpretations from the actual research findings, which in Reinhart and Rogoff’s study, were mere associations accompanied by arguably weak causal hypotheses. Allowing the subjects of bias, personal values and normative considerations to enter into the discussion of policy implications would draw the reader to question assumptions made by the researchers, and enrich the dialogue about policy implications (Amy, 1984).
As Herdon et al. point out, the Reinhart Rogoff study was used in testimony before the Senate Budget Committee, their tables were reprinted in publications and presentations of the Centre for Economic Policy Research and Peter G. Peterson Institute for International Economics, was cited in numerous popular media outlets such as The Economist, Wall Street Journal, New York Times, National Public Radio, and cited in more than 500 academic publications (Herndon et al., 2013). What responsibility do the authors have of controlling the message their research communicates to these audiences and specifying the extent to which the findings can be interpreted for policy guidance? Warwick and Pettigrew (1983) argue that social scientists should not engage in publicity campaigns for their policy preferences as part of disseminating research results, should emphasize limitations, and again, separate empirical and normative issues. Audiences of mass media publications will be unable to distinguish between normative and empirical issues unless they are made clear in reporting; these audiences cannot be expected to review research methodology and design of every reported finding in order to ascertain for themselves the research design limitations. This problem relates to the social science research profession’s roots in technocratic optimism, and a desire to see the social scientist as someone who is value-free and therefore capable of advising policy without being influenced by values (Amy, 1984). This false perception of the researcher degrades the ability of policymakers and public stakeholders to engage in informed debate about public policy issues.
References
Amy, D. J. (1984). Why Policy Analysis and Ethics Are Incompatible. Journal of Policy Analysis and Management, 3(4), 573–591.
Herndon, T., Ash, M., & Pollin, R. (2013). Does High Public Debt Consistently Stifle Economic Growth?: A Critique of Reinhart and Rogoff. Political Economy Research Institute. Amherst, MA.
Krugman, P. (2013). Grasping at Straw Men. The New York Times: The Opinion Pages. Retrieved April 29, 2013, from http://krugman.blogs.nytimes.com/2013/04/26/grasping-at-straw-men/
Reinhart, C. M., & Rogoff, K. S. (2010). Growth in a Time of Debt.
Warwick, D. P., & Pettigrew, T. F. (1983). Toward Ethical Guideliens for Social Science Research in Public Policy. In D. Callahan & B. Jennings (Eds.), Ethics, The Social Sciences and Policy Analysis (pp. 335–369). Hastings-on-Hudson, New York: The Hastings Center.
RSS Feed