Modern cliometric studies use dummy variables to measure the effects of institutions. The dummy variable approach can be misleading, as illustrated by recent research on the impact of colonial rule on borrowing terms. We show how trying to measure a 'colonial effect' without an analysis of the financial consequences of political subjection can be misleading. The main effect of the British Empire was to remove the default risk. Establishing how this was done, and with what effects, should take us closer to a proper understanding of the effect of empire.e hr_536 385..407 I don't care to be part of a club that accepts people such as myself.Groucho Marx T he goal of this article is to provide an assessment of methodological pitfalls in works aimed at measuring the economic benefits of the British Empire. Due to the increasing interest that economists have taken in institutions, the study of colonial legacy has attracted renewed attention in recent years. Thanks to contributions by Acemoglu et al., research has recently revived, in the language of formal models and with the tools of modern econometrics, issues that have occupied historians' minds for a long time, such as the influence of imperial institutions on countries' long-run economic performance. 2 These questions had been discussed in the past, in a different but related guise, not least in Davis and Huttenback's seminal book on the political economy of the British Empire. 3 Among the varied issues raised in their work was the influence of empire subjection on borrowing costs. Computing average borrowing rates in British colonies and in sovereign countries, Davis and Huttenback found that colonies tended to have much lower interest rates than sovereign nations and suggested that this was a benefit of colonization. Combining this number with the amount of capital transferred to colonies provided a measurement of the social surplus that colonies secured. Social surpluses have traditionally been used to compute the effect of switching to a 'superior' technology (such as railways) and abandoning an 'inferior' technology 1 Frédéric Zumer was closely involved in related and earlier work and is gratefully thanked for the time and insight he contributed to this spin-off from our earlier collaborative research. We are grateful to Knick Harley, Larry Neal, and conference participants in Oxford, Paris, and Edinburgh for their comments and suggestions.We also thank Jane Humphries and four referees for their advice in preparing the final draft. All misinterpretations remain ours.2 Acemoglu, Johnson, and Robinson, 'Colonial origins'.

No abstract