This paper asks and answers the question of whether Kahneman's planning fallacy or Hirschman's Hiding Hand best explain performance in capital investment projects. I agree with my critics that the Hiding Hand exists, i.e., sometimes benefit overruns outweigh cost overruns in project planning and delivery. Specifically, I show this happens in one fifth of projects, based on the best and largest dataset that exists. But that was not the main question I set out to answer. My main question was whether the Hiding Hand is "typical," as claimed by Hirschman. I show this is not the case, with 80 percent of projects not displaying Hiding Hand behavior. Finally, I agree it would be important to better understand the circumstances where the Hiding Hand actually works. However, if you want to understand how projects "typically" work, as Hirschman said he did, then the theories of the planning fallacy, optimism bias, and strategic misrepresentation - according to which cost overruns and benefit shortfalls are the norm - will serve you significantly better than the principle of the Hiding Hand. The latter will lead you astray, because it is a special case instead of a typical one.
I want to thank the editor for giving me the opportunity to respond to the comments made above by Philipp Lepenies, Graham Room, and Lavagnon Ika on my paper "The Fallacy of Beneficial Ignorance: A Test of Hirschman's Hiding Hand," printed in World Development,vol. 84. First, I will briefly mention some of the points on which I agree with my critics. Second, I will respond to their main criticisms.
I agree regarding the richness and originality of Hirschman’s work and that he was a leading 20th century intellectual and economist, well worth reading today. His ideas may still open up a range of productive and unsuspected new vistas, as maintained by Room (pp. 1,2). I also agree that many of Hirschman’s observations regarding the principle of the Hiding Hand, and the fundamentally anti-rationalist worldview they represent, are highly innovative, for instance that success may be more often stumbled upon than carefully planned for, as argued by Lepenies (p. 5). Or that the World Bank’s push for cost-benefit analysis as a synoptic project evaluation method might have some use, but was ultimately a misguided attempt at comprehensive quantification, as pointed out by Ika (p. 38). Finally, I agree that “the Hiding Hand is a possible empirical occurrence and it does happen,” as observed by Ika (p. 4), i.e., creative project managers are sometimes able to generate benefit overruns that are larger than cost overruns, securing the viability of projects.
My critics have the grace to similarly point out many instances where they agree with what I say. For example, Ika (p. 4) acknowledges that until he began working on his rejoinder to my paper he was under the impression that the principle of the Hiding Hand was not empirically testable, but now he shares my view that the principle is not only a theory for explanation but also an empirically testable hypothesis. Room (p. 2) says he finds it difficult to disagree with me that Hirschman had insufficient empirical evidence for affirming the Hiding Hand, just as Ika (p. 14) agrees with the limitations I identify for Hirschman’s work, including his small sample of 11 projects. “Admittedly, these methodological limitations plagued Hirschman’s findings,” Ika (p. 14) concedes.
Finally, Ika (p. 4) further agrees with my critique of Hirschman for having sampled on the independent variable in his study of the Hiding Hand.
When it comes to dissent, my critics disagree not only with me, but also among themselves, which weakens their critique, needless to say. For instance, Ika (p. 4) accepts my claim that the principle of the Hiding Hand is a testable theory, as mentioned above, whereas Lepenies (p. 7) rejects this. Furthermore, Lepenies (p. 2) takes at face value Hirschman’s success story about the paper mill in Pakistan, and uses this to argue that the Hiding Hand works. In contrast, Ika (p. 8, note 6) rightly observes that the paper mill turned out to be “a disastrous development failure.” This is the case, too, for several other projects that Hirschman identified as successes and took to support the Hiding Hand, but that turned into failures shortly after Hirschman completed his study, falsifying the Hiding Hand. Surely this is a problem, as pointed out in my paper. But Lepenies glosses it over, as did Hirschman when confronted with the facts.
Lepenies (p. 6) further confuses normative and explanatory theory when he quotes Hirschman as having a “dislike for general principles” and theory. Hirschman, as quoted by Lepenies, was explicitly talking about normative theory, in terms of “prescriptions,” “recipe [s],” and “therapy” (p. 6;Hirschman 1998: 88, 110). It is correct that Hirschman was against theory used in this manner, i.e., for design, where law-like ideas are used to prescribe and plan social action, or even the social order. Hirschman generally saw such social engineeringwhether driven by experts or revolutionaries -as highly problematic and bound to fail. But Hirschman had no issue with explanatory theory, not for himself and not for other social scientists. And the principle of the Hiding Hand was developed by Hirschman as explanatory theory, with truth claims and causal mechanisms. By not distinguishing between normative and explanatory theory, as Hirschman does, Lepenies misses this important point and is led to wrongly claim that Hirschman was against theory as such, when he was only against normative theory. Hirschman (1994: 277-78) explicitly referred to his work as “theory building” and added, “I bristle a bit when I am pigeonholed as ‘atheoretical’ or ‘antitheoretical’.” Hirschman would bristle, and feel pigeonholed, if he read Lepenies’s take on his work. Room (p. 2) claims I have “a particular view of theory -as hypotheses that can be subjected to quantitative empirical assessment in large datasets.” This is incorrect. If Room had read my paper more carefully, he would not have had to guess at my view of theory, and get it wrong. He would have found that I
This content is AI-processed based on open access ArXiv data.
