The Valence of Unintended Consequences

09.08.2024 01:00

How should we think about the problem of unanticipated consequences? And what are the implications for the possibility of unintended consequences regarding top-down, technocratic policy initiatives that aim to mitigate targeted social problems?

For example, I’ve occasionally heard it argued that we shouldn’t be too worried about unanticipated consequences of interventions, because unanticipated consequences don’t have to be bad. They might be good!

Albert Hirschman made this claim in his book The Rhetoric of Reaction, where he advanced two claims – the idea that “purposive social action” leads to adverse unintended consequences only “occasionally,” and that “it is obvious that there are many unintended consequences or side effects of human actions that are welcome rather than the opposite.”

In his book Power Without Knowledge, Jeffrey Friedman argued that Hirschman’s case falls flat on both points. To start, “Hirschman’s first claim is a generalization of naïve technocratic realism. It tacitly appeals to the reader’s agreement that if we tally up our first-order assessments of technocratic wins and losses, technocracy comes out ahead, begging the epistemological question by assuming the reliability of these tallies.” Given that the ability to accurately tally such things is the very point under dispute, trying to resolve the dispute by appealing to those tallies would indeed be a textbook case of question-begging.

The second claim Hirschman makes might provide a basis for defending technocracy, but Hirschman fails to adequately defend it, Friedman argues:

To counteract worries about the adverse unintended consequences of technocracy he would have had to contend that the unanticipated consequences of technocrats actions will tend to be beneficial, not merely that they may be beneficial. Thus he would have had to argue not that “there are many unintended consequences or side effects…that are welcome,” but that, even though policymakers may be ignorant of the side effects of their actions, something or other ensures that these effects will be more welcome than unwelcome overall. This claim would not be naively realistic, as it would gesture toward a second-order factor or factors that might explain the on-balance beneficial valence of unintended consequences. However, since Hirshcman does not specify what this factor or factors might be, it is hard to imagine how the claim could be supported, saved through a quasi-religious providentialism.

That is, Friedman argues that if one wants to salvage the argument in favor of technocracy in situations where technocrats lack what Friedman called “type 4 knowledge” – knowledge that the costs of a technocratic policy (consisting of both the costs of implementing the solution as well as any unanticipated and unintended costs) will not be higher than the costs of the initial problem – merely pointing out that unanticipated outcomes could in principle be beneficial is simply inadequate. One would need to provide some positive grounds for believing that unintended consequences will have an overall tendency to be beneficial.

In his book, Friedman simply adopts the fairly modest premise that “while the tendency of unintended consequences might be either more harmful than beneficial or more beneficial than harmful, we do not know which is the case…The question, then, is whether our ignorance of the valance is more damaging to epistemological criticisms of technocracy or to defenses of it.” He argues that the simple fact of uncertainty is fatal to the argument for technocracy, and to say otherwise “would fly in the rationalist face of technocracy, for it would license the adoption of policies that – like policies pulled from a hat – are justified not by knowledge, but by hope.” Appealing to the mere possibility that unintended consequences might be beneficial as a defense of technocracy actually rebuts the argument in favor of technocracy.

Friedman left the question of how to judge the valence of unanticipated consequences unexamined – his case didn’t depend on making a positive case that the valence will be neutral or even negative. But I want to look a step further than Friedman did – do we have reason to think that valence of unintended consequences will tend to be positive, neutral, or negative? And on what basis would we examine such a claim?

Friedman argues (correctly, I believe) that we need to make a second-order argument on this issue. A second-order argument is one that focuses on systemic reasoning about the workings of a system, rather than first-order arguments where one attempts to tally up points on a case by case basis. For example, one might argue that government operates inefficiently compared to market activity by first-order means, perhaps pointing out that building a public restroom consisting of a “tiny building with four toilets and four sinks” cost the taxpayers of New York City over two million dollars, while by contrast “privately managed Bryant Park, in the middle of Manhattan, gets much more use and its recent bathroom renovation cost just $271,000.”

But the same article also makes a second-order argument about the systematic differences under which state and private enterprises operate, arguing that since “government spends other people’s money, it doesn’t need to worry about cost or speed. Every decision is bogged down by time-wasting ‘public engagement,’ inflated union wages, and productivity-killing work rules.” So we can distinguish between the first order argument (examining specific cases) and the second order argument (comparative institutional analysis). Thus, the article uses a first-order case as an example of government being wildly wasteful and inefficient in what it does, and also offers a second-order argument for why this sort of disparity is systemic rather than random.

In my next post, I will be considering a second-order argument about the valence of unintended consequences, and whether we should expect them to have a tendency to be positive, neutral, or negative.

(2 COMMENTS)