Policymaker Ignorance…So What? (Part One)
How to Respond to Policymaker Ignorance: The “Epistemic-Mechanistic” Approach
This is the first part of a multi-part series that considers potential responses to the problem of policymaker ignorance, possible methods of mitigating policymaker ignorance, and appropriate attitudes to adopt with regard to ignorant policymakers.
In this newsletter, I consider one method, which I call the “epistemic-mechanistic” approach, whereby political systems might be reformed, at least in principle and perhaps also in practice, to minimize the consequences of policymaker ignorance.
In another newsletter that I hope to post this weekend, I will consider another method, the “constitutional” approach, which might be deployed independently of or in tandem with the epistemic-mechanistic approach, to address the problem of policymaker ignorance.
In the final newsletter of the series, to be posted early next week, I will defend “policy skepticism” as the appropriate attitude to adopt toward ignorant policymakers and their policy machinations.
I am occasionally asked some variation on the following question:
“Yes, of course, policymakers are ignorant…so what? What are the practical implications of the problem of policymaker ignorance? How should we respond to the fact that policymakers are ignorant, either in general or in some particular case, regarding some particular policy goal?”
The first point to note – and immediately set aside – is the philosopher’s traditional worry about the relationship between descriptive premises and normative conclusions, and the fact that the latter cannot be directly inferred from the former without the addition of some normative premises. My analysis of policymaker ignorance is purely descriptive. Unless one adds some further normative assumption(s), e.g., that political action should always aim to promote goals that constituents demand (or, alternatively, goals in constituents’ interests), nothing about how one should respond to policymaker ignorance, either in general or in a given case, can be inferred from the descriptive analysis in isolation.
This pedantic point explicated and some similar normative premise duly assumed, the question becomes how we should react to the fact of policymaker ignorance, given the normative requirement to (e.g.) aim to promote goals in constituents’ interests.
The next point to emphasize is that the problem of policymaker ignorance cannot be solved once and for all. The problem of policymaker ignorance is just a specific instance of the general and eternal problem of human ignorance, the manifest fact that human beings are neither omniscient nor omnipotent. There are always goals that humans are too ignorant or incapable to realize on their own reserves. As I’m constantly reminding readers, policymakers are mere human beings. There are, therefore, potential policy goals that policymakers cannot deliberately realize given their limited epistemic and practical capacities. A full-fledged solution to the problem of policymaker ignorance would mean making policymakers omniscient and omnipotent. Human beings cannot be made (generally) omniscient and omnipotent, so the problem of policymaker ignorance cannot be universally and permanently solved.
As the parenthetical qualification suggests, however, we might be able to make policymakers functionally omniscient and omnipotent with respect to specific problems or particular policy goals. Indeed, if the social problem or policy goal is simple enough, policymaker ignorance may be no problem at all.
The problem of policymaker ignorance has two distinct, but mutually reinforcing, sources. The problem exists not only because policymakers, like all other human beings, are cognitively limited, neither omniscient nor omnipotent. The problem arises also because constituents tend to demand (or be interested in) the realization of policy goals that are decidedly not simple, which are, in fact, extremely complicated, to realize. Even if policymakers were much less ignorant and far more capable than they in fact are, policymaker ignorance would remain a problem inasmuch as constituents continued to demand the pursuit of goals the requirements of which eclipsed policymakers’ knowledge and abilities.
In the last analysis, the problem of policymaker ignorance is a mismatch between the knowledge requirements of the kinds of goals that constituents tend to want policymakers to pursue and the knowledge that policymakers possess. Were constituents to limit their demands to relatively simple policy goals, policymaker ignorance would be less of a problem than it is.
So, the problem of policymaker ignorance might be mitigated from either of two directions, as it were. We might address the problem by improving policymakers’ relevant knowledge and capacities, or we might address the problem by constraining constituents’ policy demands to relatively simple objectives (or we might combine these methods and approach the problem from both directions).
In the fifth and sixth chapters of my book, F. A. Hayek and the Epistemology of Politics: The Curious of Economics, I discuss, respectively, the epistemic-mechanistic and constitutional approaches to mitigating the problem of policymaker ignorance, which map on to these two “directions” from whence the problem might be addressed. I discuss the first approach here and will save discussion of the second approach for the next newsletter in this multi-part series.
The Epistemic-Mechanistic Approach to the Problem of Policymaker Ignorance
The epistemic-mechanistic approach builds on F. A. Hayek’s arguments concerning the price system as an epistemic device that provides individuals with much of the knowledge required to adapt their individual plans to changing economic circumstances. Thanks to the price system, in order to adjust their demand for, say, coffee, to prevailing circumstances, individual coffee drinkers need know relatively little about the current state of the coffee market, whether, e.g., demand has increased because a novel use for coffee beans has been discovered or supply has decreased because of a drought in the coffee-growing regions of South America. They need only observe the changing prices of their favorite coffee (and other) drinks and adjust accordingly. The price system economizes on the knowledge that would otherwise be required for individuals to adapt rapidly and efficiently to changing economic circumstances.
This basic Hayekian idea – that there are (or can be) epistemic devices that help individual actors adapt their plans to changing circumstances relevant to those plans – has been applied beyond the narrow realm of economic decision-making.
Contemporary Hayekians, William Butos and Thomas McQuade, have applied this idea to science. They argue that science’s publication-citation-reputation (PCR) system, provided it is not manipulated by exogenous (most often, government, but possibly, corporate) actors, provides to individual scientists much of the knowledge they need to adapt their plans to changing circumstances in relevant scientific domains (i.e., scientific disciplines, fields, research programs).[1] Thanks to the PCR system, in order to adjust their relevant activities, individual scientists need know relatively little about the current state of the domains in which they operate. They need only observe the effects over time of new publications on authors’ positive and negative citation counts, and the concomitant effects of these citations on the authors’ reputations. The PCR system economizes on the knowledge that would otherwise be required for individual scientists to rapidly and efficiently adapt to changing scientific circumstances.[2]
A democratic election is also a kind of epistemic mechanism, albeit a far less efficient one than either the price or PCR systems. By observing the results of elections, policymakers learn a bit about constituents’ preferences, particularly regarding contests between specific candidates, but not nearly as much as policymakers need to know to deliberately realize goals associated with their constituents’ preferences.
What is required, but lacking, are epistemic devices that communicate in an efficient manner to policymakers not only 1) the interests, policy preferences, and demands of their constituents, but also 2) how to realize goals associated with these interests, preferences, and demands.
Existing democratic mechanisms are utterly impotent in the latter regard. Whatever they express to policymakers about constituents’ preferences, they communicate exactly nothing to policymakers about how to realize goals in line with these preferences. If policymakers do not already know how to realize the goals that their constituents demand, they cannot acquire the missing knowledge by watching election results.
So, part of the epistemic-mechanistic approach to the problem of policymaker ignorance involves the development of epistemic mechanisms that convey both propositional know-that and non-propositional know-how from constituents to policymakers.
However, this would solve only part of the problem. Such devices would facilitate the communication of relevant knowledge from constituents to policymakers, but we would also need epistemic mechanisms to convey relevant knowledge from policymakers back to their constituents.
In particular, we would need devices to convey to constituents knowledge regarding 1) exactly which goals policymakers were in fact pursuing and 2) how well their policies were realizing these goals at any given time.
Without such mechanisms, it is exceedingly difficult for constituents to monitor policymakers and their policy successes and, perhaps more significantly, their policy failures. It is always possible, especially in our modern media-driven world, for policymakers to pretend to pursue and, up to a point, to pretend to realize, goals associated with their constituents’ interests.
Even this would probably not be enough to adequately mitigate policymaker ignorance, however. Whatever epistemic mechanisms we might conjure, policymaker ignorance would remain a problem as long as constituents continued to demand the pursuit of goals the realization of which exceeded policymakers’ epistemic capacities.
This is another respect in which existing democratic mechanisms fail as epistemic devices. Constituents learn little to nothing about policymakers’ relevant knowledge and capacities, or, more to the point, about their relevant ignorance and incapacities, from observing democratic politics.
So, in order to encourage constituents to demand only the pursuit of goals the realization of which fell within policymakers’ ken and control, some mechanisms would be required to convey knowledge from policymakers to constituents concerning the former’s epistemic and practical limitations.
Epistemic mechanisms such as those described here would promote a kind of equilibrium between constituents’ demand for and policymakers’ supply of specific policies. They would encourage both constituents to demand only policies that policymakers could effectively supply and policymakers to supply only policies that constituents demanded.
Unfortunately, the purposeful development of such epistemic devices would seem to fall well beyond policymakers’ existing epistemic capacities. That is, in other words, mitigating the problem of policymaker ignorance via the epistemic-mechanistic approach would likely fail, unless spontaneous forces intervened, due to the same problem of policymaker ignorance such an approach would be meant to mitigate.
More to come anon. Stay tuned.
[1] In collaboration with Butos and McQuade, I have applied this analysis to the history of nutrition policy in the United States.
[2] Interested readers should check out Butos and McQuade’s recently published book, Hayekian Systems: Research into the Structure of Social Interaction