• Economics cannot define uniquely correct policy

    Though I’ve written before about the limitations of welfare economics, it’s valuable to be reminded of them. Paul Kelleher brought to my attention a brief paper by Daniel Hausman and Michael McPherson (pdf) that serves that role.

    Although the value of freedom lurks inarticulately within the standard economic argument for cash benefits, the argument remains within the terms set by orthodox economic theory. There is no mention of needs, of the presuppositions of individual dignity, of opportunity, of rights or of fairness. There is no concern with the moral reasons that make individuals willing to pay taxes to provide such benefits. Are people motivated by a general concern to satisfy the preferences of others, or do they instead see themselves as obligated to help others in need? Might they regard people as having rights to food or medical care which justify taxing others? What freedoms and opportunities do justice demand? These are hard questions even to ask within the framework economists employ.

    In the background is the familiar trade-off between efficiency and equality. Moreover, those terms–efficiency, equity–are themselves not uniquely definable. Most of Hausman’s and McPherson’s paper is devoted to exploring what underlies efficiency, namely what “utility” or “satisfaction of preferences” means and is taken to mean by economists.  They write of the domain of neo-classical welfare economics, which offers one sense of efficiency (that which would be maximized by perfect markets) that is broadly invoked today in defense of markets and limited government.

    But few are exclusively interested in economic efficiency, however defined. There are some things we value that extend beyond its purview: freedom, rights, fairness, and the like. For that reason, it is never valid to argue that action A is the correct one because it leads to greater economic efficiency than action B. It may be so that A is more efficient that B. But that does not make it “correct.” If A also restricts rights we cherish or reduces freedom we hold dear, many would reject it.

    Noncontroversial examples are not hard to find. So long as I’m not a convicted felon, my freedom to cast a vote for president is an inviolable right. I can’t legally sell my vote for president to you, for example, though I might prefer the cash you’d be willing to offer in exchange for it (the hallmark of an efficiency increasing transaction). Some degree of health care is also a right. A hospital doesn’t legally turn away a penniless, uninsured patient in mortal danger no matter what choices he may have made in the past. Even though the hospital exchanges a service for a price below cost (something that would not happen in a free market), very few (though not zero, perhaps) people would find it acceptable otherwise.

    What other rights and freedoms do you wish to protect even at the expense of economic efficiency? Though your list may not agree with mine, so long as our lists are not empty we have demonstrated that economic efficiency is not the only metric of value. Economics offers useful, if limited, tools. They should neither be ignored nor over-applied. At a minimum, they are insufficient to determine anything like uniquely correct policy.

    PS: If you’re following the welfare economics discussion between Paul Kelleher and Julian Jamison, Paul has a follow-up post. Links to earlier posts are here.

    • Great paper. Thanks for posting it.

      Given all of the glaring limitations inherent in neoclassical economics – complete, perfect markets inhabited by generic automatons who use perfect information and share a generic set of completely rational preferences etc, etc, etc –(concise overview here: it’s astonishing that it became a template for evaluating different modes of delivering healthcare, where subjective value judgments in a wide variety of different dimensions are at the heart of nearly every medical decision.

      Astonishing until you realize that any attempt to coordinate supply and demand in the healthcare market using a centralized bureaucratic mechanism *has* to use such models. You have to naively assume a generic set of preferences shared by everyone in order to make the optimization of all of the variables at play mathematically tractable in whatever simplified model you have constructed to do so.

      The assumption that everyone has a generic set of preferences is silly, even without factoring in the knowledge and coordination problems that make would make models based on them unworkable in reality even in a society composed of completely identical people.

      What’s truly nauseating and offensive is the unstated premise that whomever is constructing the centrally planned system built on these absurd models has the right to over-ride every individual’s desires, values, preferences and convictions. The Fatal Conceit strikes again.

      One more reason why the resources and incentives at the heart of medical decision making should be left in the hands of patients as much as possible. I’ll take my chances with vouchers and cost sharing over an algorithmic straight-jacket constructed from a bastardized econometric simulation any day.

      • I understand your sentiments about the limitations of neoclassical welfare economics, but I don’t understand how the conclusions follow. Since the economics theories are more supportive of markets than government, how does rejection of them also lead you in the same direction?

    • From my vantage point, neoclassical economics, conceived by Keynes, formalized by Samuelson, and tweaked by the likes of Stiglitz late in the game only supports markets inasmuch as “markets” are conceived as maximally efficient mechanisms for utility optimizing agents to satisfy their preferences.

      Once you identify the conditions that optimize whatever arbitrary set of preferences you define as “welfare,” the game is over, and it doesn’t follow from the premises in the model that you need to do much more than force everyone to behave in a way that is consistent with your model to achieve your ends.

      Unfortunately for the likes of Samuelson and his disciples, humans don’t seem to behave exactly like rational utility maximizing agents that consume and produce units of GDP, and the outcomes that emerge when real humans “truck, barter, and exchange” according to their own sets of values and preferences result in sub-optimal outcomes relative to their models. This seems to have lead many of them to conclude that democracy and market efficiency, as narrowly defined above, conflict with one another. It has also lead many of them to embrace mechanism that centralize authority in the expectation that doing so will correct the inefficiencies that emerge when people are free to make their own choices.

      There are few domains where all of the above tendencies are more clearly evident than health care. First and foremost is the arbitrary determination of what the optimal level of spending on health care should be relative to GDP. Then there’s the collection of high-level statistical aggregates that supposedly correlate with “welfare,” and a cataloging of the divergence from an arbitrary optimum achieved elsewhere – all of which are marshalled into a set of argument for constructing centralized pricing and destribution mechanisms that will force doctors and patients into behaviors that will result in real outcomes that are more consistent with “maximum welfare” as the people conducting the analysis define it.

      It’s possible to reject both single payer and ostensibly “market” friendly policy prescriptions derived from neoclassical premises on both technical and moral grounds, for largely the same reasons.