Economics: Art or Science?

Last week a good friend emailed me this NY Times opinion piece by David Brooks. I'm not a big fan of Brooks, but I think this is worth weighing in on because I fear it may be the beginning of a concerted effort to discredit economics just as we teeter on the verge of an important course correction.

The text supplied by the NY Times under the link in the email read:

After the crash, economics will have to acknowledge its methodological failures and come to resemble an art rather than a science.

Gosh, I hope not.

Brook's article is pretty apocalyptic stuff. Here are his final incendiary paragraphs:

...economists are taking baby steps into the world of emotion, social relationships, imagination, love and virtue. In Act V, I predict, they will blow up their whole field.

Economics achieved coherence as a science by amputating most of human nature. Now economists are starting with those parts of emotional life that they can count and model (the activities that make them economists). But once they’re in this terrain, they’ll surely find that the processes that make up the inner life are not amenable to the methodologies of social science. The moral and social yearnings of fully realized human beings are not reducible to universal laws and cannot be studied like physics.

Once this is accepted, economics would again become a subsection of history and moral philosophy. It will be a powerful language for analyzing certain sorts of activity. Economists will be able to describe how some people acted in some specific contexts. They will be able to draw out some suggestive lessons to keep in mind while thinking about other people and other contexts — just as historians, psychologists and novelists do.

At the end of Act V, economics will be realistic, but it will be an art, not a science.

Blow up their field??? Spoken like someone who doesn't understand the science or the math or the value of either.

Let me give you an example.

Several years ago, I found myself among a group of bioethicists, most of whom seemed frankly hostile to economists. One of the issues that was paramount in their discussions was the ethical allocation of scarce resources. While they were mostly concerned about what I think of as "bedside" ethics, they were grappling with a recurring issue on the nation's health agenda: the elimination of health disparities.

One recurring phrase was we need to "maximize societal health." They seemed to think this would involve giving more health producing inputs to health disadvantaged people. It seemed almost nonsensical to me for several reasons. First, if I wanted to "maximize societal health," I would allocate more resources to the people who were most efficient and effective at producing health, those whose marginal product from the resources is highest. Assuming that people who are more efficient at producing health are also more efficient at producing wealth (and vice versa), this would almost certainly mean that I would be giving more health producing resources to higher income people who are already healthier, not necessarily those whose health is least good.

That comes out of the math.

So whatever it is they want to do, it's probably going to be more complex than a simple transfer of resources. Knowing something about the math might have helped them think about this. And if they really meant what they said, they would probably end up denying resources to some who ethically should have them. I don't think "maximizing societal health" means what they think it means, at least in the short-run.

To be fair, they all understood the complexities of what they wanted to do, and most understood that one of the things that must happen is that those with low marginal product of health need to be helped to become more efficient and effective at producing health. They also understood that resources should be directed where marginal product was likely to be highest, regardless of ability to pay. But as long as there is heterogeneity in health production, "maximizing societal health" will allocate more resources to those who are best at producing health, probably the healthiest.

Another thing that comes out of the math is that if I take resources away from health advantaged people in order to give them to health disadvantaged people, the health advantaged people will be harmed, at least in terms of their level of health. This has ethical implications, unless of course, there is some region of the health production function where marginal product is negative for the most healthy. This could occur if the health advantaged have managed to capture more than the optimal allocation of health producing resources. Then taking some away from them and giving the resources to the health disadvantaged would make everyone healthier and would "maximize societal health." Win-win.

Unfortunately, the production of health is complex. Environment, nutrition, education,  and exercise matter as much if not more than medical care. So it probably matters which resources you re-allocate. It most likely won't be enough to re-allocate just medical care.

That comes out of the math and the theory.

It comes out of the theory because we as a society probably don't want to maximize societal health anymore than most of you reading this are "maximizing" your own health. Instead, you're maximizing something, known only to you, that provides you with what you judge to be a "good life." It is a combination of work, education, health care, nutrition, leisure, relationships, and, for some of you, tobacco, alcohol, and drugs. (The last three might actually harm your health.) We economists call what you're maximizing "utility." Utility can be thought of as happiness or satisfaction. We call it "utility" because the term implies something a bit more complicated than short-sighted pleasures. In its original sense, there was a sense of virtue or interconnectedness with others that would imply a brake on pure hedonism in it's shortsighted modern sense. Implicit in this conceptualization is the notion that you are the best judge of what gives you utility and that you have all the information you need to make that judgement.

What we might want to maximize then is the sum of all our utilities, which is one way to think of maximizing societal well-being. It turns out it's difficult to figure out how to do this. So instead we could assume that societal well-being is indirectly reflected by the value of the total amount of a nationally produced mixture of publicly provided goods and services and privately provided goods and services. If we were to adopt this proxy measure of societal well-being, we would want that "mixture" and the amounts to match pretty accurately people's preferences and everything that yields benefits or harms to them.

It turns out that if private markets are competitive (i.e., no externalities, no information asymmetries, no agency problems, many small firms, etc), they will provide the mixture and amounts that match people's preferences. If those preferences map accurately to individual utility (and rightly or wrongly, economists assume they do under certain conditions), societal "utility" or well-being will be maximized.

There's a problem though. For many of our most important privately produced goods and services, say, health care and financial services, the conditions for perfect competition are not even close to being met. So we can't be sure that the right (societal utility maximizing) amounts and mix of those services are being produced. This means that our proxy measure of societal well-being will not accurately correspond to actual well-being.

Other important private goods and services, like mother's milk and child or elder care by (unpaid) family caregivers, aren't traded in markets, which makes it very difficult to measure the valued-added by maternal and family caregiver time and effort (what they add over and above the final goods and services they purchase to provide the unpaid services). For this reason, their contribution to societal well-being will not be accurately reflected in our proxy measure of societal well-being.

And then there are public goods, like clean air and water, disease control, national defense, and public recreational facilities. In the absence of markets, it's difficult to know for sure what mix and amount of public goods will "maximize societal well-being." If our political process does not yield the socially optimal amount of final goods and services used to produce these public goods, it will tend to bias the correspondence between our proxy measure and actual societal well-being. This is one reason why it is so important to have a well-functioning political system and an informed electorate.

For these reasons, "perfect competition" in private markets does not solve all of society's allocation problems nor does GDP (a measure of national output) accurately reflect societal well-being. Most economists know all of this and some are attempting to develop better measures of national well-being.

The point of all this is to say that much of the science of economics is quite sound and quite useful for informing public policy. Recent events in the financial sector have stimulated much needed soul searching in macroeconomics, but that doesn't mean the whole discipline is wrong or corrupt. Nor does it mean that all macroeconomists are fools. Some of them got it right and continue to get it right. Even those who got it wrong are entitled to be wrong. Scientists make mistakes. When they learn they have made a mistake, they revise their theories. It's an important feature of "science."

If there is a flaw in the science and math of economics, I believe it is in the narrow ways we have tended to conceptualize and relate our theories to real-world phenomena. I think that Brooks is right that we have tended to downplay externalities of affect or commitment. We have also tended to be uncritical about the normative meanings of our positive assumptions. For example, we tend to assume that a household's budget constraint is endogenous in hours worked and that workers are paid their marginal revenue product and that therefore wages (and profits and household incomes) are always and everywhere just. And we have tended to assume separability of consumption bundles even among those whose incomes are so low that they are substituting heat for food and food for children's health care.

I don't know that these positive assumptions are necessarily fatal in accurately estimating average effects (causal or otherwise), but I frequently worry that public policy based on them may be fatal (literally fatal) to some groups (groups that are not average) that the policies purport to help. That has normative implications.

Another empirical and merely technical issue I worry about is our strong preference for instrumental variables methods when a policy variable of interest is self-chosen or otherwise correlated with the error term. These methods frequently yield purportedly unbiased, but often highly imprecise estimates. As the limitations of instrumental variables have become more apparent, instead of trying to "fill in" the unobservables, we have resorted to ever more complex "instruments" derived from so-called "natural experiments" that yield "instruments" that may or may not be exogenous and that produce results that may or may not be unbiased estimates of causal effects. Then we make public policy recommendations based on the results without regard for the possibility that human lives and well-being hang on the accuracy and precision of those estimates.

But these are not flaws of the underlying science, they are flaws of application and translation. They are the kind of flaws found in peer-reviewed research in other social and behavioral sciences. If we use this period of soul-searching well, we will begin to recognize, teach, and remedy these and other deficiencies of translation.

I think Brooks is wrong. We will not blow up the field. We will improve it while remaining a science. There are plenty of economists who have been simultaneously scientists and moral philosophers (and maybe even artists) who will lead the way. Some of them have even won Nobel prizes. They just haven't gotten much attention until recently. The market for their ideas has been "depressed."

Let's hope recent events in housing and finance (and soon health, I'll bet) will change that.

Comments

[reason]

To be snarky, it is neither it is a religion.

But to read the rest of your post I see some substantial biases. First, it is far too concerned with consumption and to little with production. You are not unique in this, it is a bias in economics since the earliest days. One of the most substantive criticisms of economics, is the lack of real (as against suppositional) productive function research. Demand curves are also suppositional, but we have actually much more evidence about them. But supply issues matter (not the least of those is the resouce cost - especially over the longer term).

Economics needs serious reform, and many well regarded economics are anything but scientific. Please, don't deny that, it puts you on the wrong side of reality.

But yes, Brooks is an idiot. Feel free to ignore him.

[The Raven]

A great deal of biology, for instance, is not mathematical. That does not make it unscientific. By "art," Brooks seems to mean the production of narratives--histories. The mathematics of economics is still far short of what is needed, but to abandon the idea that anything can be measured at all in the field is to abandon economics as a critical tool. Are we then to return to the economics of the 19th century, when the powerful chose the economic narratives based on their power, to excuse their abuses? (And why is economics only the concern of the wealthy?) The rich, cruel, and powerful, and sometimes also the poor, scared, and greedy (e.g. tea partiers) badly want economics to go away, because it unsettles their moral certitudes. People like Brooks don't want their privilege and prejudices questioned. And Brooks, yet again, is here a voice for the rich, cruel, and powerful. Has he no shame? But the answer, apparently, is no. It is very hard to shame the privileged.

[Peter T]

I can't speak for Brooks. I can see a place for the maths - it is, after all, a routine tool in ecology, history and other social disciplines.

BUT - to take your own argument, why assume "utility" is a property of individuals? And that it can be "summed". Why assume allocations reflect marginal revenue product? When we look around, we see issues at the level of individuals, small groups (families), large, structured groups (companies), and very large groups, and lots of behaviour that only makes sense if interpreted at the appropriate level. We see intricate, long-lasting structures that allocate rewards and shape individual's life-paths (and are shaped by them). We see brains that are literally formed by constant interaction with others, and so reflect multiple group pasts and priorities.

Incorporating these realities would make economics much less certain, in the sense of less calculable.

Your first point illustrates this. I am not sure who misunderstood "maximising societal health". Does it make sense to see health as an external property - which the healthy "produce" or as a state? In what sense can the very "healthy" persons "health" be weighed against the health of say 50 others? I suspect the doctors had a very different idea of health in mind.

[dogheart]

Your idea of economics isn't at all conducive to our continued survival as a species. It forces unnatural conditions on natural systems thereby completely distorting the true value of EVERYTHING. Money poisons absolutely. What it wants is not what we need.

[Bruce Wilder]

What I would be inclined to say, first, to the fatuous David Brooks, is that Economics is not, properly, a study of human nature, or even human society, but, rather, a study of a human, social problem: the economic problem of reconciling individual, material needs, wants and desires, with the organization of production to satisfy those material needs, wants and desires.

Because it is about a material, instrumental problem, it is perfectly appropriate for Economics to emphasize and, indeed, prescribe rationality in trying to solve that problem, or judge proposed solutions to that problem, by a rational, material, and even objectively measurable standard.

As to the math, I'm sorry to say that I have known many Economists, who knew much more of the math, than of the economics.

If you had honestly explained the Economic theory of production to the bio-ethicists, I suspect that they would have laughed at you. You would have been like the economist in the joke about the physicist, engineer and economist shipwrecked on a desert island with only a crate of canned goods and no way to open the cans. The physicist proposed building a fire, and heating the cans till they exploded; the engineer proposed rigging ropes and pulleys in a high palm tree, and lifting the cans to the top, and then dropping them from there to a rock below, in the hope they would break open upon impact. The economist considering their proposals, thoughtfully interjected: "no, no, you've got it all wrong. The way to think about this problem is this: First, we assume a can opener . . ."

That's the essence of the economic theory of production, the theory that output is a (mathematical) function of inputs. When you reason about re-allocating resources to improve health outcomes, that's the analysis you are using.

Of course, a minute or two of critical thought would lead to the insight that output is NOT a function of inputs. Economic theory gets past that little logical difficulty by assuming a can opener, that is by positing that "maximum" output is a function of inputs. By simply assuming that output is maximized, Economics sweeps aside as already solved all the problems of management, organization, engineering, technology. By doing so, Economics allows itself to ignore the technical and managerial challenges of organizing production, to focus itself exclusively on the problem of *allocative* efficiency. If those other sticky problems have been solved, and output is maximized, than the economic problem resolves itself to a problem of allocating inputs. Which is, of course, exactly the terms in which you addressed the problem of "maximizing societal health" -- you talked about re-allocating resources (inputs) among those already producing health, to change output. [As an aside, what Peter T said]

If you had clarified that your frame required assuming that all the technical and managerial problems of medicine and public health were solved, so that existing productive structures were "maximizing" output in a technical sense, in order to reduce the problem to one of allocative efficiency, I dare speculate that your bio-ethicists might have been less than impressed.

Contra David Brooks, I think Economists, in focusing on the economic problem, have good reasons to avoid the deeper muck of human nature. I am more skeptical of the neoclassical decision to parse the economic problem further, to arrive at only a consideration of allocative efficiency. You don't recover from the entailed disabilities, by later allowing that things are complex.

The truth is that business men and engineers and medical doctors do not spend much of their days considering how to "allocate resources" in isolation from the daily struggle to solve the managerial and technical problems of production. Their common sense idea of efficiency is mostly a matter of administrative or technical efficiency, not the allocative efficiency of the Economist. In the abstract, it is a matter of bringing a production process under control, in the cybernetic sense, and reducing waste and error from there. If they are "allocating resources" at all, it is likely to be whether to make a specific sunk cost investment in reducing (future unit) costs of output. The businessman isn't going to be worried about Pareto optimality in making such an investment (or allocation); he's going to be worried about finding a business model that allows him to recover that sunk cost investment -- something the alert Economist probably knows cannot be done "in perfect competition", and therefore judges should not be done.

Much of the real functioning of the economic world is hidden by this neoclassical parsing down to allocative efficiency. If output is already maximized, in a technical and managerial sense -- that's all settled business -- then marginal product is fully and uniquely determined. But, if we live in a world of genuine uncertainty, where technical and managerial problems are only partially and imperfectly solved, we might argue that each factor will still be paid its "marginal product", but that "marginal product" will not be uniquely determined. Rather than the marginal product determining the wage, the wage may well determine the marginal product -- and, we're in the somewhat topsy-turvy world of the efficiency wage!

If the neoclassical delusions of, say, Clark, that marginal product is just, get blown up, it will be none too soon, imho.

The fault is in the math, in the unsustainable (and under assumptions of genuine uncertainty, practically undefinable) assumption that technical efficiency is "maximized". What should have been a handy way to analytically distinguish allocative efficiency from managerial or technical efficiency was adopted as a crutch, and, then, became a handicap.

[tjfxh]

Robert H. Nelson, Economics as Religion (2001)

http://www.psupress.org/books/titles/0-271-02095-4.html

[Ron]

Brooks is not entirely wrong about the current state of economics and some of the causes, but he really misses the biggest points of all. One of the best ways of dealing with uncertainty is to pay careful attention to the certainties and just be mindful of the uncertainties and maybe even a little theoretical over-reach is fine, but never ignore the certain.

What on earth ever even made anyone think that a sustainable economic recovery could be built on a real estate asset bubble financed by loose credit. That was simply insane. Whe I took out an equity line of credit in 2004, I could have taken a fixed rate, but I chose a variable rate at prime plus zero, because I knew the bubble would burst and how the Federal Reserve would need to respond. I did not know about credit default swaps until 2006, but I saw the conventional loan rate increases lag over a year behind the Fed Rate increases that started in mid-year 2004 and knew something was hinky. That was just too much CDO/MBS originating outside of Fannie and Freddie, but it gave me clue as to why Fed Rates and prime rate did continue to rise on the back of a faux recovery through 2006. I figured out all of that just by watching basic information available to anyone and using a little common sense fundamental economic wisdom. I am only a high school graduate. Shouldn't a PhD economist be able to do at least that well?