relayed by Scott Alexander:
There’s an old philosophers’ saying – trust those who seek the truth, distrust those who say they’ve found it.
Can anyone source this? I see Andre Gide quoted as saying “trust those who seek the truth; doubt those who find it,” but I cannot find an essay or book to cite.
Take a typical consumption function where consumption depends on current income and other things. Income is endogenous. In CC models using 2SLS, first stage regressors might include variables like government spending and tax rates, possibly lagged one quarter. Also, lagged endogenous variables might be used like lagged investment. If the error term in the consumption equation is serially correlated, it is easy to get rid of the serial correlation by estimating the serial correlation coefficients along with the structural coefficients in the equation. So assume that the remaining error term is iid. This error term is correlated with current income, but not with the first stage regressors, so consistent estimates can be obtained. This would not work and the equation would not be identified if all the first stage regressors were also explanatory variables in the equation, which is the identification criticism. However, it seems unlikely that all these variables are in the equation. Given that income is in the equation, why would government spending or tax rates or lagged investment also be in? In the CC framework, there are many zero restrictions for each structural equation, and so identification is rarely a problem. Theory rules out many variables per equation.
Pointer from Mark Thoma.
I am afraid that Ray Fair leaves out the main reason that I dismiss macroeconometric models, namely the “specification search” problem. As you can gather from the quoted paragraph, there are many ways to specify a macroeconometric model. Fair and other practitioners of his methods will try dozens of specifications before settling on an equation. As Edward Leamer pointed out in his book on specification searches, this process destroys the scientific/statistical basis of the model.
I have much more to say on this issue, both in my Science of Hubris paper and in my Memoirs of a Would-be Macroeconomist. In the latter, I recount the course that I took with Fair when he was a visiting professor at MIT.
Other remarks:
1. On DSGE, I think that the main vice is the “representative agent” consolidation. It completely goes against the specialization and trade way of thinking. Fighting the whole “representative agent” modeling approach is a major point of the Book of Arnold, or at least it is supposed to be. (I may have been too terse in the macro section of my first draft.)
2. VAR models are just a stupid waste of time. As I said in a previous post, we do not have the luxury of saying that we construct models that correspond with reality. What models do is allow us to describe what a possible world would look like, given the assumptions that are built into it. VAR models do not build in assumptions in any interesting way. That is claimed to be a feature, but in fact it is a huge bug.
I think that the project of building a model of the entire economy is unworkable, because the economy as whole consists of patterns of specialization and trade that are too complex to be captures in a model. But if you forced me to choose between VAR, DSGE, and the old-fashioned stuff Fair does, I would actually use that. At least his model can be used to make interesting statements about the relationship of assumptions to predicted outcomes. But that is all it is good for, and for my money you are just as well off making up something on the back of an envelope.
Paul Romer says
Ultimately, the test of the model is its correspondence with the world. If we use certain frameworks, you can understand a much richer set of facts about the world. Growth is a difficult area to work because you’re addressing questions about the very long run, so you don’t have an abundance of data. You’re trying to invoke evidence from history over the very long run. We needed to come up with a way to think about all of these facts about the broad sweep of human history.
Math can be a very clear, concise, effective way to communicate ideas. What I saw in some of the people I was criticizing for mathiness was an almost obstinate adherence to positions, and then a use of any kind of mathematical argument that would support that position. What was missing was one of the characteristics of good science, which is to say “Well, given these new arguments, I may have been wrong before.”
Pointer from Mark Thoma.
Earlier, a reader’s comment brought me to Itzhak Gilboa’s review of a book by Mary Morgan.
Clearly, there are many instances in which economic analysis yields qualitative predictions, providing robust insights that allow us to predict trends, compare economic systems, and so forth. Yet, economics is not considered to be a successful science when quantitative predictions are concerned.
There is, however, another view of economics, by which it can have other successes: it is a field of enquiry whose goal is to critique reasoning about economic phenomena
This is an idea with chewing on. The purpose of a model, either theoretical or empirical, is not to provide a definitive “correspondence with the world,” as Romer would have it. Rather, it is to point out possibilities that deserve the attention of economists and those interested in economic policy.
I am reading Robert Murphy’s new book, Choice. Think of it as an English translation of Human Action. I am very happy with it so far. I can see how much of The Book of Arnold can be found in Mises, and yet. . .
I still do not buy into Mises on epistemology. Murphy writes,
Mises shows that economic laws are not obvious and that they do indeed enlarge our body of knowledge, even though economic laws do not need to be verified with empirical observation.
The Anglosphere is largely empiricist, or logical positivist. We tend to believe that there are two types of truths. There are tautologies, which are embedded in language; and there are truths about the world, which are learned by observation.
A claim that 2 + 2 = 5 can be falsified using logic. A claim that pigs know how to fly can be falsified using observation.
Milton Friedman takes an empiricist view of economics. For Friedman, the economist makes predictions about the world, and those predictions are verified or falsified on the basis of observation.
Empiricists give no epistemological status to anything that appears to be a claim about the world that cannot be falsified using observation. Such claims are classified as dogma or nonsense.
Mises was no empiricist. Murphy writes,
From the starting point that humans act, the economist could logically deduce–thereby forming a tautology, it’s true–that individuals have subjective preferences with ordinal rankings, that choices come with opportunity costs, and that the value of second-order capital goods is dependent on the value of the first-order consumer goods that the individual believes they have the technological power to produce.
The way I read Murphy/Mises, economic laws are derived from insight into human nature. At least some insight into human nature comes not from observation but from introspection. The insight that comes from introspection is not falsifiable. For example, suppose I find it inconceivable that I would make choices on some basis other than benefits and costs at the margin. This makes it inconceivable to me that other people would make choices on some other basis. Hence, I appear to arrive at a statement about the world–people make choices on the basis of benefits and costs at the margin–that is not falsifiable by observation.
My take on this is that a statement such as “people make choices on the basis of benefits and costs at the margin” falls into a category that I might term “guiding dogma.” We will use a guiding dogma to make predictions about the world. However, the guiding dogma is not testable. If our predictions go awry, we will not discard the guiding dogma. Instead, we will look for something else that made our prediction go wrong.
“Guiding dogma” may be synonymous with Kuhn’s notion of “paradigm.” In physics, there are some spectacular cases in which a guiding dogma came to be replaced by a new guiding dogma.
The interesting predictions are those which go beyond a guiding dogma. For example, a prediction that a rise in the minimum wage will reduce employment is based in part on the guiding dogma of the Law of Demand. However, the prediction about the effect of the minimum wage is falsifiable empirically. Suppose that a rise in the minimum wage does not produce a decline in employment. Will we throw out the Law of Demand, or will we look for some other factor at work? My claim is that we will do the latter.
Speaking of the minimum wage, consider this sarcastic assault on Larry Summers by John Cochrane:
Never mind centuries of supply and demand, centuries of experience with minimum wages and other price controls, or even the current controversies. Never mind that who works for what business and how many do so is a little bit endogenous. Larry has a new and very clever theory about monopsonistic wage setting in the presence of recruitment and motivation costs. (One that apparently only holds at the lower end of the wage scale where minimum wages bite?)
Thus, if we were to find that an increase in the minimum wage does not reduce employment, then we would credit something like “a new and very clever theory about monopsonistic wage setting in the presence of recruitment and motivation costs” rather than reject “centuries of supply and demand.”
Incidentally, the laws of probability are also not easy to fit into the empiricist framework. When we say that the probability of a coin landing on heads is 1/2, that sounds like a statement about the world, but it also might be thought of as the definition of a fair coin. Once again, the phrase “guiding dogma” comes to mind.
To many environmentalists, sustainability means leaving the world the way you found it. I think that this may reflect the instincts of a hunter-gatherer.
If you are a hunter-gatherer, how much you can eat is limited by the natural rate of replenishment. If you eat game or plants faster than they are replenished, your tribe will die.
Modern human welfare is not governed by replenishment. We use knowledge to add value to our environment. Cultivation of crops means that we can grow more food than we could obtain by gathering. And we apply ever-increasing ingenuity to this cultivation.
Sustainability of modern life is thus much more complex than sustainability of hunter-gathering. Our modern ancestors have left us the gifts of their ingenuity, so that what they took out of nature has not hurt our welfare. And we are likely to do the same for our descendants.
During my time at MIT, Robert Solow was harshly critical of the new classical macro models pioneered by Robert Lucas, dismissive in a way that seemed to me to skirt uncomfortably close contempt. I recall hearing the same type of criticism from Frank Hahn, who must have been visiting MIT. Looking back, perhaps I misinterpreted them because I was not familiar with the sarcasm and put-downs that were a part of British intellectual life that Solow had to confront in his exchanges with Joan Robinson. But if it sounded like contempt to me, others may have heard it the same way.
…The alternative to derision would have been for skeptics to embrace and extend. This was what Stan Fischer and Rudi Dornbusch, who were supervising almost all of the Ph.D. students at MIT doing anything related to macro, were quietly doing at this time. Fischer, Dornbusch, and their students absorbed the rebel critique of traditional macro, saw what something was missing in the first generation of rebel models, and set about extending them. As a result, Fischer and Dornbusch trained a cohort of Ph.D. students at MIT who put the tools of modern macro to work and as Krugman has observed, turned out to be unusually influential. If Dornbusch and Fischer had set the tone for the response to Lucas and his followers, things might have turned out differently. But because of the inherent instability of acrimony, grievance, and factionalism, they and their students could not undo the effect of the more hostile response.
Pointer from Mark Thoma.
I disagree with this so much that I can actually feel my anger.
1. Romer’s point is that Solow set a bad tone for macro, and all difficulties in the subject flowed from that. I call baloney sandwich. Solow did not set the tone for discussions in macroeconomics in this period. As Romer points out, by this point Fischer and Dornbusch dominated macro at MIT at this point.
2. Solow’s problem with Lucas was that Solow thought that reality should take precedence over microfoundations. Solow equated Lucas’ approach to macro with deciding that because one’s theory could not explain how a giraffe could pump adequate blood to its head that one had proven that giraffes do not have long necks.
3. I think that one problem in macro is that there are many theories that are consistent with observed reality. Freshwater macro happens to be one of the few that does not reconcile with reality. It deserves Solow’s disdain.
4. Romer thinks of Dornbusch and Fischer as heroes. To me, they are villains. They pushed the representative-agent, rational-expectations nonsense that is good for nothing but mathematical, er, self-abuse.
5. Even though Solow is a Keynesian partisan and I am not, I still feel connected to him because we share a view of what is wrong with the way macro has been pursued since the 1970s.
Someone recently emailed me that I should put my memoirs of a would-be macroeconomist on Amazon. For now, it’s freely available, and I think that Romer and others interested in his posts should read it. In fact, if you want to read it on Kindle, I am pretty sure that this file will work.
Noah Smith updates us on the Paul Romer’s “mathiness” critique. He translates (without necessarily agreeing with) Romer as saying
If I could tell the freshwater economists just one thing, it would be that the rest of economics is doing things differently. Really. We’re out here being honest with each other, trying to get to the truth together, not politicking for our own pet theories. We’re being scientists. You can too. If you get outside your bubble, you’ll see I’m telling the truth.
Pointer from Mark Thoma. My comments:
1. Just as a point of clarification, in the debate between freshwater and saltwater economists, I am certainly not an advocate for freshwater economists. I am definitely in the “plague on both your houses” camp.
2. For this post, I would like to stipulate that saltwater economists are as detached and objective as they claim to be. Put aside my doubts about that for now.
3. What math can do is rigorously connect assumptions with conclusions. You manipulate the equations to show that the conclusions follow from the assumptions.
4. As Noah has pointed out on other occasions, economics differs from physics in that physicists usually can undertake direct tests of their assumptions, while economists generally cannot.
What point (4) means is that when we prove that assumptions a, b, and c together imply outcome X, and we instead observe outcome Y, we have no way of independently testing which of assumptions a, b, and c is untrue in the real world. Because there are so many plausible assumptions available to economists, this means that real world does not constrain economic models nearly as much as it does in physics.
Assumptions persist in economics as they get copied from paper to paper. That is, because of a combination of convenience and path dependence, not because of empirical verification.
I think that this makes the claims of “science” in economics quite dubious, and in macroeconomics downright fraudulent. Even if you think that there is a group of economists who is unbiased and objective, they are not entitled to don white lab coats.
But r = 0.23 means the percent of variance explained is 0.23^2 = ~5%. If some Social Darwinist organization were to announce that they had evidence that who your parents were only determined 5% of the variance in wealth, it would sound like such overblown strong evidence for their position that everyone would assume they were making it up.
His point is that some pundits have used a recent study to claim that inherited wealth is really, really important, even though numerically the study fails to show that.
Suppose we define mathiness as people making misleading, ideologically loaded claims about what their theorems prove. It seems fair to suggest that statisticiness is a similar problem.
You may recall my essay on MIT economics. Now we have Paul Krugman in response.
Robert Shiller makes some of the same points that John Cochrane was making at lunch about the way that housing markets can get out of line.
Pointers from Mark Thoma.