The Macro Wars: Inside-out vs. Outside-in

I remember reading once that it is still not understood how the giraffe manages to pump an adequate blood supply all the way up to its head; but it is hard to imagine that anyone would therefore conclude that giraffes do not have long necks. At least not anyone who had ever been to a zoo.

Robert M. Solow

Solow wrote those words at the height of the macro wars. I was very much on his side at the time, and this post will explain the sense in which I am still on his side.

Think of the task of macroeconomics as completing a mineshaft between the “outside” (what we observe in the world) and the “inside” (a mathematical model that is “pure” in its microfoundations). The Old Keynesians, including Solow, took an outside-in approach: let’s work from what we observe, build a crude model to handle that, and maybe eventually we can dig deeper and find the microfoundations. Start from the fact that there is a giraffe, and try to figure out how it maintains its blood supply. Do not start from a model of blood supply that precludes the existence of giraffes.

For the Old Keynesians, macroeconometric models were a tool with which to observe the world. They provided the starting point for the outside-in approach. Then Robert Lucas came along with his “critique,” which said that if you took an inside-out approach that included rational expectations, macroeconometric models would break down. The Lucas Critique launched the macro wars.

Lo and behold, macroeconometric models did break down. However, I do not think that the Lucas Critique had much to do with it. You can get more on my perspective by reading this paper and by reading my macro book.

The New Keynesians took up Lucas’ challenge by adopting an inside-out approach. Stan Fischer’s course at MIT was 100 percent inside-out theory, and I viscerally hated it. At the start of one class, I stood up, proclaiming loudly and sarcastically to Fischer and my fellow students how much I enjoyed the topic of “monetary growth models,” which was the particularly pointless mathematical, er, self-abuse that he was teaching us that week.

I chose Solow as my dissertation adviser, and I wrote an outside-in thesis, working backwards from what we observe to a theory of price rigidity. Not having a thesis that focused on rational expectations and not having Fischer plugging for me were career-altering. I was doomed to failure if I tried academia, and so I wound up on a different track. I don’t think I was the one who lost out on that deal.

So if you are trying to follow the methodological discussions among Mark Thoma, Paul Krugman, Noah Smith, and others, you will find me still on the side of the Old Keynesians. I still despise inside-out macro, and I still prefer the outside-in approach.

What has happened to me since I left MIT is that I no longer think that macroeconometric models provide a valid lens into observing the real world, and I no longer think that Keynesianism is the One True Way. The real world is still out there, and I still think it should be our starting point for digging the mineshaft. I still respect the Old Keynesian approach of starting with observations about the world rather than starting at the bottom of the mine with a “pure” model. However, I am willing to entertain theories that differ considerably from the Old Keynesian one. Hence, PSST, which you can also read more about in my essays/papers.

33 thoughts on “The Macro Wars: Inside-out vs. Outside-in

  1. Some large fraction of econ is closed off if one doesn’t follow a single ideological position? Really? That doesn’t seem to jibe with the caplans and blocks I’ve read claiming its a good risk. I guess one must dance to the piper’s tune elsewhere but it is not claimed otherwise.

    • Actually, I just changed my mind, in industry for example, noone expects you to pretend you really believe in what you are ordered to do.

  2. Kling, with a Ph.D. from MIT, could (should?) have found a tenure track position at a Ph.D. granting university. Maybe, a top 10 school. Princeton, Yale, Harvard, etc. Without a thesis based on rational expectations and Fischer plugging for him, that didn’t work out.

    I don’t think Block or Caplan are really telling you that going for a Ph.D. in economics will likely result in a tenured position at Harvard. There is a good chance for a tenured position at a more teaching-oriented college or university.

    Going for a Ph.D. in History or Philosophy makes even that low probability.

  3. Is there a word for trying to be too precise in either a model or simply in the analysis of something. I know there are words like overfitting, overspecified, too many degrees of freedom, too many significant digits, etc but I’m looking for one word or phrase that encompasses all of it.

  4. I’m with Arnold on this one. Went to Stanford expecting to study macro. Took the courses and seminars, then wrote a micro dissertation. Macro is now a hobby. Search my blog for “healthcare” and you’ll find two articles I posted 3 years ago that correctly predicted some of the outcomes ACA is foisting on us today. I did not, however, manage to predict the vast quantities of incompetence and stupidity. For that I am truly sorry. :)

  5. I’m really sympathetic to your point of view. But to use your own words, this type of blog post is “nihilistic and unhelpful”. I hope that in the future you’ll spend more time explaining why your alternative to evaluating macroeconomic policy is better than the approach you despise.

    Economics is all about evaluating decisions against the alternative. Let’s hear more about the alternative to which you’re comparing modern macro!

  6. As an individual trained in engineering, computer science, and accounting, I’ll provide you with a slightly different perspective. While the experience is not exactly in the world of macroeconomics, you may find it analogous to this situation.

    I should mention that I like to tell my sons that in my younger days I was enamored with the mathematical models of operations research and management science… until I discovered that people are rarely rational decision makers.

    When designing and implementing advanced management systems for manufacturing companies, we frequently talk about ‘top-down’ design and ‘bottom-up’ implementation. That is, you start with a set of goals with top management and work out a system design to achieve those goals. Nice theory, but it does noy really address the dynamic nature of knowledge and technology. Namely, it’s not practical to design a system if you’re not aware of what is realistic or the capabilities of new technology. Hence, the need for reviewing bottom-up insights.

    As a result, many years ago I focused on defining our methodology for iterative development that involves cycles of top-down design checked against bottom-up reality. In fact, we use the expression ‘outside-in’ to describe this methodology… So without sounding like the great ‘compromiser’, perhaps a macroeconomic model that uses one side to check the other, is a practical approach.

  7. Macroeconometrics is an attempt to explain away some inconvenient but easily observable phenomenon.

    “All Keynesians devolve into liars. I give you Krugman.”

  8. Don’t believe me? Check out the pedigree of those that think ObamaCare was workable economically. Not all economists can do basic arithmetic.

  9. Paul Krugman(D) isn’t a Keynsian… he’s a Democrat. To them, Keynes is an excuse… and you can’t “prove” them wrong because they don’t understand Keynes in the first place.

  10. I did graduate work in mathematical economics in the ’70s and was far more impressed with the work being done in micro than in macro. Keynesian and post-Keynesian macro just never made much sense to me. Not nearly as much sense as the non-mathematical work of people like Hayek and von Mises. It seems to me that what we call macro should probably be seen more as a matter of political economy – since it involves the ordering of the economy by political means, whether that ordering is ‘hands off’ or dirigiste – than what has passed for macro for the past 75 years.

    I did enjoy the really theoretical work in general equilibrium theory, but that was probably because it was elegant and I was learning some really neat mathemtics like topology and measure theory while thinking about all of economists’ assumptions that drive non-economists nuts, such as convexity, continuity, and twice-differentiability – which patently don’t obtain in real life -and how to relax the assumptions.

    I also remember the fad for “input-output” economics which was supposed to recast the socialism debate by making it possible, with big enough computers, to make central planning work. I even still have a copy of Leontieff’s book somewhere.

    Sigh. I still enjoy reading economic thought (if not the overly mathematical stuff), but I am very glad I don’t do it for a living….

  11. Did you ever look into the Austrian School (Menger/Mises/Hayek) approach to the artificial divisions of micro and macro economics? They pretty much evicerate both sides of this argument

    It is a shame that so many bright minds spend some much time and treasure studying and writing about “macro economics”, a field that does not surpass astrology in its scientific rigour.

  12. Not an economist, so feel free, perhaps even obliged, to dismiss my comments. My exposure to economics was in one ECON 101 course at Penn State, many, many moons ago, under a Dr. Thomas Fox, who was (if memory serves) an economic advisor to President Johnson and a devoted Keynesian. That, and many hours of reading popular books and articles on the subject by Milton Friedman, Thomas Sowell, George Gilder, Henry Hazlitt, Alan Reynolds, and other libertarian/conservatives. Friedman got pretty short shrift in Fox’s class, as I have given Keynes pretty short shrift in my recreational reading.

    Why? I’ll just state it flat out: the Keynesian view never made an logical sense to me. Maybe my complaint isn’t with Keynes himself, but with what his disciples made of him, I dunno. The model, at least as Dr. Fox presented it, was mechanistic to the max. He even provided a computer “simulation”, which “showed” that the more the government spends, the bigger the economy.

    Well, hell… why not go all in, then? Let government spend 100% of the GNP.

    Here’s why not: nobody really believes that model, which would suggest that North Korea ought to be the richest country in the world.

    I think economists have deceived many people, if not quite themselves, that economics is scientific, mechanistic.

    No. Newtonian physics is. Chemistry is. But not economics. Economics is the science of trying to read the tea leaves in a giant cup of steaming-hot decision-making as it unfolds. Atoms and molecules are predictable because they do pretty the same things every time, just adjust the parameters a bit and they’ll do that instead of this. Thow a beaker of sulfuric acid or caustic lye at something organic and it burns. It will happen every time.

    But people and the decisions made by people are the atoms and molecules of the economy. Throw a “stimulus package” at a reeling economy and you’ll get something different than when everything is running well. Economies tend to run well when people have faith in the general predictability and impartiality of the law, and confidence in the future. When the government tinkers with stuff, making economic decisions for political reasons, people’s behavior changes because they lose trust, and thus confidence and any sense of personal control of their economic destiny. It breaks these simplistic models. These atoms and molecules now do *this* when they used to do *that*.

    If that’s how chemistry and physics behaved, a physicist or chemist would tell you, all bets are now off. That’s the “sinking sand” model I learned in Sunday School, many years ago. “On Christ the solid rock I stand, all other ground is sinking sand.” Science requires a bedrock of unshakeable faith in empirical methods, if not precisely in Jesus. When a chemist drops sodium in a tank of water and it explodes, chemists do not stand around and say, “Hmmm, that’s interesting.” But the government is the proverbial bull in the economic china shop, and it is its conceit that people will behave they way it wants them to. They expect an explosion and all they get is a cold hissing sound. People are the elements and molecules of the economy, and they can form opinions about what the chemists and physicists are up to, and respond accordingly.

    I posted the following on my blog in March 2009, when the Obama Presidency was just starting to grind the gears that Bush broke…

    > “…the game has changed so much that you will no longer need to wonder, when ten different economists are trotted out to articulate ten different opinions, which one happens to be right. Now you will know none of them are right. They were all trained in economics as it used to be, the old game, according to the old rules. Those rules are dead. Nobody knows how things will work, or if they’ll work, from here on out. Nobody’s forecasts are worth a dime. Nobody’s advice is worth soliciting. New rules, new game.”

    http://reformedtrombonist.blogspot.com/search?q=blount

    Turns out I was right. Every month since, whenever bad economics news comes out, the word “unexpectedly” appears in the text somewhere. Economists are going “Hmmm, that’s interesting” and scratching their beards.

    You can’t change the rules in physics; that’s why physics is a science — whereas, economics is a black art. It relies on observing behavior and trying to predict its results. You can affect behavior, but you can’t necessarily predict how things will turn out. I’m a computer wonk and when I change a single line of code on something I wrote myself, I can’t necessarily predict how things will change — that’s why we test before we deliver the updates. Economists don’t have a “test” economy. All they can do is test their theories in the real world, on us, and when they break stuff, all they can say is, “Ooops,” presuming they’re honest enough, and scratch their beards.

    And things tend to break when you tweak something as complicated as the economy.

    Sorry for the rambling on. My two cents, anyway.

    • You have noted something I have long notices, also. Economics is not really a science like the hard sciences are. Hard science lives in a world of permanent, inflexible rules. We have no say over those rules. In economics the rules are often what we want them to be. Tax rates, fads, what information is chosen to be presented and how it is presented………You get the idea. So many variables in economics are merely random numbers with no basis in reality. Add to this the complexity of what you are studying and you see the problem at hand. Years ago, chaos theory was supposed to be the salvation of economics. Guess what, chaos theory only applies to real world phenomena that have physical basis. The rules at all levels must be fixed. The rules will never be fixed in economics.

      • I think the closest we can come to “fixing the rules” in economics is to make them as simple and predictable as possible, and to reduce government meddling as much as possible. It’s not just the amount of government meddling that’s the problem, but the philosophy behind it. Economics is a competition, like it or not. We have scarce goods, everyone wants all of them, we settle for what we can get. All competitions need rules and referees. No, you may not pull a gun on that clerk for his Sony Playstation 3. No, you may not pollute the river with your effluent wastes. No, you may not enslave your employees. Rules to call, and someone to call them. That’s the proper role of government. I don’t think it even matters so much what the rules are, so long as they’re a) predictable, b) universally applicable, and c) gives the players a reason to bother with playing, which we call ‘incentives’.

        Today’s government, however, doesn’t referee anything. It takes sides. You win the game by paying off the referees. Once upon a time, you got rich by manufacturing cars people wanted; today, you get rich by failing to manufacture cars people want. And all because the ref has an idea who he wants to win.

        If that’s how football worked, would anyone play? Would anyone watch?

  13. One problem with Keynesianism was that it was based on observations that did not hold up over time, like slow information flow and myopic decision-making. The structure of the economy itself changes over time. The idea behind the Lucas critique was that we should understand the behavior that causes the structural change or we will never be able to anticipate it. We will always relying on historical relationships that are no longer valid. The newer approach has turned out to be too theoretical to be practical, so what now?

  14. From Ralph Raico, Was Keynes a Liberal?:

    “Throughout Keynes’s career, however, clear indications appear of his longing for
    a much more radical social order—in his words, a “New Jerusalem” (O’Donnell
    1989, 294, 378 n. 27). He confessed that he had played in his mind “with the
    possibilities of greater social changes than come within the present philosophies” even of thinkers such as Sidney Webb. “The republic of my imagination lies on the extreme left of celestial space,” he mused (1972, 309). Numerous statements strewn over decades shed light on this somewhat obscure avowal. Taken together, they confirm Joseph Salerno’s (1992) [1] argument that Keynes was a millennialist—a thinker who viewed social evolution as pursuing a preordained course to what he conceived to be a happy ending: a utopia (O’Donnell 1989, 288–94). [2]”

    [1] Joseph Salerno, 1992. The Development of Keynes’s Economics: From Marshall to Millennialism. Review of Austrian Economics 6, no. 1: 3–64, link to PDF: http://mises.org/journals/rae/pdf/RAE6_1_1.pdf

    [2] Ralph Raico, Was Keynes a Liberal? The Independent Review, v. 13, n. 2, Fall 2008, ISSN 1086–1653, Copyright © 2008, pp. 165–188. link to PDF: http://www.independent.org/pdf/tir/tir_13_02_1_raico.pdf

  15. My view is that GDP is the wrong number entirely.

    Everything depends on disagreement. If you value something I have at more than I do, and vice versa I value your thing more than mine, we trade and we both come out ahead.

    The standard of living of the nation rises by the amount of our disagreement, not by the market-clearing price of the goods traded.

    Add up all the disagreements of the nation, and the standard of living of the nation rises by that amount.

    The rule for maximizing disagreement, and hence the standard of living of the nation, is specialize and trade. The specialist values his output at much less than his customers do.

    So long as there is disagreement that’s not under water, that job exists.

    Tax and regulate and disagreements go underwater. Jobs disappear.

    Macro constraint appears: things that lower the standard of the nation, that is things that require underwater transactions to happen, don’t work.

    That includes digging holes and filling them.

    Don’t use the GDP for this problem.

  16. Its not that your ‘outside-in’ method is fundamentally flawed, but like any approach to data interpretation there are powerful caveats that should be considered. Such a macro approach is very likely to be more subject than the alternative to things like confirmation bias, arguments from authority, and group think. IE- you already know what a giraffe is, so nobody is asking many questions about giraffes, and if an established giraffe expert wrote something about them 40 years ago that everyone believes, its very unlikely you will question that. Everybody knows that swans are always white.
    Now the ‘inside-out’ approach is also subject to this, but less so.
    When you are breaking new ground, you really need to examine all the new parts and how they fit together. There is less established theory to wrangle with (or not). Moreover, just the example of giraffes should be your best friend. If your theory doesnt conform to reality, that is a good thing, your theory is wrong. IE- this is all about falsifiability. A theory created from the ground up will be tested all along the way, if its consistent with reality it will be robust. A theory built from the outside in is by nature less robust. You know what you know within certain parameters, but since you dont have much idea of WHY, you open yourself to rather nasty surprised outside of your well trodden ground. You are very susceptible to thinking you know more than you actually do, and particularly in economics that leads to making bad assumptions.

  17. “Outside-in” like other approaches is an attempt to make order out of what appears to be chaos, but really which is just a very complex and dynamical system having actors who have varying and inconsistent degrees of rationality. It is further complicated by the unconscious biases of the people trying to figure it out. The system is so immense and the apparent chaos is so massive that several theories that propose models of the economy sound plausible, but are not founded on honest rigor. Each in its turn shows its defects even as those who propose them rush to patch them up.

    For me, Keynesian macro was an excuse to justify socialism. It only made sense when that end point was kept in mind. How many trillions of our wealth and credit have been squandered and how many future generations have been made serfs by the simple formula of C + I + G? This formula used by champions of big government to tell us with a straight face that we would be better off if they borrowed money and spent it on digging holes with one crew and filling them up with another.

    If some foreign power were attempting to impose Keynesian economics on an otherwise free and prosperous people, we should consider it an act of war.

  18. Thank you Reformed Trombonist for writing what I have always thought about economic theory. I am a student of human behavior and how the brain works. I would assume that micro-economic deals with human behavior, both rational and irrational?

    • Thanks for your kind words. My basic economic assumption is that most of the people, most of the time, do what they consider to be in their personal economic self-interest. Humans are cooperative but they’re also competitive, so often what’s good for Jim, in Jim’s estimation, isn’t good for Jack. That makes it a bit of a free-for-all. Many humans seek predictability, and the hurly-burly nature of the free market doesn’t reassure them. This leads those people to prefer government involvement, because that makes it easier for them to keep their eye on the ball, so to speak. So a different class of people strikes it rich. The age of push is over and we enter the age of pull.

      But people are not simple atoms and molecules, and at some point we realize we’re being played. And some of us change our game, and others just quit playing.

  19. As an undergrad Econ major in the mid-seventies I was smart enough to build macro-econometric models ( Fortran/WatFive on IBM 370 ), but not enough to pull rabbits out of chaos, so I moved on. Two failed and two successful businesses later, I find power and truth sufficient to steer a large community of rational/emotional human actors through the Keynesian meltdown within the insights of Von Mises, Rothbard et al. For this defunct economist, “outside-in” works well only at the micro level, where the “model” embraces chaos, disagreement and un-modeled unknowns; the sphere which, come to think of it, the Founders sought to encourage, not macro-diminish. My 0.23 cents ( .02 adjusted for inflation ).

  20. I’ve always thought Keynesian theories (old, new, etc) were like an Ice Cream and Pizza weight loss diet. Dangerous because it gives people justification for giving into their temptations, even when the results are going to be negative. Politicians are strongly tempted to appropriate more and more control over the economy, because it gives them more power and importance. It also gives the less scrupulous (and we should realize that sort of person will be overrepresented in politics) a chance to skim more money.

    So Keynes will tend to find favor with the political classes, because it justifies giving them more power and wealth. Has any Keynesian model ever tried to account for that?