Recent Posts

Monday, January 8, 2018

The Death of Mainstream Macro

The American Economic Association's big conference - ASSA - was the occasion for a limited amount of introspection about the state of economics. On the surface, it looks like mainstream economists congratulating themselves on what a great job they are doing. However, my interpretation is that my views were confirmed: mainstream macro is essentially a dead body of theory. Anyone hoping to advance our state of knowledge of the macro-economy needs to look elsewhere.

Although I wish to avoid spending too much time on economic squabbling, the discussion here might be useful if I ever get around to writing an obituary for mainstream macroeconomics.

Sure, Economics Is Just Not Macro

One standard complaint of mainstream economists is that critics are all wrong about what economists do: they do not all do macro forecasts. That is probably true. However, that is an utterly irrelevant point for most people: they do not really care about the non-macro parts of economics. To be more precise, the average person cares about non-macro economics as much as they do about any other field of the humanities: not a whole lot. This is perhaps unjust, but I think anyone in any other branch of the humanities could have explained this to economists a long time ago.

If I take my personal example, I had contact with hundreds of economists over the course of my career. Furthermore, I probably read articles or saw television interviews of hundreds more. Every single one of those economists' day jobs was macro analysis. I am not some weird outlier; anyone in finance or who watches financial television, or reads the financial press, would have pretty close to the same outcome. Probably the only people who might experience otherwise are academics in economics departments.

(These mainstream academics make a big deal about whether economists produce forecasts or not. Technically, not all macro analysis is forecasting; in fact, I do not do forecasts. However, if you are a mainstream economist and argue that economic outcomes are the result of optimising behaviour along with accepted deviations from optimality, macro analysis is tied to some entity doing macro forecasts.)

So if any mainstream economist reads this, yes this is not a critique of your non-macro academic work.

Paul Krugman: Mainstream Macro Is Dead

Paul Krugman published an article: "Good enough for government work? Macroeconomics since the crisis" (the paper is currently available for free, but I think that will only be for a month).

I worked as an academic in a real academic field. One of the requirements of being minimally competent was being able to read another researcher's paper, and compare what they demonstrated within the body of the paper versus what the abstract says the paper accomplished. One of the distinguishing characteristics of the DSGE macro literature is the gulf between the claims in the abstract and the actual analysis done. If you take article abstracts at face value, every single problem in macroeconomics has been solved at least five times. The fact that all of these "solutions" do not agree with each other tells us all we need to know.

Therefore, it is a mistake to look at Paul Krugman's conclusions about the state of macro, rather, one needs to look at the details of his argumentation.

In summary, he argues that we could just use IS/LM (with a correction for the effect of the zero bound, which he coincidentally has a paper about) to analyse the policy response to the Financial Crisis. Basically, we have two dials: "monetary policy" and "fiscal policy", and we move them to manage aggregate demand. In other words, we can use the "state of the art" 1960s undergraduate textbooks (which he used when he was a student, plus of course, Krugman's zero bound paper) to solve all of our macro problems.

There is a grain of truth to this. Faced with an economic collapse, ramping up government spending will help speed up a recovery. For that particular situation, that's perhaps all you need to know. However, everyone other than ideological extremists knew that already. That said, knowing that an increase in government spending will most likely raise nominal GDP does not answer most of the important questions. How large is the effect? What are the costs? Stop-go aggregate demand management has a deservedly bad reputation. The insight has extremely limited value in other contexts, such as judging the effect of a universal basic income, or Job Guarantee.

It also runs into a rather obvious problem: everybody working in the macro side of finance has access to those textbooks (and may have even studied from them in university). However, the IS/LM model (or whatever variant Krugman is discussing) rarely shows up in discussion -- because it is largely hopeless. It consists of two arbitrary lines on a chalkboard, that can jump around for any number of imaginative reasons. You can use it to predict almost any possible outcome -- which is not a useful property of an applied model.

Even if one insists on sticking closer to Krugman's interpretation of his analysis, the existing body of mainstream macro looks to be in a very bad state. He notes how mainstream analytical tools failed to offer useful clues for the direction of inflation after the crisis (the large negative output gap implied deflation). The entire theoretical core of DSGE macro is based on price determination; the theory of inflation it provides is normally seen as the main theoretical advantage versus heterodox approaches. Admitting that we have no idea how to forecast inflation -- when the main triumph of mainstream economics was to put in place central banks whose entire job description was to stabilise inflation -- is a serious admission of failure.

Look at Us, We're Doing All This Amazing Stuff!

The other line of defense of mainstream macro is that young researchers are doing all this amazing new work. However, given the track record of the people involved, why would we expect any better outcome with the latest iteration of mainstream macro?

One argument is that mainstream macro is more empirical. Yes, so was the Reinhart and Rogoff government debt analysis. Their empirical analysis was useless even in the absence of the spreadsheet error since they commingled data from free-floating sovereigns with those in currency peg regimes. You cannot do useful empirical work in a complete absence of a theoretical framework. Analogies to physics do not cut it: we are working with complex systems, and you cannot infer the behaviour of complex systems from stand alone analysis of individual components.

The other line of defense is that researchers are working on relaxing the various obviously wrong assumptions embedded in representative agent DSGE macro. As is obvious, I am not following that literature, but from what I have seen, the usual tendency is to relax one assumption at a time, and leave the rest of framework intact. That's known as "bolting on epicycles," and one does not need a doctorate in the history and philosophy of science to guess what the outcome will be.

The other thing to keep in mind that to the extent the new literature works, it is jettisoning existing mainstream macro theory. They are starting off from scratch, and refusing to look at the literature that diagnosed the problems with mainstream macro decades ago. How likely is it that they will develop any conclusions which were not observed much earlier by post-Keynesian economists?

Concluding Remarks

As should be clear, I pay almost no attention to the latest developments in mainstream macro. Perhaps some discoveries will surprise me, and end up being useful. If the reader wishes, they are free to read the literature, and draw their own conclusions.

Nevertheless, I remain cautious. The mainstream has a long history of announcing that problems are solved, or great progress has been made ("the state of macro is good!"). Even by the standards of mainstream economics backers, the new research areas show little sign of producing insights that were not already obvious to someone with a slight familiarity with post-Keynesian economics. (For example telling us that people's behaviour is not entirely rational would not surprise anyone paying attention while reading Keynes' General Theory.) Unless you are being paid to keep up with the latest research fads, it is probably safe to wait until some form of new consensus appears among researchers before actually reading the papers. (If you see anything that sounds plausible, feel free to send me questions about it.)

(c) Brian Romanchuk 2018

18 comments:

  1. In the realm of Keynesian debates have you read this critique of MMT by Thomas Pailey? He argues with some conviction that MMT is an inferior version of neo-Keynesian ISLM models.

    http://www.thomaspalley.com/docs/articles/macro_theory/mmt.pdf

    Footnote 11 on page 12 concludes with this statement: "Indeed, Post Keynesian economics is now unwittingly re-inventing the stock – flow consistent ISLM model as evidenced by the much cited work of Godley and Lavoie (2007). The principal innovation in their framework is the extension of the stock-flow consistent ISLM model to include endogenous money, inside debt, and inside debt effects on AD."

    The link below is his reply to a response by MMT authors:

    http://www.thomaspalley.com/docs/articles/macro_theory/mmt_response_to_wray.pdf

    The main attack on MMT seems to be an argument that the central bank must curb inflation generated at full employment and MMT does not say so. I am not familiar with all MMT papers, have read Hyman Minsky more carefully, and I think full employment without inflation would be possible with a flexible federal job guarantee putting a price floor under wages and with central bank ability to curb inflation when necessary by adversely impacting the pace of credit formation in the private sector. Congress should be pressed hard to tune federal spend, credit, and tax policies but due to political and other practical limitations it would be the price floor under wages with a job guarantee and monetary policy that drive the economy forward at the margin with some degree of price stability and approximate full employment.

    ReplyDelete
    Replies
    1. I read the Palley critique years ago. He has the sads because not everyone is using his version of the IS/LM. It was not a very constructive critique.

      Can we have “full employment” and stable prices/inflation? It depends upon what is meant by “full employment” is the first problem. If people are even semi-rational, it should be possible, given a policy framework that is consistent with price stability. In order to get the result that it’s impossible, you have to believe that firms will mindlessly bid up the price of labour in order to fill every open position. It is possible, but it seems unlikely.

      Delete
    2. A question:

      It is easy to view a wage-price spiral as an arms race, as employers and employees vie for a bigger share of GDP. However, a barter economy with real wages ought to work OK, right? What, then, is the problem in a monetary economy with nominal wages indexed to inflation? Thanks. :)

      Delete
    3. If you index everything, any price shock creates a “positive feedback” loop. I have sympathies with the view that this explains why energy price shocks caused the 1970s inflation, and not since.

      Delete
    4. Thanks, Brian. (For some reason my OpenID is not working now, so I am anonymous. But it's me. :)

      1) Only wages are indexed, not prices. Nobody asks for a raise. If we also assume that social security and other government benefits are indexed, then the people who receive those benefits are just along for the ride, n'est-ce pas?

      2) There is a lag, so that a general rise in prices produces a later general rise in wages. That lag means that employers may be able to increase their share of GDP before wages catch up, but, since wages are only catching up, why should the rate of inflation not converge?

      3) Employers and employees are not the only agents. And, in fact, by indexing wages the employees are not agents as such, but they are agents as consumers. Even if inflation induces them to spend, the lag in the rise in wages limits their capacity to do so. Also, the government may be able to take steps to increase overall saving, which will make price rises less effective, and that might lower the rate of inflation. Note that the inflation expectations of employees as employees has no bearing on future wages, because of indexing. So employers, observing a drop in sales after price hikes because of increased saving, will moderate price increases, which will in turn moderate wage hikes in the next round. In such a case, where is the positive feedback?

      4) In human systems positive feedback loops that threaten the stability of the system often occur, but then cease for reasons unknown. One or more of the agents stops participating in the contest. A large economy has a good deal of inertia, which means that there are a lot of negative feedback loops that could prevent runaway inflation. For instance, suppose that inflation expectations lead consumers to buy new cars instead of putting the purchase off. Do we think that they will then trade their new car in the next year? Also, if they take out loans to buy the cars, then paying back the loans will decreases their future consumption. Negative feedback. If people stock up on canned goods, their pantries will get full. Negative feedback. Many employers, having made relative gains in wealth because of the lag between price and wage increases, may become satisfied and not raise prices as much in the next round. They may gain market share as a result, and that may then induce other employers to reduce their prices. Negative feedback.

      Delete
    5. 4a) Sheer randomness of human behavior could provide inertia. For instance, a certain randomness in price hikes could have the same kind of moderating effect as employer satiation, in the example above.

      Delete
    6. If wages are indexed, prices will be. Everyone will shift to an “indexed” psychology.

      All you need is a big spike in energy prices, and then it starts the thing moving. The fact that there are time delays is what ensures that the loop keeps spinning. The oil price spike of a few months ago raises wages now, raising production costs, so prices are hiked.

      I grew up in the 1970s. Even as a kid, I knew how it worked. In an inflationary era, you become very CPI-aware. It’s politically toxic.

      Delete
    7. If both wages and prices are indexed, then neither employees nor employers are agents, except as consumers. Also, the rate of inflation with dual indexing should me constant. Furthermore, wage and price controls by the government should be easy.

      In the 1970s the government of Taiwan, facing double digit inflation, raised prices by 75% (IIRC) and then held them constant. I never found out how that worked out.

      I also lived through the 70s. Weren't inflationary expectations important in stagflation? Nixon's attempts at direct control did not work out very well.

      BTW, back then I read an explanation that made sense to me at the time, but I don't think so anymore. Here it is. OPEC raising oil prices increased prices in general. Petrodollars left the US for the Middle East, but then returned to the New York banks, because Middle East banking was not sophisticated, given the Koran's injunction against usury. Instead of investing the returning petrodollars in the US, which would have protected and generated US jobs, Citibank and others decided to make risky loans to Latin American countries, on the theory that countries do not go bankrupt.

      Delete
    8. These mechanical operations stories are not helpful when it comes to handling inflation.

      The mechanism is simple, it only appears complex because mainstream economists follow a crazy model of price determination.

      When oil prices go up in a shock, all users of oil have to raise their prices, or go out of business. This is faster than predicted by indexation, as the shock will not yet appear in the CPI index. These raised prices then hit the CPI. This then raises wages that are indexed to the CPI. Firms then have no other choice, but to raise prices to cover the rise in CPI. Thus, the wage-price spiral is born.

      Mainstream economists wrongly say that this cannot happen; other prices will magically instantly fall to offset the oil price rise. Pretty much all the confusion in economists’ discussions of inflation follow from the fact that this offset does not occur, so they have to invent stories to get around this fundamental problem.

      Indexation of prices just means that businesses will raise prices automatically in line with CPI, but they will also revise them upward by more if needed. You have to, if you want to stay in business, and your input cost rises by more than the CPI. That’s how inflation accelerates. The government cannot stop this, other than by wage and price controls. Some post-Keynesians get all misty-eyed by wage and price controls, but the mainstream consensus is that they do not work, and I would not waste political capital arguing with them on that score. Wage and price controls work in wartime, when you also have the ability to ram rationing down the population’s throats. In peacetime, nobody is going to put up with rationing.

      Delete
    9. Some years ago, when I first heard about NAIRU and other theoretically accelerating inflation, I tried to obtain it with indexed wages, and was unable to do so, except by assuming first, that employers set prices as a constant multiple of costs (that was my father's rule of thumb), and second, by assuming that all costs, not just wages, increased by the same ratio. Color me skeptical about runaway inflation without truly exceptional circumstances, for a number of reasons. The question is not necessarily one of staying in business, but more likely, one of taking a smaller share of the surplus. Business is a non-zero sum game.

      Also, despite the oil shocks and inflationary madness, wasn't the 70s a decade of increasing real GDP? (So I have heard. I didn't check it out.) Didn't the inflation scare set the stage for the stagnation of real wages?

      Delete
    10. To put it another way, an economy is not a clock. It is a complex adaptive system. To predict sustained maladaptive behavior, even from an oil shock, is asking too much. You have to make too rigid assumptions about economic behavior.

      Delete
    11. You don’t need it to accelerate forever. It just needs to accelerate from 2% to 10+%, which obviously happened in the 1970s.

      The whole point is that some prices rise, such as oil, which is determined in world markets. That sets off the mechanism, as it is passes through. Assuming that all input price rises are equal is equivalent to assuming that there is only a single good - an assumption which eliminates the possibiliy of wage-price spirals.

      Delete
  2. Macro for retarded economists
    Comment on Brian Romanchuk on ‘The Death of Mainstream Macro’

    “Since every act of spending results in income for somebody else, total spending for the economy as a whole equals total income. This is true by definition and is a basic building block in macroeconomics.” (Cooper)

    Both, orthodox and heterodox economists subscribe to this statement as the self-evident rock bottom truth of all of economics. Too bad that this statement is materially/logically false.

    The foundational error/mistake/blunder consists in the methodological fact that the two most important magnitudes of economics — profit and income — are ill-defined.#1 In order to see this one has to go back to the most elementary configuration, that is, the pure production-consumption economy which consists of the household and the business sector.#2

    In this elementary economy, three configurations are logically possible: (i) consumption expenditures are equal to wage income C=Yw, (ii) C is less than Yw, (iii) C is greater than Yw.

    In case (i) the monetary saving of the household sector Sm≡Yw−C is zero and the monetary profit of the business sector Qm≡C−Yw, too, is zero. The product market is cleared, i.e. X=O in all three cases.
    In case (ii) monetary saving Sm is positive and the business sector makes a loss, i.e. Qm is negative.
    In case (iii) monetary saving Sm is negative, i.e. the household sector dissaves, and the business sector makes a profit, i.e. Qm is positive.

    It always holds Qm+Sm=0 or Qm=−Sm, in other words, at the heart of the monetary circuit is an identity: the business sector’s deficit (surplus) equals the household sector’s surplus (deficit). Put bluntly, loss is the counterpart of saving and profit is the counterpart of dissaving. This is the most elementary form of the macroeconomic Profit Law. It follows directly from the profit definition and the definition of household sector saving.

    Loss or profit are NOT income. Only distributed profit is income. The profit theory is false since Adam Smith.#3 As a collateral damage all I=S or IS-LM models are false.

    Economists are too stupid for the elementary mathematics of accounting.#4 The statement total income equals total spending is simply false because of the all important phenomenon of credit. Equipped with credit the household sector can spend MORE than its period income (= dissaving) or in the opposite case LESS (= saving). Total spending and total income are NEVER equal, the foundational intuition of macroeconomics is false ― and so is all the rest. Macroeconomics is dead since Keynes.#5

    Egmont Kakarot-Handtke

    #1 For details see ‘How the Intelligent Non-Economist Can Refute Every Economist Hands Down’
    https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2705395
    and ‘Keynes’s Missing Axioms’ Sec. 14-18
    https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1841408

    #2 The elementary production-consumption economy is given by three macro axioms: (A1) Yw=WL wage income Yw is equal to wage rate W times working hours. L, (A2) O=RL output O is equal to productivity R times working hours L, (A3) C=PX consumption expenditure C is equal to price P times quantity bought/sold X.

    #3 See ‘Essentials of Constructive Heterodoxy: Profit’
    https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2575110
    and cross-references Profit
    http://axecorg.blogspot.de/2015/03/profit-cross-references.html

    #4 See ‘The Common Error of Common Sense: An Essential Rectification of the Accounting Approach’
    https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2124415

    #5 How Keynes got macro wrong and Allais got it right
    https://axecorg.blogspot.de/2016/09/how-keynes-got-macro-wrong-and-allais.html

    ReplyDelete
    Replies
    1. So you have proven that cash expenditures on investment by the business sector are truly an expense! You are a visionary, sir! We shall notify the Sveriges Riksbank “Nobel” Prize Committee immediately!

      Delete
    2. Brian Romanchuk

      You say “So you have proven that cash expenditures on investment by the business sector are truly an expense!”

      I have proven nothing of the sort but I have indeed proven that macro is dead since Keynes and that you have not realized it until this very day.#1

      Egmont Kakarot-Handtke

      #1 Note that nominal magnitudes Yw, C, Qm, Sm are normally NOT identical with cash payments. For the relationship between the nominal flow magnitudes and the stock of cash see
      The creation and value of money and near-monies
      https://axecorg.blogspot.de/2017/12/the-creation-and-value-of-money-and.html

      Delete
  3. Brian Romanchuk

    Your discussion of macro is the usual echo chamber economics. You do not ask what the correct approach is but are content with stating that current macro is crap and watching what your clueless peers are doing and opportunistically waiting who fetches the most likes on Twitter or Facebook. You argue:

    • “I had contact with hundreds of economists over the course of my career.” There is no use of talking with people who have not even realized that supply-demand-equilibrium is proto-scientific rubbish.

    • “One of the requirements of being minimally competent was being able to read another researcher’s paper, and compare what they demonstrated within the body of the paper versus what the abstract says the paper accomplished.” This does not even prove minimal competency but only how low the scientific standards in economics are.

    • “In summary, he [Krugman] argues that we could just use IS/LM … to analyse the policy response to the Financial Crisis.” If you have not realized until now that Krugman is not a scientist but a soapbox economist nothing can help you.#1, #2

    • “The other line of defense of mainstream macro is that young researchers are doing all this amazing new work.” Macro is axiomatically false and the new generation is busily but senselessly digging at the same wrong place as the old generation.

    • “One argument is that mainstream macro is more empirical.” Yes, but this does not help if the theory is axiomatically false to begin with.

    • “As should be clear, I pay almost no attention to the latest developments in mainstream macro. … Unless you are being paid to keep up with the latest research fads, it is probably safe to wait until some form of new consensus appears among researchers before actually reading the papers.” Very smart, indeed. What about getting off your ass and figuring things out for yourself?

    “The highest ambition an economist can entertain who believes in the scientific character of economics would be fulfilled as soon as he succeeded in constructing a simple model displaying all the essential features of the economic process by means of a reasonably small number of equations connecting a reasonably small number of variables. (Schumpeter)

    What is your simple macro model? If you cannot answer this question you are out of economics and out of the discussion.

    Egmont Kakarot-Handtke

    #1 Krugman is not an economist
    https://axecorg.blogspot.de/2016/09/krugman-is-not-economist.html

    #2 Mr. Keynes, Prof. Krugman, IS-LM, and the End of Economics as We Know It
    https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2392856

    ReplyDelete
    Replies
    1. Take a look at my “Bitcoin series”, (published just after this article). Note how I explain how “simple” models fail when applied to the straightforward question of valuing Bitcoin. How well is a simple model going to do when applied to all prices?

      Delete
    2. Brian Romanchuk

      Thank you for the reference to your Bitcoin articles.

      I have only one thing to criticize: the issue is macro and with Bitcoin you switch to partial analysis in good old Marshallian tradition. What your analysis, first of all, shows is that microeconomic price theory does not work and, worse, that the one-size-fits-all explanation with supply-demand-equilibrium explains in fact nothing and never has.#1 Neither has general equilibrium theory, the very core of economics. In other words, 200+ years after Adam Smith, economists still do NOT know how the price mechanism works.

      You say “Note how I explain how ‘simple’ models fail when applied to the straightforward question of valuing Bitcoin. How well is a simple model going to do when applied to all prices?”

      Indeed, the explanation of the price mechanism has to start from macrofoundations and NOT from microfoundations or partial analysis. So we are back at macro.#2

      The elementary macroeconomic Law of Supply and Demand is shown on Wikimedia.#3 This price formula gets, of course, longer with the increasing complexity of the economy. All these details are not needed at the moment.

      The elementary macroeconomic Law of Supply and Demand says:#1

      (i) An increase of the expenditure ratio rhoE=C/Y leads to a higher market clearing price (the Greek letter rho stands for ratio). An expenditure ratio rhoE greater than 1 indicates credit expansion, a ratio rhoE less than 1 indicates credit contraction. Credit expansion/contraction, in turn, affects the quantity of money.
      (ii) An increase of the ratio of wage rate to productivity W/R leads to a higher market clearing price.

      Roughly speaking, the macroeconomic Law of Supply and Demand explains the price level and its development over time. The equation contains only measurable variable and is therefore testable in principle. Starting with one market, the way forward is differentiation.

      The crucial differentiation is between primary markets (= perishable consumption goods from current production) and secondary markets (= durable goods). Both markets run on entirely different principles.#4 The formula above holds for the primary market.

      In brief, the nominal anchor of the whole price system is unequivocally given with the macroeconomic Law of Supply and Demand. But from primary markets to secondary markets and then to Bitcoin is a longer analytical way.#5 In any case holds, if it isn’t macro-axiomatized it isn’t economics and microfounded price theory is dead.

      Egmont Kakarot-Handtke

      #1 The monstrous utility-supply-demand-equilibrium failure
      https://axecorg.blogspot.de/2016/04/the-monstrous-utility-supply-demand.html

      #2 This is the correct core of macroeconomic premises: (A0) The objectively given and most elementary systemic configuration of the (world-) economy consists of the household and the business sector which in turn consists initially of one giant fully integrated firm. (A1) Yw=WL wage income Yw is equal to wage rate W times working hours. L, (A2) O=RL output O is equal to productivity R times working hours L, (A3) C=PX consumption expenditure C is equal to price P times quantity bought/sold X.

      #3 Wikimedia, Law of Supply and Demand, pure production-consumption economy
      https://commons.wikimedia.org/wiki/File:AXEC101.png#{{int:filedesc}}

      #4 Primary and Secondary Markets
      https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1917012

      #5 The creation and value of money and near-monies
      https://axecorg.blogspot.de/2017/12/the-creation-and-value-of-money-and.html

      Delete

Note: Posts are manually moderated, with a varying delay. Some disappear.

The comment section here is largely dead. My Substack or Twitter are better places to have a conversation.

Given that this is largely a backup way to reach me, I am going to reject posts that annoy me. Please post lengthy essays elsewhere.