19 Oct 2013

Accounting Identities Versus Economic Theories

Chicago School, DeLong, Eugene Fama, Krugman, MMT 24 Comments

OK I’m going to crowdsource this one, because I’m guessing in the comments there will be two camps who will eventually reach the truth (at least in my mind). In a previous post I alluded to Brad DeLong bringing up a 2009 Fama essay on the impotence of stimulus spending. Upon seeing DeLong’s post, Krugman reminded everyone that it was precisely this essay from Fama (and a similar one from John Cochrane) that had made Krugman announce we were once again in a Dark Age of Macro.

Here we can summarize the offending passages from Fama, in the eyes of our intrepid Keynesian bloggers:

There is an identity in macroeconomics. It says that in any given year private investment must equal the sum of private savings, corporate savings (retained earnings), and government savings (the government surplus, which is more likely negative, that is, a deficit),

PI = PS + CS + GS

Even when there are lots of idle workers, government bailouts and stimulus plans are not likely to add to employment. The reason is that bailouts and stimulus plans must be financed. The additional government debt means that existing current resources just move from one use to another, from private investment to government investment or from investment to consumption, with no effect on total current resources in the system or on total employment.

So I think most libertarian Austrians will recognize such an argument; I’m sure I made it myself on the radio when the stimulus package was big news. I definitely remember using the line–which I don’t think is in Fama’s piece, but it’s in Cochrane’s–that “If the government spends $1 billion on infrastructure, that money has to come from somewhere.”

So, DeLong and Krugman hate this. Specifically, Krugman argues that it fallaciously relies on an accounting identity to derive causality, which you actually need an economic theory to do. As a general rule, Krugman is right: I used the same approach to critique the MMT camp, as well as Krugman himself (on international trade, believe it or not).

Back to Fama: I’m tempted to resurrect his position by saying, “OK sure, you need one more element than just the accounting tautology. You need to assume that saving must precede investment, which strictly speaking you wouldn’t know just from the identities. But give me a break, of course saving must precede investment. Thinking otherwise is precisely why the MMT guys and the Keynesians get into trouble.”

OK so now my question: Is that really the way to handle this, and can I sleep comfortably at night? I wonder, is this still actually a fallacy, which would be apparent if we switched contexts? For example, suppose someone named Gene Famished were to argue, “We’re stuck in a deep recession and we will never get out of it. Sure, investors might spend money opening a new store and hiring 1,000 people, but every dollar they spend on the store just means there’s one less dollar spent elsewhere. It won’t create jobs on net, just move jobs around.”

See what I’m saying? I think “our side” is fine, and just needs to be a little less sloppy in our arguments, but I’m obviously biased.

Discuss.

24 Responses to “Accounting Identities Versus Economic Theories”

  1. Enopoletus Harding says:

    Gaah! Where did the post on the Slate article go?
    Also, how does one know when a comment on this blog was posted? This is probably the only blog I read where I’ve found no way to find when any comment was posted.

    • Bob Murphy says:

      I took the whole post down E.H. in response to your comment. I hadn’t actually read “Dear Prudence”‘s original thing, only the backlash. I didn’t see any problem in the quotes the article took from her piece, but the stuff you posted made me understand more why the feminists were saying she was blaming the victim. So, rather than get into minutiae I just took the whole thing down.

  2. DMS says:

    I don’t think the response needs to center on the accounting identity as presented – the response is more foundational than that. In my humble opinion, bad economic thinking often has two basic problems. One is an implied “ceteris paribus”, when in fact the world is dynamic and responsive. The second is to ignore the informational context, and the informational discovery aspect of economic activity. Of course, Keynesian bad economic thinking is rife with an implied “ceteris paribus”, and seems to willfully ignore informational implications.

    On this specific topic, Austrians in particular, but other non-Keynesians as well, understand the dynamic and information-specific nature of economic relations, while the Keynesians simply do not. The seemingly plausible notion from the Keynesians that government spending will be stimulative when resources are idle simply ignores the following issues:

    1. Which particular resources are idle, and what specific capacity is “slack”? How would we know with clarity, and could we agree? How does government spending flow effectively to these areas when the information is muddled (as it always is), or when Public Choice issues obtain (as they always do)? Obviously the Austrian capital theory directly relates to this.

    2. If the government spending doesn’t take away from the private sector because it is merely employing otherwise “idle” capital, what will be the response of the holders of the idle capital after it is redirected? If an individual or an institution constructs his/her/its portfolio with a certain element of idle cash (however defined and for whatever reason), and if that idle cash is expropriated (taxed away, or lured away with government bonds) how could we possibly assume no portfolio rebalancing, and a subsequent draining of capital out of “non-idle” pursuits? It seems clear that the burden of proof should be on the Keynesian set of assumptions, as it flies in the face of common sense and economic theory.

    3. If we put aside vacuous phrases such as the “Fetish for Liquidity”, or “The Paradox of Thrift”, and “Animal Spirits”, and if we instead assume that resources and capital are idle for some reason, would not forced government employment of those resources necessarily confuse and corrupt the informational discovery process necessary for ultimate recovery? That is, should we assume some untestable psychobabble such as a “fetish for liquidity”, or should we assume that there is relevant and necessary signalling from consumers, investors, and employers and that the idle resources are a consequence of economic choices – a symptom, not a disease? Is not the proposed government “stimulus” far more likely to result in iatrogenic medicine, rather than in a recovery of the patient?

    Instead I would assert that the concept of employing otherwise idle resources by governmental fiat cannot possibly be stimulative in a dynamic and responsive market with complicated informational content, even if one believes involuntary unemployment is rampant, itself a dubious assumption.

    • Tel says:

      If you bring up the accounting identity:

      PI = PS + CS + GS

      So where are the idle resources? How do they show up in this picture?

      Try it with other macroeconomic accounting identities like the MMT thingy with GDP. You still can’t find those idle resources.

      As you correctly point out, whether a resource is idle or not, can only be determined by personal opinion. Accounting can’t provide such opinions, it merely tallys up transactions and that’s all it does.

    • Rick Hull says:

      Tel brings up a great point about where the idle resources are represented in the accounting identity.

      I can’t help but respond to this, as much as I’d like to never again feel compelled:

      > One is an implied “ceteris paribus”, when in fact the world is dynamic and responsive.

      Ceteris paribus is a technique for thought experiments. The very premise of “holding” ceteris paribus implies that it does not hold in the real world. To discuss this device any further risks a complete derailment. i.e. if you have to ask, you might just never know. As far as I know, the only worthwhile critiques (kritiks, even) involving c.p. usage are when the very thing in question implies that c.p. cannot possibly be true. e.g. The sun going nova will, c.p., be good for the American economy.

      Which is not to say that c.p. cannot be misused. However, criticism along those lines must focus on the misuse, not a general critique of c.p. usage itself.

      • Tel says:

        I think there’s a pretty big difference between the thought experiment, “What if there were idle resources and government could do something about that?” and the typical Keynesian claim to practice scientific empiricism based on careful measurement.

        Since two people might disagree with regard to the exact same resource (whether it was “idle” or not) how can any accounting system accurately describe this? If your chosen method of data collection cannot even represent the quantity in question, don’t worry about any claims to empiricism, because nothing has been measured here.

        That was my complaint, and I don’t get how c.p. can bypass such a fundamental problem.

        If you want a practical example, take a look at unemployment measurements, we don’t even agree on a definition for what makes someone unemployed, let alone what factors cause it. The techniques used to gauge unemployment change every decade or so to ensure no historical continuity, and yet it is probably the most often quoted economic statistic.

  3. Cosmo Kramer says:

    If we focus purely on identities, then we can conclude that nothing is happening at all times. So MMT will cut back one layer and astound audiences with the sectoral balances chart……. Of course it is true that MMT can theoretically create and sustain full employment, but this is only true if it improves productivity.

    “but every dollar they spend on the store just means there’s one less dollar spent elsewhere. It won’t create jobs on net, just move jobs around.””

    Since we can conclude that new fiat derives its purchasing power from the existing stock, we can deduce that “nothing” is happening. But this isn’t true. Our wealth is the total sum of goods & services we can buy. All the spendthrifts want (real terms) is to shift goods & services from the productive sector of society to the unproductive sector. The steel will go into making bombs instead of cars etc etc. We are where we are because humans have put resources to use in an efficient way.

    That is the easy part. The hard part is proving to them that they are creating moral hazard.

    I think the Fama focuses far too much on the accounting identities. He only briefly mentions the theory behind why gov’t stimulus won’t do any good. There is nothing wrong with using accounting identities. I suppose it is how they are used. Krugman’s critique was a bit unfair.

  4. Lord Keynes says:

    “You need to assume that saving must precede investment, which strictly speaking you wouldn’t know just from the identities. But give me a break, of course saving must precede investment. Thinking otherwise is precisely why the MMT guys and the Keynesians get into trouble.”

    That just conflates real saving with financial saving, and sets up a straw man.

    The assertion that prior real resources must exist for investment or consumption is not denied by Keynesians or MMTers.

    But where unemployment, idle capital goods, unused excess capacity, unused inventories and stocks, other idle resources, and resources available by international trade exist, prior “monetary” saving is not necessary for investment or consumption: banking systems can create new money or governments can monetise deficits.

    • Cosmo Kramer says:

      Totally incorrect.

      “banking systems can create new money or governments can monetise deficits.”

      New money just allows the new money holder to re-allocate EXISTING resources. New resources weren’t summoned into existence.

      “But where unemployment, idle capital goods, unused excess capacity, unused inventories and stocks, other idle resources,”

      Entirely bogus. Unemployment means more competition for jobs. You ignore this. There is no such thing as an idle anything. The money in my bank account means x less $ to compete with those that are currently spending(more effects exist). and then this leads us to the obvious question: at what point does something become idle? How long does $ in my bank have to sit there before a Keynesian calls me a miser? All $ is owned at all times. So what is the correct money velocity?

      You believe that new money does not take its purchasing power from existing money. You foolishly do not understand that printing $100 and giving it to Unemployed Urkle only forces the private sector to give up goods and services to U. U. Providing him a shovel to uncover buried jars of $ can only make us worse off.

      • Lord Keynes says:

        New money just allows the new money holder to re-allocate EXISTING resources. New resources weren’t summoned into existence.

        And nobody said new money “summons” new resources “into existence” ex nihilo. What is said above is that prior *monetary* saving is not required for a business to obtain credit via new money and engage in investment.

        “There is no such thing as an idle anything”

        Yes, I expect people like you could explain this — with a straight face — to the millions of unemployed. I expect you also believe that “empirical evidence never proves anything,” and so on.

        • Cosmo Kramer says:

          What is the correct money velocity?
          What is the correct money velocity?
          What is the correct money velocity?

          “Yes, I expect people like you could explain this — with a straight face — to the millions of unemployed.”

          Yes, I am very serious. 10 people seeking 1 job has what effect on the wage rate offered? How about 20 seekers per job?

          Now what happens if you force prices to double and give all of the unemployed a minimum wage rate?

          The purpose of a wage is to be able to obtain goods & services either now or in the future. Employing everyone that “wants” to work doesn’t summon into existence extra goods and services for all of these individuals. Instead, it takes away from everyone else. Employing everyone wouldn’t solve the problems with the economy.

          “What is said above is that prior *monetary* saving is not required for a business to obtain credit via new money and engage in investment.”

          New money derives its power from the existing stock. Your statement ignores what happens in real terms.

          There is no free lunch. Employing supposed “idle” resources is not free or a negative cost.

        • Major_Freedom says:

          Theoretically speaking, all available resources could be in use while there are people looking for employment but cannot find any, because their asking nominal wage rates is too high relative to the nominal demand for labor.

  5. Tel says:

    If the government spends $1 billion on infrastructure, that money has to come from somewhere.

    Wasn’t this the thing that got you into trouble with Mosler and his subway tokens? Of course the tokens are issued by whatever issuing authority first and afterwards someone spends them. There’s no constraint on where the money comes from, which has been solidly demonstrated by the Fed and their waves of QE money from nowhere.

    However, if the government deploys 1 million bags of concrete to build a bridge, the concrete has to come from somewhere. The Fed can’t print concrete.

    Remember that accounting serves two purposes: stock accounting, and financial accounting. It’s just that economists tend to ignore the stock. The money is nothing more than a token that makes it easier to track the stock, just like subway tokens are meaningless unless the entity issuing the tokens first owns a subway.

    The Keynesian argument is that they have a special nose with the ability to sniff out idle resources. They turn those idle resources into non-idle resources, and presto a million bags of concrete appears from nowhere. The accounting identity can’t prove this right or wrong, you just have to take it on faith that these “idle resources” will spring into action.

  6. Keshav Srinivasan says:

    Bob, how would you respond to this old post from Brad Delong, criticizing John Cochrane “the money has to come from somewhere” argument?

    http://delong.typepad.com/sdj/2009/01/time-to-bang-my-head-against-the-wall-some-more-pre-elementary-monetary-economics-department.html

    He gives a pretty simple example that he claims shows that Cochrane’s point is wrong.

    • Bob Murphy says:

      Let me check it out Keshav, it’s possible he’s doing exactly what I was worried about in this post. thanks

      • DMS says:

        DeLong’s example illustrates a functioning private economy, and that is a wonderful thing. But that is not the issue. The issue is what if there is an abstract entity G (obviously for government) that originally borrows the $500 that Carol was previously “hoarding”, and injects those dollars at Beverly, and DeLong’s example unfolds exactly as he write?. Is this a good thing? Per my original post:

        1. How does G know that Beverly should be the recipient of funds? How can G possible know that what will unfold is the scenario that DeLong narrates? Why couldn’t an equally probable scenario be one in which Beverly engages in wasteful, non-productive enterprises?

        2. If Carol was “hoarding” the money because she was fearful of the environment, and G taxes it away (or borrows it by offering above-market terms), does her fear of the environment go away? Won’t she need to replace her cash hoard, and isn’t an equally plausible narrative alternative to DeLong’s one in which she borrows back the money from the system, perhaps even directly from Beverly?

        3. What if Carol is “hoarding” her cash because she is displeased with the offerings of the marketplace, perhaps feeling there is an excess of home improvement and home catering, and she wants no more of it. Then when G effectively borrows and lends Carol’s money into DeLong’s example, hasn’t G just exacerbated the issues that Carol’s behavior was highlighting?

        DeLong’s comments on velocity are intellectual hand-waving and a non sequitur.

        • AcePL says:

          Well, I have a question: from where Carol had the money she loaned to Beverly?

      • EJMRTroll says:

        Cochrane is definitely wrong here, but to understand clearly why he’s wrong, you need a model. A good place to start is Woodford’s AEJ Macro paper on the government expenditure multipler – which, ironically, can be found on Cochrane’s own site: http://faculty.chicagobooth.edu/john.cochrane/teaching/Monetary%20Economics%20PhD%20course/woodford_simple_analytics.pdf

        Let’s take a very simple environment, essentially the basic NK model: an economy with a single final good, single representative household, and some price stickiness. In this environment, we can make a trivial but far-reaching statement – as long as the real interest rate does not change in response to a temporary increase in government spending, and government spending is separable from household consumption in the utility function, the output multiplier on that increase is exactly one.

        In a way, this is completely obvious – in general equilibrium, agents make decisions in response to prices. If spending by the government “crowds out” household spending, then this effect must be mediated somehow by the price signal (unless it arises from an interaction between government spending and household spending that’s directly in the utility function). But in a one-good economy, for a temporary shock the only possible price signal is the real interest rate – and if the real rate doesn’t move, there simply isn’t any crowd-out mechanism. (And, of course, at the zero lower bound the real interest rate will anything move in a way that magnifies the government spending. This is the basis for the NK multiplier being above 1.)

        I’m not saying that I think such a simple model is the final word on fiscal stimulus. I’m merely emphasizing that there must be something deeply wrong with Cochrane’s logic. He claims that it’s all about “accounting”, but in fact one can write a simple model where the “accounting” is immaculate and his assertions are false. Ergo, something is provably wrong with his logic.

        This is why models are useful.

        • Tel says:

          But in a one-good economy, for a temporary shock the only possible price signal is the real interest rate – and if the real rate doesn’t move, there simply isn’t any crowd-out mechanism.

          It’s a one-good economy, and this good is a physical good (like apples) so there’s no fiat money. That means the government spending cannot possibly be larger than government tax rake (unless they are printing apples)… seems like a pretty obvious mechanism to reduce private consumption of apples, by snatching those apples out of private hands.

        • AcePL says:

          Maybe we should leave the models to theoreticians. I’m down to earth SMB entrepreneur (with all possible emphasis on small) and I find all those models with lots of assumptions ridiculous. I would like to hear a straight answer to Cochrane’s question as If I stood in front of my bank right now. Not a description of orchard. Because the economy out there I understand. F*****g models? Give me a break…

    • Matt G says:

      DeLong says: “And all this has happened without printing any extra money.” But this is indeed an example of money printing. Those IOUs they issue are acting as money. Cochrane’s argument (at least as DeLong presents it) is assuming the absence of new money creation.

      DeLong conflates the issuance of new debt-money with an increase in the velocity of a constant stock of $500.

      Now if he had said “Cochrane’s argument is simplistic; in the real world money is created and destroyed all the time”, he might have had a better point. But that’s a different argument.

  7. Major_Freedom says:

    The best way to solve this dilemma of saving and investment is to just define the terms. Once the terms are defined, good old fashioned economic intuition and deduction can be used to make strong conclusions.

  8. John B says:

    Profitable private sector spending actually creates value by combining inputs into something worth more. Government spending may or may not create value. There is no way to know because it doesn’t go through the profit and loss test. Given government incentives, it is reasonable to assume much greater private sector efficiency.

    Because of this, the money and workers that went into creating jobs in the private sector will end up creating value or net while government spending probably will not.

Leave a Reply to Bob Murphy

Cancel Reply