Home > Main > The Useless Intertemporal Government Budget Constraint: Not Economics

The Useless Intertemporal Government Budget Constraint: Not Economics

August 30, 2011

One of the amazing things about economics is that the major flaws are defined out of existence and then just ignored by most of the profession.

I’ve already went through many reasons why monetary policy sucks. These aren’t controversial statements.  Monetary policy is a bad way to try and influence an economy even if it works.

Then, one of the most common statements in economics, I = S, gets the accounting wrong.   This is the MMT view of I = S and of most of modern macroeconomics – that it gets the accounting wrong so it can’t get the economics right.

But probably the most egregious error in economics is the Inter-temporal Government Budget Constraint.

First, I think the no-Ponzi assumption is stupidly wrong.   It’s impossible to tell ever tell if it holds, and so people can make up anything they want about the IGBC and they might be right.

It turns out that Minnesota fed president Kocherlakota agrees with me, as I’ll show later in another post.

Second, I am  sure that even if it does hold, its worthless to use anything but what we know today about inflation to make decisions about it.

(Thanks again to Peter D for the excellent summary.)

So if the assumption is a boneheaded assumption, we shouldn’t use the IGBC.  But even if the assumption holds, then we must take primary market information seriously and only use inflation as our policy guide.

Now, I have a 3rd critique of the IGBC.  The IGBC isn’t economics.   Nick Rowe says that economics requires most models to contain elements of both supply and demand to be considered to be economics.

This is not a controversial statement.  It’s a common sense observation about what is and what isn’t economics.

And by this common sense observation, the IGBC isn’t economics.

Here is a link to the cannonical Walsh derivation of the IGBC.   Go to page 136.   You can get a similar derivation in Fullwiler’s “Interest Rates and Fiscal Sustainability“, pages 7-9.

Where is the demand for (G-T) in the IGBC?  There is none.

We know:

  1. G – T is the net financial assets issued by the government by identity.
  2. There is so much demand today for the assets generated by G -T that investors are willing to knowingly lose money over a period of 5 years just to have these assets.   We have negative real interest rates in the 5 year, as of August 29th, 2011, indicating massive real demand.  People hold over $14T worth of (G-T) in our real world – indicating massive nominal demand.
So the IGBC isn’t economics.  It can’t be economics, because there is no demand for G – T in the equation.
For this constraint to be economics, it needs to reflect the staggering demand for G -T.  We can call them “net financial assets” or “government borrowing” or something like that, but we know by identity that these equal G – T.
Yet, according to Walsh:
“In most traditional analyses, fiscal policy is assumed to adjust to ensure that the government’s inter-temporal budget is always in balance, while monetary policy is free to set the nominal money stock or the nominal rate of interest. “

We live in a world where the demand for (G -T) by the U.S. government has been high for the entire life of the country.  We live in a world where right now, today, this demand is so high that people are gladly losing money to purchase (G – T).

Yet, ‘most traditional analysis’ use a model that assumes zero supply or demand for G – T.

This isn’t economics.  It’s religion.

This is seeing the world through a stunted belief regime.  Enforcing this through deliberately hiding the truth of the matter though deliberately confusing non-experts and appeals to authority is anti-democratic.

Of course, the second you put demand into the IGBC, it becomes an equilibrium condition, not a constraint.  And then you get something very much like MMT, called the Fiscal Theory of the Price level.


Categories: Main Tags: ,
  1. Clonal Antibody
    August 30, 2011 at 1:12 pm


    Another way to put it is – “demand for (G-T)” is really “demand for new money.”

    Traditional economics says that this demand is met by new credit creation by private banks. But we know that credit creation by banks does not really create new money if both the principal and the interest is paid back to the banks, even if all the interest is spent back by the banks into the economy.

    This situation leads to the necessity of real growth, and/or an ever increasing real debt burden on the society. This ultimately is always untenable, and invariably leads to economic and societal collapse.

  2. August 30, 2011 at 1:47 pm

    Fred has an intertemporal budget constraint. Call it the IFBC. Fred has a finite life, and no heirs, so we don’t have to worry about the no-Ponzi condition for Fred (unless Fred is very good about lying about his assets, Madoff-style).

    The IFBC is not economics. It is not a theory of Fred’s (or anyone else’s) behaviour. But understanding the IFBC might help us build a theory of Fred’s (and others’) behaviour. And a theory of Fred’s behaviour which violated the IFBC would upset the accountants. And the accountants would probably be right to be upset.

    Suppose initially that Fred is planning/expecting a sequence of income and expenditures that satisfy the IFBC. And everybody else is expecting Fred to conform to his IFBC too. I.e , Fred is planning to pay his debts, and everyone is expecting Fed to pay his debts, before he dies.

    Then Fred decides he wants to consume more today. I.e. Fred’s desired consumption today increases (his desired saving today decreases).

    Question: What are the consequences of this?

    Now you need some economics. Will the quantity of consumption goods produced and sold increase to match Fred’s increased demand? Will the price have to rise? Or will other people’s demand fall to match Fred’s increased demand? Will the supply of loans to Fred be able to satisfy Fred’s increased demand for loans? Will the rate of interest rise? Or will others’ income rise enough from the increased income they get from Fred’s increased demand for consumption lead them to want to lend Fred all their extra income, without the rate of interest rising? Etc.

    The IFBC cannot answer those questions. Nor can any accounting identity. Not by itself, anyway.

    But, at this point the defenders of the IFBC step in, and say “Hang on! Your question is incomplete. If Fred is planning to increase his consumption today, what about the IFBC? Is Fred: planning to cut his consumption tomorrow; planning to increase his income tomorrow; or is he planning to default on his extra debt? Or maybe, he’s a counterfeiter and he’s planning to start printing money? And equally importantly, what do other people expect Fred to do? Because if they expect him to default on his new debt, maybe they won’t lend him it in the first place? There’s a whole lotta questions here you’re ignoring, that the IFBC forces you to answer!”

    And the defenders of the IFBC would be right.

    The defenders of the IFBC are just accountants, insisting that we pay attention to an intertemporal accounting identity. They can’t answer the question: “what happens?” But they an force us to reveal our implicit assumptions when we ask that question. And they can force us to check that our answer adds up right.

    Obviously, the government is different from Fred. But the IFBC is a good place to *start* thinking about the IGBC. Then think of the differences:

    1. Governments *may* live forever. Or do they? The sun will go supernova eventually. Nerdy people (I mean no disrespect) can have really great fun here on the Chinese Hangman paradox. But the no-Ponzi assumption cannot be as simple as in a case where you know when Fred will die, and so won’t lend him any more on the day before he’s going to die, etc..

    2. Governments have power to earn income through taxation in ways that Fred can’t. (But are those powers limited, either by economics or politics?)

    3. Governments can earn income by printing money and Fred can’t. But, the expectation they will print may itself be important for the answer to our question. Plus, in real terms, there’s a political or economic limit to how much income the government can earn, in the long run, by printing. Even the Zimbabwean government couldn’t earn more than 100% of Zimbabwean GDP by printing, though it certainly tried.

    4. Fred has his own interests at heart when he decides what to do, and maybe the government has the national interest at heart.

    5. Money, as medium of exchange, is weird. It’s not like bonds. People accept money in exchange for goods because they can think they can pass it on to someone else, not normally because they plan to hold it. And it usually pays a negative real interest rate. The demand for money isn’t like the demand for bonds. We accept it, even if we don’t want it, because we all think someone else will accept it in turn.

    6. Etc.

    Sorry this is so long.

  3. August 30, 2011 at 2:19 pm

    BTW: I actually rather liked your post. Disagree with a lot of what you say, but that matters less than that I liked your questions.

    I am not an Austrian, but. On questions like this I actually find it helpful to think back to Hayeks’ way of looking at things. Hayek thought in terms of people having plans, for their current and future actions. Demands and supplies would be just one example of their plans. And their plans depend on their expectations of others’ actions. An “equilibrium” is where all those different people’s plans and expectations are mutually consistent. If one individual changes his plans, that will probably falsify other people’s expectations, or make their planned actions mutually inconsistent.

    Within that Hayekian context, we can ask: “do individuals’ plans conform to their intertemporal budget constraints? Do their expectations of others’ future actions conform to others’ intertemporal budget constraints?”

    In particular we can ask that question about whether the government’s plans conform to the IGBC, and whether people’s expectations of government behaviour conform to the IGBC.

  4. August 30, 2011 at 2:33 pm

    The IGBC is essentially the same as living for the afterlife. It requires the current generation to live like monks so that the next life will be somehow better.

    And there’s absolutely no evidence that the afterlife will be better than this one. So better to live for today, because for certain if we sit around with 7 to 20% of our productive resources idle we are not building a better tomorrow.

    Much better to get them back to work and build something from their efforts.

  5. TC
    August 30, 2011 at 2:49 pm

    Hi Nick,

    Your comments are what reasonable, intelligent people think about the IGBC – it’s a starting point, not a constraint.

    Before we get to far – I read your stuff all the time, and really like it. I don’t agree much when you get into money, but that’s because I’ve swallowed the red pill of core MMT.

    You’ve seen that some considerable amount of the posts here are inspired in one way or another by your posts. 🙂

    Nearly all of your points can be reduced to: What’s the demand function for G – T? Mainstream econ doesn’t do a demand function for G – T within the IGBC. Then it defines G – T out of the major models of interest rates as you showed.

    We can attempt to answer these good questions raised by the accounting of the IGBC by introducing a demand function.

    When we do introduce this demand function, the IGBC is no longer a constraint but an equilibrium condition. This means we get the fiscal theory of the price level – which is something very much like MMT.

    Mainstream econ hates the fiscal theory of the price level. Perhaps hate is too mild of a word.

    Many economists don’t even know this idea exists. Many of those that do refuse to recognize it as having any legitimacy at all. Kocherlakota and Buitler wrote diatribes about the FTPL.

    Check out my next post on the FTPL and what Kocherlakota’s famous paper on the FTPL means for the IGBC.

    Here’s the crux:

    The FTPL does not exist in a universe by itself. It exists in this universe as the policy state of non-Ricardian policy as opposed to the IGBC Ricardian policy. Government policy must be either Ricardian or non-Ricardian.

    My strong contention is that we cannot tell if we have entered non-Ricardian policy with any confidence whatsoever. Kocherlakota shows this is the case in his paper – its impossible to tell if we’ve entered non-Ricardian policy. Perhaps he made a mistake.

    If we can’t tell that we’re acting in a non-Ricardian way, we cannot tell if we are acting in a Ricardian way. To satisfy our claim “we’re in Ricardian Policy”, we must be able to say “we’re in non-Ricardian policy and need to get back to Ricardian Policy”. But as Kocherlakota points out, it’s impossible to tell if we’re in a non-Ricardian state.

    It’s like saying “All swans are white”, but then finding out that “I am blind and can’t tell if swans are other than white”. The statement “All swans are white” becomes meaningless because it has no context.

    This has massive consequences for the IGBC, for much of monetary economics, and for our political debate.

    Re: Bonds vs. Money: Many people in MMT regard Treasury bonds as “Tradable Government Guaranteed Term Saving Accounts.”

    lol – more soon.

  6. August 30, 2011 at 4:54 pm

    TC: “Nearly all of your points can be reduced to: What’s the demand function for G – T? Mainstream econ doesn’t do a demand function for G – T within the IGBC.”

    Fair question.


    Assume closed economy.

    Assume away all assets like land, antique furniture, etc. (That’s a biggy, and one I object to, but leave that aside)

    So the only tradeable assets are: newly-produced capital goods (“machines”); plus money and government bonds. “G-T” is the flow of new money and new bonds, (provided we include interest on existing bonds in minus T as a transfer payment).

    “Mainstream” (‘what’s that?) econ has a consumption demand function Cd and an investment demand function Id. Define Sd=Y-T-Cd. Then the (flow) demand for your “G-T” is given by Sd-Id. If the mainstream theory specifies a consumption demand function and an investment demand function, then it (implicitly) specifies a demand function for your “G-T”. Its a (flow) demand function for government bonds+money. If you are planning to do something with your disposable income other than spend it on newly-produced consumption and investment goods, then you must (by default) be planning to buy bonds and/or add to your stock of money.

    The above applies to models like ISLM and many New Keynesian models.

    We can add private bonds into the mix, if we assume they are perfect substitutes for government bonds. Someone who wants to borrow has a negative demand for “bonds”. When we aggregate up, across all private agents, private bonds disappear, because they net to 0.

    There’s a lot to object to in such a model. But it does (implicitly) have a demand for G-T.

    How does the IGBC come into this?

    1. If you are *advising* the government on G and T, you might want to incorporate the IGBC into your advice. You are saying what you would do if you were the government. And you might want to keep in mind that anything you recommend for G and T today will have implications for G and T in the future, if the IGBC is true. Make your own policy recommendations stock-flow consistent, if you like.

    2. If the IGBC is true, and does indeed act as a constraint on government the way IFBC acts as a long term constraint on Fred, then when government changes G-T today, people may expect G-T to change in the opposite way in future. In other words, when people see G-T increase today, they expect G to be lower in future, or taxes to be higher, or money-creation to be higher. And those expectations may affect their behaviour today. Ricardian Equivalence is just one example of this. The IGBC gets incorprated into the consumption demand function. This creates a “Say’s Law for government bonds”. An increase in the supply of government bonds (due to a tax cut) creates an equal increase in demand for government bonds, with no changes in income, interest rates, or prices.

    I’m rambling.

    FTPL is sort of…..too mainstream to be mainstream. Only true hardline believers believe it. It’s so ultra-orthodox it’s heterodox. Assume zero frictions, and perfectly flexible prices, and that government is just like a corporation, so government bonds are like shares in IBM, and you get FTPL.

  7. TC
    August 30, 2011 at 9:32 pm


    I have some things to do so I cannot respond in full until tomorrow.

    But I will say two quick things.

    1. Thanks again for engaging!
    2. I am not saying the IGBC is true or false. I am saying it isn’t even false. I am saying the model is misspecified, that its the equivalent of trying to divide by zero, that it isn’t even wrong. http://en.wikipedia.org/wiki/Not_even_wrong

    This is different than being wrong.

    It comes down to the fact we can’t tell it’s wrong due to ex ante vs. ex post knowledge problems. Kocherlakota runs into the same problem at a different point when he tries to apply non-Ricardian policy.

    He finds out he can’t ever know if he’s applying non-Ricardian policy. It’s not falsifiable, because we can’t demonstrate a non-Ricardian policy. He tries to get around this, but can’t, so resorts to name calling.

    In other words, the no-Ponzi (RE) is an impossible assumption for governments because it implies impossible knowledge. Therefore, the IGBC cannot be a constraint. Assuming it’s true – that it’s a constraint – is worse than actually being wrong.

    I want to be clear about this: It is impossible to show that we are violating the IGBC or that it is a constraint in any way. Impossible. No matter how much we spend in excess of taxation. Not matter how little.

    I am pretty sure Kocherlakota realizes that the inability to demonstrate a non-Ricardian policy dooms the Ricardian assumption to being non-scientific – check out the last page or so. I think it can be described by the words “flailing about”

    So saying we might want to consider the IGBC as a way to guide policy would be bad advice. The IGBC isn’t just a flawed model that has no demand – even though the lack of demand is extremely disturbing.

    It’s a misspecified model at it’s root that cannot give us any insight or help us make a decision. Specifically, the misspecified part of the model is Ricardian Equivalence applied to G – T.

    This flows through to other models in this way.

    Any models that imply demand for G – T through defining demand for Cd and Id and put any Ricardian equivalence constraint on that flow (stock) aren’t even wrong. They are mis-specified models.

    It’s mis-specified even if the RE constraint sneaks in because its applied to households that purchase either capital goods or bonds/money instead of the G – T directly.

    So when households get the RE/no-Ponzi constraint in the models, and then that’s used to work back through and get G – T, these models cannot be used. Because they imply G – T is subject to the IGBC.

    I think people have been trying to say this but haven’t been able to do so because it wasn’t clear the no-Ponzi fallacy is such a keystone piece or how this fit into the rest of the world.

    Or more likely, they did say this and I didn’t understand them. 🙂

  8. JKH
    August 31, 2011 at 8:19 am

    Demand for (G – T) has both liquidity and capital (equity) elements.

    (G – T) generates (normally) risk free, liquid assets – bank reserve deposits with the CB, currency, and treasuries.

    Perhaps we should classify all of these together more simply as reserves – held by banks and non-banks in aggregate.

    (One of the MMT themes is that treasuries are essentially as liquid as the others)

    Thus, demand for (G – T) assets is liquidity driven.

    But there’s also a capital (equity) element.

    (G – T) generates income.

    Which for the most part becomes capital (equity) for non-banks (mostly, in the sense of the broad distribution of the income effect).

    So (G – T) generates liquid assets and creates non government capital.

    Non-banks (including households), like banks, are capital constrained in spending just as banks are capital constrained in lending.

    Although banks hold reserve deposits and reserve currency for settlement reasons, both banks and non-banks hold (G – T) assets more broadly as a liquid reserve store of value.


    Nick seems overly eager to assume that accounting awareness obstructs economic thinking.

    There’s no reason to make that assumption. It is an unnecessary bias against accounting.

    Accounting awareness correctly frames economic thinking.

    As in “ready, aim, fire”.

    Don’t fire the economics before you aim the accounting.

    Speaking of “aim”, Ricardian paranoia capitalizes the negativity of future taxes for no good reason.

    Why not capitalize future growth and demand for liquid, risk free assets instead?

    • JKH
      August 31, 2011 at 8:37 am

      Perhaps the problem with monetarism is its relative neglect of the liability/equity side of the balance sheet.

      As in “money doesn’t spend itself”.

      (Or, money doesn’t lend itself, in the case of banks)

  9. August 31, 2011 at 1:20 pm

    But gold sells itself!

  1. September 30, 2011 at 8:17 am
Comments are closed.
%d bloggers like this: