The Economy almost certainly needs Bubbles.

Larry Summers shocked everyone recently with his speculations on the links between growth and bubbles. some people slightly agree, others do not. For example, Clive Crook disagrees with the premise we need bubbles for the economy to work properly.

I have been thinking about the link between bubbles and growth for nearly 3 years. Back in my Traders Crucible days, Nick Rowe wrote a post on bubbles which really caught my attention. He pointed out an old paper by Paul Samuelson gave some insight into the reason the (completely wrong) Intertemporal Government Constraint might not hold.

“There is something economists have known since 1958 that we don’t talk about much, except in private, like in economics journals that nobody else reads. It’s a bit too weird.

There are two sorts of world. In a normal world, the equilibrium rate of interest is above the growth rate of the economy. In a weird world, the equilibrium rate of interest is below the growth rate of the economy. What’s weird about a weird world is that it needs a bubble, a Ponzi scheme, a chain-letter swindle, for the economy to work well.

Maybe the world we live in is a weird world. And it needs a bubble to work well. And the economy keeps trying to create a bubble. And when it succeeds the economy does work well. But bubbles are unstable and eventually burst. Then it works badly again. Until the next bubble.”

I have been publicly silent on this idea for a while, but this does not mean I’ve stopped thinking about the relationships between g (the real rate of growth) and r (the rate on interest on publicly held government debt) and another r (the natural rate of interest).

I do not believe in an economy wide natural rate of interest. It does not seem to be observable in real time, or even easy to determine it’s general level years after the fact.

(Aside on the natural rate of interest: Here are more problems with the NRoI. The inputs to determine the “natural rate” are hard to measure. The NRoI suffers from ex-ante, ex-post problems. The NRoI doesn’t meet basic criteria for effective real time guidance. The economy is ergodic, so what use is a NRoI? It also seems mis-specified in that it targets an low quality economic equilibrium of stable prices. The Investment/Savings aggregation problems. Aggregating a yield curve into a single interest rate is nonsense. NRoI implies lending, which implies borrowers and lenders, and borrowers and lenders are not uniform at all. The NRoI highly dependent on institutional structure. I have some writings on this which will come out over the next few weeks. There are more…)

But for now, just forget the problems with the NRoI. Take the NRoI model at face value, like Larry Summers does. What does this mean in the context of the Paul Samuelson world? Do we live in this weird world, where g is higher than r (the natural rate of interest)? More specifically, do individual economic sectors (such as real estate) live in this weird world?

Yes we do. We currently live in a world where growth is higher than the natural rate of interest. Everyone who thinks there is a natural rate of interest believes the NRoI is negative right now. Miles Kimball, Brad Delong, Paul Krugman – nearly any economist who abides by the NRoI model believes the NRoI is negative.

The economy needs a bubble in order to function properly when g > r.

They also believe g, our real rate of growth, is positive today and has been positive over the last several years. The rate g might be low, but it’s well above zero, probably somewhere around 2-3%. Technology is advancing at some non-zero rate, which *must* force growth to be higher than zero. In any case, real GDP has been growing.

These circumstances put the U.S. economy into Paul Samuelson’s world where we can get a free lunch, as long as we have a bubble. We get extra growth at no cost of inflation. Our lives are better off with the bubble.

My take on this is the economy demands bubbles when r <g. That’s right – demands bubbles. If the bubble is not met, it will try and find a way to create this bubble. (See real estate, stock market, S&L, Emerging Markets) Artificially constraining the bubble forces unnecessary misery on people, and causes involuntary unemployment and unused capacity, which is bad because it causes an economic incentive for war.

How large does our “bubble” need to be for the economy to function properly? This should be an empirical question.

What does a “bubble” mean? I have a post on this, but bubbles and money are closely related. Bubbles can only happen within the structure of a money creation process. Nick talks about this in his post:

“Samuelson called the chain letter “money”. But it could be an unfunded government pension plan. Or gold could serve the same purpose, even if gold were intrinsically useless except as a store of value. Whatever it is, it’s a bubble, that has a market value in excess of any intrinsic value.

But if the marginal return on capital, after an allowance for risk, is less than the growth rate of the economy, it still needs a bubble. A chain letter, or bubble, can grow forever at the growth rate of the economy, as long as people believe in it and don’t break the chain. And it can pay a rate of interest equal to the growth rate. So people save partly in real capital, and partly in the bubble, and both pay the same risk-adjusted rate of return, equal to the growth rate. (The marginal rate of return on capital rises as the capital stock falls when people hold less capital and more bubble.)

Land may also alleviate the problem, in the same way that capital does. But land is different from capital because they aren’t making it any more. That means that land may also serve as the bubble asset, with the price of land made up of a fundamental component equal to the present value of expected rents, plus a bubble component, with the bubble component growing at the growth rate of the economy and yielding capital gains. And house price bubbles are really just bubbles in the price of the land on which the house sits, because the house itself is just reproducible capital.

Governments can create the bubble, with a national debt they never pay the interest on, but just rollover from one year to the next. A Ponzi scheme. If spending and tax revenues grow at the growth rate of the economy, and that growth rate is above the rate of interest, they can do this forever. The long run government budget constraint just doesn’t add up, literally.”

This isn’t an easy topic at all. This gets right to the heart of our economic system. It’s easy to make mistakes, and I don’t claim to have anything close to an understanding of how this works. Remember, I do not think the NRoI is something that exists economy wide, but is something which is rather sector/geographically specific.

Yet, with the discussion kicked off by Larry Summers.

More Points:

  • r is impacted by our institutional structure. We can change this with our monetary, fiscal, credit, and foreign sector setup. We could live in bubble land forever if we wanted.
  • g isn’t entirely outside our control either…
  • if r and g are sector specific, then some sectors will demand bubble or anti-bubbles (where the sector wants to destroy money). This should be extremely confusing to the “economy”.
  • While r is impossible to sum
  • This adds some color to the observation deflation incentivizes war. Deflation/Depression attempts to force g down to r, rather than bringing r up to g. People don’t like this
  • It seems like we’ve lived in bubble land since we abandoned Bretton Woods and Functional Finance in 1973. If we take the estimates of the NRoI at face value, there is a decent chance we are in bubble land for much of the time.
  • If the NRoI is negative today, then the demand for a bubble must be very large today. The spread between NRoI and real growth is very, very high – possibly as high as 5%
  • When r < g, we demand a bubble. If r < g for more than a few years, we need a bubble.

P.S. Businesses Hire when they are swamped with demand.


Update: Slackwire has been thumping around this same area. His post is a must-read in context of the comments by everyone below.

  • stone November 23, 2013 at 1:25 am

    Your make the crucial point, “r is impacted by our institutional structure. We can change this with our monetary, fiscal, credit, and foreign sector setup. We could live in bubble land forever if we wanted.”

    IMO, the need for bubbles comes because our financial system fails to reflect the reality that every unused hour of potential machine or worker time is permanently lost. Currently, we can preserve wealth over time using our financial system even if we are not using it productively- that is the key distortion that destroys our economy IMO.
    There would be no need for bubbles if a maintaining wealth over time required a strong stream of dividends in order to collect enough funds to pay an asset tax.
    Bubbles basically are a sort of asset tax in that they flush down the toilet a lot of hard earned savings. The problem is that they don’t do so evenly. Smart players actually make a killing from bubbles. The bubble world misdirects the “best and the brightest” away from developing new technology (or organizing a productive economy or whatever) to instead working for hedgefunds and investment banks creating and harvesting bubbles. The people who get fleeced by bubbles are often the people who have been concentrating on contributing to the real economy (or are just foolish I guess) and have had a dumb casual exposure to the bubble. I guess many haven’t even stopped to think that recovering from a 75% loss takes a 300% gain.

  • JKH November 23, 2013 at 1:31 am

    Interesting series of posts.

    Question on this one:

    I can see (maybe) why the economy induces bubbles in this circumstance.

    But what is really meant by saying the economy ‘needs’ bubbles?

    i.e. ‘needs’ to what end?

    • stone November 23, 2013 at 1:46 am

      I guess they mean we need the increased spending that comes from the associated malinvestment? -Jobs building ghost estates of unwanted houses, empty shopping malls, lawyers writing patents for daft pretend biotech bamboozlements or whatever?

    • Michael Sankowski November 23, 2013 at 9:30 am

      It seems like the individual sectors where g > r try to find ways to make money like structures so they can grow properly.

      I think about this as the profit incentive from g > r forcing people to pay attention. Basically, there is a economic gain possible when g > r. The endogenous nature of money wants to help this profit motive be filled.

      I didn’t get into the relationship with the other r mentioned in the piece, which is the rate of interest on government debt. But this r (call it r*) impacts how much profit can be had from g and NRoI.

      btw – some of this was directly inspired by your question on one of the last posts about the relationships in the late 1990s’. I am still working on that response, 2500+ words already.

      Think about this in relation to the late 1990′s internet bubble. In the early days of the internet bubble, people correctly estimated g >> r. Something people forget about the internet bubble is that it’s provided us with one of the great alltime inventions of mankind, and the potential for growth is still absolutely gigantic.

      When g > r, the sector is essentially begging for a money creation process based on it’s growth. Usually, these sectors are small relative to the economy. Then, the sectors don’t have a way to take their equity and turn it into money. The internet bubble was both potentially large relative to the economy and had a way to make something very money-like.

      Something else that strikes me about this is your observation there are huge amounts of assets sitting on balance sheets in the world, which are just there and don’t get revalued frequently. The internet boom created a bunch of new assets in the world almost overnight, in any case extremely fast compared to even very rapid asset growth.

      Then, what if you take Mosler’s assertion the NRoI is zero? He means something different in his vision of the NRoI, but maybe he doesn’t either. The NRoI is a strange idea where somehow things balance out without causing changes in price, but in a wide economy with lots of chaos and many goods and services and geographic seperation, how does this equilibrium signal across all of this information without being disrupted by the interactions around it?

      If the NRoI is zeroish, and g is positive, we always live in the bubble demand world due to technology and productivity gains. Technology and productivity push g to be positive.

  • stone November 23, 2013 at 1:59 am

    Does this all depend on “capital” actually suffering a net loss from bubbles such that they cause a trickle down?
    I wonder whether as more “capital” concentrates with sophisticated speculators and as the TBTF culture causes losses to be socialized, “capital” overall might actually TAKE more from the economy by harvesting bubbles than gets tricked down.
    It doesn’t even take much sophistication does it? Just cash to spare? If someone kept say 75% cash and 25% or some bubbling asset and rebalanced periodically as that asset rose to the stratosphere and then fell all the way back down to the start, then they would have profited wouldn’t they? The counterparty that was paying for their gains could be the real economy. In that way bubbles could actually end up dragging down the real economy -quite the opposite of what it is said that they are needed for.

  • Greg November 23, 2013 at 5:53 am

    Sounds like we have too many capitalists who are saying, “Unless we can win the lottery with this investment we aint investing” Which actually seems quite a rational strategy when you have lots of wealth to protect. Its been shown time and time again that moderate losses affect us more negatively than consistent small gains affect us positively (from a psychology standpoint) But this becomes self reinforcing. Everyone with wealth refuses to invest so incomes suffer and the people with wealth find that their businesses start suffering from lack of demand.

    To add to Stones point;
    ” In that way bubbles could actually end up dragging down the real economy -quite the opposite of what it is said that they are needed for.”

    I agree, its kind of like a team with Michael Jordan on it sitting back and waiting for Michael to win it. When Michaels the only one playing basketball its hard for his team to win. The housing bubble at least involved lots of sectors of the real economy, many of which did back breaking work (real production) but it was all the financialization of the mortgages and all the side bets on the default risks which made it unstable and eventually led to its demise. Stock market bubbles or any other financial bubble will always end badly in my view. We dont need any more tech stock or any other financial product bubbles

  • Ramanan November 23, 2013 at 9:27 am


    Unsure as to what you mean by “need” bubbles?

    • Michael Sankowski November 23, 2013 at 12:03 pm

      I can understand that uncertainty, because I am not entirely certain myself.

      First, just take the NRoI model at face value – which is of course very problematic. But please just think about this from Larry Summers perspective where he accepts the existence of the NRoI.

      In that model, when g > r, then the economy has a gap between the capacity generated and the capacity used. Over time, this gap can become very large, and the capacity generated need not be realized in actual buildings and capacity.

      Imagine a spaceship crashlands on a completely earth like planet. Most of the ship is destroyed except the codex which contains all the practical knowledge of the civilization, and a few rough tools like hammers.

      The GDP of this civilization would grow very rapidly from roughly zero. If they abided by a monetary economy and enforced 5% NGDPLT, then monetary hording would take place, probably so much as to slow the overall growth of the economy to something far less than if they allowed bubbles in the money supply.

      This is essentially what happens for countries like China when they allow their current account to be very positive. They have the codex, and rough tools, but can’t afford to internally finance the growth, and enough hording is taking place the growth will never happen. When they allow/force the current account to be out of whack, they are allowing the naturally demanded bubble to happen to take advantage of the growth potential inherent in g > r. This could not be fully realized without the bubble happening.

      Also – please don’t be too harsh. :) This isn’t exactly a well explored area of the map.

      How does this fit into a PK or stock-flow model? NRoI doesn’t exist in the same form, but PK does recognize interest rates are an important determiner of what happens in the economy. The other r, the real rate of interest on government debt, interacts with the entire credit surface (the yield curves for all assets). When g for an asset is greater than its corresponding market determined r for that specific situation, there should exist profit opportunity for entrepreneurs. The larger the spread, the larger the potential profits.

      If money is created endogenously, money (which has bubblish qualities as you know better than me) can be created to fill that gap. But there are other factors which might either exacerbate the want to exploit the possible profits or restrict that need. The raw gap doesn’t go away in either case.

      • Roger Sparks November 23, 2013 at 12:51 pm

        “This is essentially what happens for countries like China when they allow their current account to be very positive…………”

        Here is a spin that is a little different:
        The Chinese workers are part of the American economy to extent that they produce product to be consumed in America. Yes, they live and work in China, but they would not have jobs except for the American purchase.

        Now, while working for the American economy, they still live in China and all transactions for the workers are in Chinese Currency. The required conversion between Chinese and American currency (to make this all happen) is dependent upon government approval (and assistance if trade is unbalanced).

        Has an American caused bubble been generated? I would say ‘yes’.

  • Ramanan November 23, 2013 at 9:35 am

    I love Summers’ confusion and also Krugman – who seems to have supported Summers.

    These guys haven’t even understood what Wynne Godley was writing in 1999:

    • Michael Sankowski November 23, 2013 at 12:14 pm

      Thanks for the paper. I need to go through this in detail. I’ve been writing something to JKH which hits on nearly all of the same topics.

  • Frances Coppola November 23, 2013 at 9:47 am

    The question is though, WHY are we in this weird world where r < g and therefore the real rate of return on investment is negative? I wonder if the reason is that there are not enough activities that generate financial reward – hence rising structural unemployment and falling median wages. Or that too much of the return from growth goes to too few people – which is sort of the same thing, really.

    Worth noting, Mike, that the "step change" at the end of Bretton Woods was also the end of low stable oil prices. For me the appalling inflation in oil prices ever since has far more to do with this. I think Bretton Woods collapse was a symptom of global trends (of which one was the end of US control of oil prices), not a cause.

    I don't buy the "bubbles are necessary" argument. Bubbles are created by wealth holders desperately seeking to avoid capital erosion (on safe assets) or find some kind of return that they regard as sufficient to justify taking risk. They aren't "essential", except in so far as wealth holders really don't like losing their assets.

    • Michael Sankowski November 23, 2013 at 12:12 pm

      Agree very much on the end of BW and the resulting oil/commodity swings. I think that was part of the point of ending it, in fact, at least for the rulers.

      ” I wonder if the reason is that there are not enough activities that generate financial reward – hence rising structural unemployment and falling median wages.”

      The institutional structure we swim in can prevent activites from taking place, or from there being enough money to allow those activities to happen. Steve Waldman had some good insights along this line where he observed profit opportunity is purchasing power related. Capitalists and Entrepreneurs want to go where they can make money, so they gravitate to sectors where people already have money to spend. Nobody tries to make money on poor people because even if the demand is high because they don’t have enough purchasing power to justify the risk.

      I’ve been thinking about the bubble process for a while now. Seems to me as though true bubbles need some sort of money creation process to be true bubbles. Sheltering assets is part of the puzzle on the savings side, right?

    • Oilfield Trash November 23, 2013 at 12:53 pm

      “I don’t buy the “bubbles are necessary” argument. Bubbles are created by wealth holders desperately seeking to avoid capital erosion (on safe assets) or find some kind of return that they regard as sufficient to justify taking risk. They aren’t “essential”, except in so far as wealth holders really don’t like losing their assets.”
      I agree with this except I would add the Greenspan and Bernanke Puts using monetary policy to support asset markets and the financial sector when the real economy was not asking for monetary stimulus creates conditions such that eventually there are only two states possible for the economic system – a bubble or a collapse.

  • Roger Sparks November 23, 2013 at 10:00 am

    “………… g (the real rate of growth) …………..”

    Can I safely assume that “g” can be measured by the rate of change of GDP? (that is, GDP as included in Federal Reserve data.)

  • JKH November 23, 2013 at 1:24 pm


    A permanent zero rate sure is interesting to think about.

    (i.e. permanent by design rather than circumstance)

    Although I’ve always been surprised that Mosler wrote something called “the natural rate of interest is zero”.

    He among anybody should know that zero is a central bank choice – just like one or minus one.

    “It’s always about price” doesn’t experience a sudden discontinuity when the administered rate flips from positive to negative.

    (Perhaps we could get Scott Sumner to confirm the second half of that statement.)

    Moreover, zero isn’t even a natural number.

    A chip off the old paradigm.

    • Tom Hickey November 23, 2013 at 7:12 pm

      JKH: “He among anybody should know that zero is a central bank choice – just like one or minus one.”

      The point of the paper is that if the cb doesn’t set the rate by paying IOR or adjusting quantity of rb, e.g., by OMO, the rate falls to zero. Since deficits are the norm, the normal rate is zero toward which the rate naturally gravitates with excess reserves in excess of demand (which is why the Fed is paying IOR now).

      In this case the cb is choosing to set the rate to zero by letting excess reserves take it there — I was told this by one of the MMT economists under circumstances that I cannot recall. I think I had stated a position incorrectly and was corrected.

      Mathias Vernengo pointed out that this is the “normal rate” and wonders about the choice of “natural rate,” which has a different meaning anyway. He thinks the point of the paper is essentially correct but the choice of words unhelpful.

      On the natural rate of interest one more time

      • JKH November 23, 2013 at 7:37 pm

        I know what the point of the paper is supposed to be.

        My point is that zero is a rate of interest – its a central bank policy choice.

        • Tom Hickey November 23, 2013 at 8:11 pm

          You think Warren is disagreeing with that? That’s a central SCE and MMT position.

          • JKH November 23, 2013 at 8:31 pm

            No I don’t think he should disagree or does disagree with that.

            As I said, that’s why the paper he wrote surprised me. (I know he wrote it a long time ago.) It’s inconsistent with that basic position.

            The premise of the paper would seem to assume that the central bank is run by zombies who have no capacity to think about what rate of interest should be charged on the deposits it issues as a monopoly supplier of those deposits, and that no interest is paid as a result of that natural zombification of management. You might say the same thing about a commercial bank that gave no thought whatsoever to paying interest on its deposits.

            The fact of the matter is that central banks are active managers of short term interest rates and they will design reserve systems and reserve pricing within those systems to manage that process – whether that can be done with a zero interest rate on excess reserves (e.g. pre-2008 system) or with a non-zero interest rate (e.g. corridor and floor systems).

            There’s nothing natural about looking the other way in the sense of not making a decision on what interest rate to pay on reserves, or paying no interest by default as a zombie non-decision. Central banks decide what system of reserves they’ll provide for their commercial banking systems, and they’ll set the interest rate structure on reserves accordingly. That’s their job.

            You can’t assert zero is the natural rate if the job of central bankers is to set the short term interest rate that prevails in the markets. The reserve system must be congruent with that objective. It’s not natural if it requires management zombification to allow that zero “natural rate” to take hold. That’s inconsistent in a system where the reserve mechanism itself requires a process of intelligent creation in order to meet central bank objectives in the broader sense.


            I think a better argument might be something like this:

            Suppose there was no private banking system, and that banking consisted of a single central bank that offered deposits similar to those offered by commercial banking now. There are obviously no bank reserves in such a system. Then the question would be – what rate of interest would you pay on those deposits and why? If you can come up with an argument that a “natural rate” of interest of zero should be paid on those deposits, that would be interesting to hear. Because in that case, maybe there’s a follow on argument that the central bank should set its policy rate and the interest rate on reserves at zero in the system that we actually have. But that’s a completely different chain of argument than what is presented in the paper. IMO, the argument shouldn’t be based on the behavior of the interest rate just because it comes crashing down to zero as a result of excess reserve creation and when a zombie central banker forgets he has a job to do in deciding what rate of interest should be paid on reserves in the system he is responsible for.

            • Tom Hickey November 23, 2013 at 11:53 pm

              True. If there is confusion possible, this is not made as clear as it should be. I also agree with Matias that “normal” would have been a better choice than “natural,” which is too clever a play on the accustomed meaning of “natural rate of interest.”

              It seems that the fact that the cb sets the policy rate is assumed. I stated somewhere based on my understanding then that if the cb doesn’t pay IOR or target a rate by adjusting quantity of rb, then the rate falls to zero. I got a correction is that this is how the cb sets the rate to zero. Whatever happens, the cb is setting the rate, even if by default — even if those in charge are zombies — because the cb sets the policy rate whether it realizes it or not. So it seems that the presumption is that under normal conditions the cb sets the rate to zero by not acting to set a higher rate, either paying IOR or adjusting rb.

            • JKH November 24, 2013 at 2:42 am


              The natural rate as “defined” by Mosler and any natural rate otherwise found in the literature are two entirely separate things with two entirely separate explanations. Duplicating terminology like this is not helpful, even if the mainstream idea is highly questionable.

              So there are two separate arguments for at least two separate ideas.

              My point is that Mosler’s natural rate concept is internally inconsistent even from an MMT perspective, in the following way;

              - The CB sets the short term risk free rate? – check
              - The CB has all the tools it needs to set the short term risk free rate? – check
              - Those tools include restricting reserves through bond issuance, so that the risk free rate falls out as a function of the CB’s desired rate setting + an inelastic demand curve at the level of reserves supplied (i.e. pre-2008)? – check
              - And those tools include payments of interest on reserves if that is desired? – check
              - The CB therefore has a choice as to whether the short term risk free rate will be zero? – check
              - Is there anything natural therefore about letting the risk free rate fall to zero because of a sloppy excess reserve setting and the CB abandoning its purpose and responsibilies?


              That’s the contradiction with calling zero the natural rate

              I suspect Mosler and MMT have better arguments than that for setting interest rates at zero, although there would obviously be a big debate about that. But the argument for it should not start with the bank reserve system. It should start from first principles as to why the government might want to set that rate as policy. My personal view is that from a policy perspective it’s comparable to a gold standard in the sense that it is a net loss in policy flexibility.

              There’s lots of literature on the other mainstream concept. I think of it as the theory of a virtual Ouija board that sits in the middle of the FOMC board table, and instructs them on where to set the Fed funds target. I gather there are other theories as well.

              Two very different things. That they are very different shouldn’t be a source of confusion. Nor need the MMT view be a source of confusion (a bit of housecleaning preferred). I think the mainstream view could be very confusing though – Ouija boards are like that.

              • Tom Hickey November 24, 2013 at 10:06 am

                Thanks for clarifying your view.

              • Cullen Roche November 24, 2013 at 11:04 pm

                Just one of many definitions in MMT that are changed. Full employment, net saving, “money” (which includes T-bonds – WTF?), etc.

                What MMT consistently fails to discuss is the history behind the reserve system and why it even exists. Before the Fed the nation used a system of private clearinghouses that acted a lot like the regional Fed banks do where payments could be centrally cleared. The Fed system was formed after the panic of 1907 and modeled specifically after the NY Clearinghouse. The Fed is basically a big clearinghouse designed around private bank money to help support it. Yet MMT acts like the reserve system is some sort of system that usurps bank money and effectively nationalizes it rendering banks as users of state money. The whole thing was designed AROUND bank money to support private banking. Yet MMT acts like the circuit of money creation starts with the Fed/Tsy because of the creation of this clearinghouse. It’s a misrepresentation of the entire history of the system and money. And then they add on the concept of banks “leveraging” reserves which is basically just a more sophisticated version of the money multiplier. In the end MMTers make the same mistake all neoclassicals do by focusing excessively on reserve money and building a model that places it squarely in the middle. And that’s the core source of Mosler’s thinking about the “natural rate”.

                The more I get to like Post-Keynesian economics the more I wonder why we need these big distractions from MMT. Why do we need to change all these definitions, alter the history of money and pretend there’s some sort of new “paradigm” with MMT? There’s literally nothing MMT can explain that I can’t explain using consistent definitions and basic understandings of the monetary system without having to pretend there’s this whole new paradigm. And in fact, a lot of the explanations and causes of things seem muddied by this thinking so I don’t see how this is helping.

                • stone November 25, 2013 at 1:06 am

                  Cullen, isn’t it true though that the Fed is just one of many central banks worldwide and they have each moved to much the same final set up from quite different starting points? MMT is not just about the USA. Bill Mitchel is Australian. I thought that the Bank of England displaced a system based on Tally Sticks which was pretty much like your platinum coin system. It was all about seigniorage of tax token money (in the form of Tally Sticks). The Bank of England came about because the English parliament gained the power to obstruct the setting on new taxes by the King. That caused a break down of the Tally Stick system with Tally Sticks trading at wide discounts
                  Also of course, the Fed was not the first US central bank. The Fed came in after the US had a time without a central bank but it had previously had a central bank.

                  • Cullen Roche November 25, 2013 at 1:19 am

                    Well, that gets into an even more detailed discussion. I am referring to the USA, but yes, I would agree that it very much depends on the central bank and monetary system you’re discussing. When you’re understanding the monetary system in different systems you must account for the political arrangement that is created around that system as well. In the USA we have a specific monetary design where the banks are the dominant money issuers and the central bank acts mainly as a facilitating entity around it. This is totally in keeping with the Constitution and the way our govt is specifically designed to disperse power. In this case, the Executive Branch is specifically firewalled off by the Central Bank from having a monopoly control of money. But in the MMT model they just go and ignore this little inconvenience, consolidated the two entities and give the Executive Branch a money monopoly by consolidating the Fed into the Tsy. Nevermind that “money is a creature of law” and the laws are specifically designed to ensure that there is no money monopoly. How can anyone just ignore these things and claim to be describing the operational realities of a specific system?

                    Of course, if you went to China and tried to make the same argument it would be foolish. The systems are totally different. And yes, I am aware that Randy says it’s all the same just because he says the accounting is the same (which is a very silly way to abuse the use of accounting). MMT is more true in a system like China than it is in the USA. So yes, it depends.

                • winterspeak November 30, 2013 at 6:09 pm

                  Sorry Cullen, this is one of those areas where Warren is much clearer, I think, than I’ve seen you being.

                  Warrens point on setting the interest rate to “zero” — whether you want to call it “natural” or “normal” or “optimal” or whatever, is that the reserves are primarily used to settle transactions between banks in a system where individual depositors are not expected to take individual bank credit risk into account. FDIC (in part) establishes the latter, in which case the former follows as a point of logic. In such a system, the price an individual bank sets for credit extension should be a function of it’s own capital position, it’s own business model, and it’s own calculation of credit risk in the would-be borrower. That will be the “correct” interest rate, and not that there is no reason for the Fed rate to be anything other than zero.

                  The entire ladder of interest rates can be built of this permanent zero fed rate just as it has been since the US moved to ZIRP 6 years ago.

                  The Govt maintains a number of levers to manage economic activity overall that are more effective and have more robust mechanisms that the obviously flawed interest rate model — including capital regulation, other regulatory levers around credit (including forbearance), and of course, fiscal policy. Bank quality would be managed not by looking at individual repo rates and such for that institution (which have well known problems), but audits to the Fed’s own credit model (for better or worse).

                  It does not help your case to misrepresent Warren or pick straw men arguments. Even people you don’t like can have correct points sometimes, you know.

                  • Cullen Roche December 1, 2013 at 12:45 pm

                    WS, I find it continually odd how you so vehemently defend MMT’s positions yet you still don’t seem to understand that your rejection of the JG is rejection of their most important policy position as well as the core understanding that holds the entire theory together. If you don’t buy into the JG then you don’t buy into the notion that the govt is the price setter of money and the monopolist who causes the unemployment in the system where they introduced the currency. You’re defending a theory you don’t seem to fully accept. It’s very strange. Anyhow….

                    I am not sure I “misrepresented” anything. I simply stated a fact – MMT does not tend to focus on the history of how the Fed actually came into existence as an entity that serves as a clearinghouse to support private banking. Instead, MMT cuts corners and consolidates entities that, in reality, are distinctly different entities operating under very specific legal parameters. MMT “misrepresents” our entire reality by simply ignoring the actual institutional designs in place or brushing them off as “self imposed” or something. If anyone “misrepresents” anything then it’s MMT.

                    It doesn’t help the MMT cause to constantly accuse everyone, who points to potential errors in their views, as “misunderstanding” or building “strawmen”. Those two responses seem to be the primary way MMT responds when confronted with someone who actually understands their theory and the way they present it. Why not actually respond to the point I made rather than just calling it a “strawman”?

                    And I get Warren’s point entirely. I just don’t agree with the way MMT builds a world view around the reserve system as though it’s the top of some “hierarchy of money” within in a consolidated Fed/Tsy view when the reality is that the reserve system was created to facilitate and support a money system built around private bank money. Am I allowed to disagree with MMT without being attacked and accused of being dishonest?? Are we allowed to present the monetary system as it actually is and not be attacked by people who don’t like our views because it muddies their policy ideas? I am sorry to be so frank, but MMT’s entire “general case” is a misrepresentation of the way things actually are so when an MMTer accuses someone else of “misrepresenting” things time and time again, it starts to get a little old. The theory you defend so vehemently is based on a misrepresentation….Maybe you’re willing to overlook these details, but it’s not fair to accuse everyone who disagrees with MMT of “misrepresenting” things just because they’re getting into more detail than MMTers prefer to be shed on their views….

                    Anyhow, I’d prefer not to argue with you. You seem like a reasonable guy. And also, I rather like Warren so it would be a “misrepresentation” to claim I don’t like him. :-)

                    Happy holidays.

                    • winterspeak December 2, 2013 at 8:37 am

                      Hi Cullen

                      I don’t think there was anything vehement in my post. Historical antecedents can be interesting, but as we know, things change and not always in a logical way. The Fed we have now is very different from the Fed we had in 2007, and I don’t think anyone really knows just how temporary our current IOR regime is and if we will ever go back to the old system.

                      To address your specific point, to the extent I understood it, I think the Executive/Fed/Treasury, while distinct, have strong areas of influence and overlap with each other. It’s a short cut to consolidate them, but not an unreasonable one under some circumstances.

                      It is an error, however, to think that current institutional arrangements are how they will always be. Congress has changed this in the past, has the power to do so now, and surely will in the future. And the direction of how these institutional arrangements have changed has gone from being a gold-standard system built around supporting regional banks to a fiat-money system built around a strong central bank with close ties and revolving door executive relationships to a handful of super banks, both investment and commercial, built around supporting securitized debt.

                      My understanding of Mosler’s reasoning behind the natural rate of interest being zero is that reserves, aside from their payment settlement function, just aren’t that important (contra your assertion that MMT builds its theory around reserves). Certainly the function they are normally thought to have — a mechanism to set interest rates which then manage economic activity/inflation — doesn’t work and has no real mechanism by which it can work.

                      So to take up JKH’s point, setting FFR to zero and leaving it there may take a policy option off the table, but it’s a ineffective policy options so there is no real loss. And to the extent this removes a distraction from meaningful policy levers — particularly bank regulation and fiscal policy — it may be a real gain.

                      Lots of counter arguments to this, but I think the main policy thrust needs to be something like we want the Govt to control real estate pricing through interest rate channels (which is the primary mechanism we see). That’s not an easy argument to make, and I think would be very difficult if we were beginning from a excess reserve permanent ZIRP policy position.

                    • Cullen Roche December 2, 2013 at 11:13 am

                      Hi WS,

                      I agree with you there. It’s important to understand that the system is changing and will change, but we should get ahead of ourselves here. Joan Robinson famously criticized neoclassical economists for their ideas that the system had evolved into something that it wasn’t just yet. I fear MMTers do the same thing in many cases by consolidating the govt and thinking about the govt as a self funding entity. I think that’s why me and the MR guys are so critical of what appear like some sloppy claims about how things actually are today.

                      Anyhow, the criticism seems to be forcing MMT economists to think hard about these matters which is a good thing.


  • Nick Rowe November 23, 2013 at 1:38 pm

    Mike: Thanks!

    Yep, the Larry Summers stuff is very much about this. Samuelson 58 and all that.

    “This isn’t an easy topic at all.”

    Yep. It’s damned hard, conceptually.

    Empirically, we are talking about an interest rate from now to infinity. Ultra long. A consul yield. And the growth rate from now to infinity.

    I have revised my views a little, since that old post. If land exists (and it does) that can make a big difference to the Samuelson 1958 model. Here is my more recent post in the same genre:

    • JP Koning November 23, 2013 at 7:20 pm

      Just to make sure I get this. The “growth rate” g refers to the rate of population growth, right? Looking through the old Samuelson paper…

      …he seems to be talking about population growth.

      • Nick Rowe November 23, 2013 at 8:24 pm

        Yes. But in Samuelson’s model, there is zero productivity growth, so it’s the same thing. If we generalise, we need to redefine g as GDP growth, not population growth.

  • Nick Rowe November 23, 2013 at 1:54 pm

    I skimmed the Godley paper. I think it’s totally off-topic. He doesn’t seem to be aware of Samuelson 58. Same for Warren Mosler. You need to have at least some vague idea of overlapping generations models (like Samuelson 1958) before you can even get started on this. If people lived forever, this problem wouldn’t arise. Nobody would save for their old age, if they never got old.

    • stone November 23, 2013 at 2:19 pm

      Nick Rowe, ” If people lived forever, this problem wouldn’t arise. Nobody would save for their old age, if they never got old.”

      Nick Edmonds has just done a post about that:

      Some people people do save for reasons other than that though. Steve R. Waldman has pointed out this this study:
      Why Do the Rich Save So Much?
      Christopher D. Carroll
      “The paper concludes that the simplest model that explains the relevant facts is one in which either consumers regard the accumulation of wealth as an end in itself, or unspent wealth yields a flow of services (such as power or social status) which have the same practical effect on behavior as if wealth were intrinsically desirable.”

      I’m not sure that saving is even typically a conscious decision. Sometimes the default -go with the flow- result once someone has a great fortune, is for that wealth to gather more -pretty much by itself.

    • Ramanan November 23, 2013 at 2:21 pm

      Well I posted that because Summers and Krugman seems clueless still!

      About your Samuelson and rg … it is sideshow to the issue. One can have r>g and yet have fiscal expansion to sustain aggregate demand. There is nothing Ponzi about r>g and in fact was shown by Wynne Godley and Marc Lavoie.

      Of course not all processes are sustainable and one has to be careful. The r><g is a sideshow.

      • Michael Sankowski November 23, 2013 at 3:07 pm

        Money from nearly any source is Ponzi-like in many attributes. The fiscal expansion is a “bubble” in that there is nothing behind the money – just good intentions. ;)

        Nick -Godley is operating from a entirely different perspective than Samuelson. Not slightly different, but fundamentally different. People would save for other reasons, like planning for uncertainty. This results in a similar outcome where a bubble improves the economy.

        But the processes by which this “extra” is created can be unstable. I don’t see what the problem is with these observations.

        Here is something on real estate and money.

        • Ramanan November 23, 2013 at 3:11 pm

          I do not believe that the overlapping generations model is the key to the theory of money – James Tobin

          • Michael Sankowski November 23, 2013 at 4:01 pm

            To be clear – I don’t either. Just saying money *can* fill the gap implied by OLG because it has the bubble-ish properties which are required to fill up the gap left by g > r.

            Also, we should bring Scott Fullwilers paper on Fiscal Sustainability into this as well, because he pointed out it even within the NRoI model, as long as g > r* (real rate of interest on govt debt) there was no need to balance the budget.

            I was just pointing out that within the world Krugman thinks is the real world, the model implies bubbles are necessary when g > r as is the case now.

          • Nick Rowe November 23, 2013 at 8:29 pm

            Of course it isn’t. Money does not change hands once per generation. (Though it might neverthelessbe useful, understood very very metaphorically.) But as a theory of lifetime savings, that’s a different story. Most of our saving is for retirement, unless you believe in a Ricardian world of intergenerational bequests.

      • Nick Rowe November 23, 2013 at 8:25 pm

        If r > g, Ponzi schemes are unsustainable. When r < g, Ponzi schemes are sustainable. That's what samuelson 58 is all about.

        • Michael Sankowski November 23, 2013 at 10:48 pm

          When r> g, anti-ponzi systems make sense. It is not that Ponzi schemes are unsustainable. It is that anti-Ponzi schemes are sustainable.

        • Nick Edmonds November 24, 2013 at 3:27 am

          Is a sustainable Ponzi scheme really a Ponzi scheme at all? Same issue for the term “bubble”. If we’re talking about something that is sustainable, even if it is only based belief about the future, is that really what we mean by a bubble?

          • Tom Hickey November 24, 2013 at 11:01 am

            Greenspan and Bernanke’s excuse for not intervening in the greater fool trade?

            • Nick Edmonds November 24, 2013 at 11:22 am

              Maybe. But there’s no excuse for mistaking an unsustainable process for a sustainable one.

              • Michael Sankowski November 25, 2013 at 3:20 pm

                I was just using the terms as given. You are entirely correct – it’s not a Ponzi because it’s sustainable.

                Good catch.

    • Tom Hickey November 23, 2013 at 7:19 pm

      The purpose of a monetary production economy is accumulation of wealth, real and financial. People save not only for future consumption and provision against uncertainty, but to accumulate financial wealth as an end it itself, perhaps to pass along to heirs.

      “In a monetary economy, production occurs not to satisfy ‘needs’;, but to satisfy the desire to accumulate wealth in money form. Production is not undertaken by a Robinson Crusoe type agent who is both a producer and consumer; instead, there are those who own private property, and those who do not — and so must work for wages. However, the existence of propertyless workers extends market demand, and extends the use of money as a medium of exchange. Unlike production in, say, a tribal society, capitalist production always involves money. The capitalist must hire workers to produce the goods that will be sold on markets. As production takes time, the capitalist must pay wages now, before sales receipts are realized. Furthermore, because the future is uncertain, sales receipts are uncertain. This means that interest must be paid on liabilities and that capitalist production is only undertaken on the expectation of making profits. Thus, capitalist production always involves ‘money now, for more money later’. Since money contracts always include interest, and because contracts always are of the nature of money now for more money later, this means that monetary contracts will always grow over time at a rate determined in part by the rate of interest…. This generates a logic of accumulation: all monetary economies must grow. If they do not, accumulation falters and nominal contracts cannot be met. The logic of monetary production, then, requires nominal economic growth. It cannot be constrained by a fixed money supply, nor by a commodity money whose quantity expands only upon new discoveries.” (Wray 1999: 180).

      Wray, L. R. 1999. “The Development and Reform of the Modern International Monetary System,” in J. Deprez and J. T. Harvey (eds), Foundations of International Economics: Post-Keynesian Perspectives, Routledge, London and New York, p. 171–199

      cited by Lord Keynes, The Monetary Production Economy and Fiduciary Media

      • stone November 24, 2013 at 2:02 am

        Tom Hickey, I think there might be a mistake in that logic about the need for expansion. That is only true if the creditors permanently save the interest and use it to make more loans. If instead they draw upon savings to live as rentiers, then it can be a stable system even with no expansion. A grim clear example are people living as debt peons (eg some Bangladeshi construction workers in Dubai today). The debt peon has constant debt servicing charges that are balanced with their wages (paid by the creditor) such that the debt burden neither expands nor gets paid off. What the creditor receives is a lifetime of labour. The creditor does not get an exponentially increasing portfolio of loans owed by the debt peon. In a nicer example we can all be debt peons whilst of working age and rentiers in retirement and overlapping generation effects similarly allow a stable state with no requirement for expansion. Alternatively everyone could simultaneously be a debtor and a creditor as a way to spread about enterprise risk.
        I wrote a post trying to get my head around this:

        • Tom Hickey November 24, 2013 at 9:53 am

          How does that effect rb?

          The issue is excess rb. Existing rb just shift from one bank to another on the cbs spreadsheet and there are still excess reserves in the payments system that will drive the rate to zero unless the cb intervenes with IOR or by adjusting quantity of rb. Conceivably, the cb could also increase the reserve requirement to reduce excess rb, too.

  • Ramanan November 23, 2013 at 4:43 pm

    Another one by Tobin:

    “Although setups of this kind (the Samuelson [34] overlapping generation model, for example) are promising and already generate instructive parables, they are still so abstract and arbitrary as to be useless for policy analysis and econometric model building.”

    Appeal to authority ;-)

    • Nick Edmonds November 24, 2013 at 3:35 am

      I’d agree that these models present some formidable problems in applications to real data. But I’d say it’s worth persevering with, because I think that they hold the key to a lot of important things that are going on. My own interest in OLG models, for example, came out of trying to look at the impact on generational wealth distribution from the rapid increase in real estate prices in the UK and how that affected household saving.

      And what’s the alternative? We shouldn’t just stick to models because they’re easy to work with, if we suspect they may be giving a biased view.

      • Ramanan November 24, 2013 at 4:18 am


        And for Nick Rowe: “If r > g, Ponzi schemes are unsustainable. When r g ie the rate of interest on public debt is higher than the growth rate and yet the public debt/gdp doesn’t explode and sustains.

        • Ramanan November 24, 2013 at 4:21 am

          Damn whenever I include the greater than/less than sign, a part of the comment disappears and in this case has mixed up badly.

          What I wrote was I can write a model where the rate of interest on public debt is higher than the growth rate of output and without the public debt/gdp exploding ie it is sustainable.

          And I can write a model in which the assumed rate of interest is lower than the growth rate and yet have debts exploding.

          • Nick Rowe November 25, 2013 at 9:25 am

            Ramanan: sure. I can do that too. But that’s not the question that matters here. That question is: if the government borrows an extra $100 (or whatever), and borrows more to pay the interest on that debt, and never raises taxes or cuts spending, but keeps on rolling it over, does that debt/GDP ratio explode or not?

            • Ramanan November 25, 2013 at 10:20 am

              A fiscal expansion may lead to a higher deficit but as it starts acting, it raises economic activity to bring in more taxes without necessarily changing tax rates.

              (of course not advocating not raising taxes because I think it matters on questions of distribution).

              Such an action can be met with constraints due to inflation and the external sector in which case the powers would be limited but the standard story told is that fiscal policy is impotent except perhaps conceding that it is effective in the short run.

              Let me ask it the other way round – take a closed economy: are you saying that a fiscal expansion will necessarily make the public debt/gdp and interest payments/gdp explode?

              Take the standard debt sustainability analysis. It says that if the budget is in deficit then for g>r the public debt/gdp explodes. While true it is misleading. This is because higher activity will lead to increased tax flows putting the budget balance in surplus. People are paying more taxes – but guess what their incomes are also higher.

              The way you pose the question assumes that the $100 has permanently increased the debt but ignores that more taxes are coming in.

              • Oliver November 26, 2013 at 8:14 am

                Re: A fiscal expansion may lead to a higher deficit but as it starts acting, it raises economic activity to bring in more taxes without necessarily changing tax rates.

                Yes, finance is a cycle that can only be qualified if and when it is closed. Of course one wouldn’t see that if the model suggests an infinite time horizon and if one doesn’t take into account that something real is being financed, i.e. there is (should be) a real counterpart to every financial transaction.

                The nominal value of this real counterpart is determined precisely by the amount used to finance it. So, when govt. (or anyone else) spends $100 on something real (i.e. non-financial) and new, this raises ngdp by $100 and simultaneously determines that that which is being financed is thought to be worth $100 to consumers. Expected rgdp, one could say.

                This is true irrespective of r. A higher r just makes that expected rgdp harder to attain because it requires a much higher velocity for the money to find its way back to the initial debtor to be repaid.

                A bubble is when expected rgdp turns out to be smaller than effective rgdp upon repayment, i.e when someone takes a loss. And vice versa, so if there is such a thing as a bubble, there must also be such a thing as an anti bubble.

                The main implication is that a bubble can only be declared as such with 100% certainty after the fact. (Note: I’m not implying there aren’t any good indicators before the fact, such as large scale financing of non-productive and / or speculative activity. But, the closer finance is to production the less chance for a bubble there is, I’d say.) And for there to be a fact, there needs to be a day of reckoning. An agent that can roll his debt over forever doesn’t qualify. So I would say there are always potential bubbles in the making, but not all are equally likely to materialize. And materialization is a function of multiple factors that influence each other. r is only one of them.

                The only way I can explain the idea that a higher r leads to less bubbles, or indeed that there is some correct r vis g, is that its followers believe that less borowing also means better borrowing on average, so that the smaller numerator (borrowing) is more than made up for through higher denominator (quality of output). It’s a highly questionable assumption based on a general ‘no pain, no gain’ view of the world and it also completely neglects factors such as labour utilization that surely must be considered in a qualitative assessment.

  • jt26 November 24, 2013 at 1:38 pm

    Mike, let`s try some numbers to see if I understand. Assume we need 15 years of savings for retirement; retirement at 65; flat age distribution. Thus, we require 7.5 GDP years of savings.

    (Household net worth)/GDP~4 and (federal debt)/GDP~1, so we are short by 2.5 GDP years, so we need “bubbles”. Is that right?

    (Actually, it may not be that bad since retired people typically only spend 60% of the average, so we would need ~4.5 GDP years of savings.)

    • jt26 November 24, 2013 at 1:40 pm

      BTW I wonder if we can use this as an investment strategy. Calculate the bubble required by current demographic trends and compare to the actual. Buy when bubble is small. ;-}

    • jt26 November 27, 2013 at 4:09 pm
      • Michael Sankowski November 28, 2013 at 10:58 am

        Actually awesome, eh?

        • Michael Sankowski November 29, 2013 at 11:29 am

          I’ve been thinking about this very much over the last few days. I will probably end up using this chart in a post.

          • jt26 November 29, 2013 at 6:31 pm

            BTW maybe the direct link is more useful for you.
            I’ve also been thinking rather than using just GFDEBTN, that I should subtract some amount for assets held by households already counted in their net worth; I’ll have a look at the Z1. I think I’ll also take a closer look at how they calculate the net worth; maybe I can construct a real-time bubble indicator (via SPY,AGG,Case-Shiller), corrected quarterly using the Z1.

            • Michael Sankowski November 29, 2013 at 9:54 pm

              Ha! Already reconstructed the graph and saved to my Fred Account. It is a good start. This should probably be in percent gain terms, which could then be related to positive or negative stock market/asset market “bias”.

              There is something here…

  • JW Mason November 26, 2013 at 3:03 pm

    Great post. You are right, the concept of the “natural rate of interest” is one of the biggest obstacles to clear thinking about these questions.

    I am working on a new post about this now. but it’s important to relize that there is no connection between the yield on a debt contract (the thing we call an “interest rate” in daily life and the exchange rate between similar goods in different periods (the thing called an “interest rate” in a model like Samuelson’s.) They are distinct and basically unrelated concepts.

    The distinction between the “credit interest rate” and the “intertemporal interest rate” (my terms, not JMK’s) was a major component of the Keynesian revolution. But it seems to have been forgotten. It’s sadly ironic that in equating the “natural rate of interest” that will maintain full employment to the equilibrium intertemporal rate, Krugman is repeating exactly the position of Hayek, and opposing the position of Keynes.

    • Oliver November 27, 2013 at 12:57 am

      Interesting. I had always wondered about the relationship between the two.