« On the meme of the 'surprisingly short and mild recession' | Main | What makes charities special? »


Feed You can follow this conversation by subscribing to the comment feed for this post.


I beleive that MMT does not accept the crowding out effect. The private induced crowding out or constrained by leverage capacity (there is also tha pproach of the leverage cycle but that is for another comment beuond MMT) is not caused by discretionary monetary "stimulus" since MMT does not believe in the money multiplier concept. According to MMT it is banks/private units that expand credit and the CB validates the settlement function of deposits with appropriate reserves.

Question (maybe a naive one)

When the Fed Govt pays interest on its securities, does it deficit spend to do so?


"If you have a link to the right thread in billyblog I'd appreciate it."

I spent some time looking, and that blog is too big! Bill Mitchell is a profligate author, and I also posted too much there. I'll look again later. Maybe someone else here was following along and found it.

"I don't know what you mean by "RA" model, but I agree,"

In representative agent models, in order to close the model you need a "no ponzi" condition that controls the growth rate of debt. The typical choice is to say that the amount of debt is growing more slowly than the rate of interest, but this makes the model non-competitive.

We observe periods in which assets grow more slowly followed by periods in which they grow more quickly, but on average, the growth rate of assets should be the interest rate. In general, you would expect the capital markets to drive all assets to the same return, r, and that return would equal the growth rate of the asset, so that starting with A, next period, you have (1+r)A. Of course, you could sell some of your assets and consume the proceeds, but then the buyer gets the growth. In aggregate, rate of growth of an asset *is* the correct interest rate = expected return.

You can argue that these types of models by design make themselves unable to model debt growth in a modern economy or competitive capital markets, which is unfortunate.

Moreover, because these models are trying to add up the utility of an infinitely lived agent, there are all sorts of problems with infinities requiring assumptions of intertemporal discount factors and interest rates that are different from observed values.

This is similar to the problem of infinities observed in physics optimization problems -- say when you are trying to minimize the action of the path of a light particle, but the path is open-ended -- but the physicists solve it in a better way using renormalization.

I.e. the physicists allow experiment to determine the constants and create mathematical techniques that provide a sensible answer, rather than selecting the mathematical techniques first and then deciding what are "allowable" intertemporal discount or leverage factors.

"Still, the Treasury currently sets quantity to manipulate price (instead of setting price and letting quantity float) and that sounds like "market price" to me."

I don't believe that the treasury can set price and let quantity float.

That is another point of disagreement with the good folk at BillyBlog.

The whole concept of the monopolist controlling one of P/Q only works for consumption goods. As soon as goods can be resold -- e.g. bonds -- then even if someone is the monopoly producer of the good, they are not the monopoly supplier. In that case, you need to be both the marginal buyer and the marginal seller in order to control the price. It is not enough to be just one.

The fact that Mosler doesn't understand this and refers to the government issuing money like an electricity company selling electricity is mindboggling.

It is for this reason that nations that peg need forex reserves. If you believed in the P/Q view, then as the monopoly issuers of their own currency, they should not need to keep reserves.

That is why the government can easily control the interbank interest rates, because in the interbank market, they act as the marginal buyer -- by being able to drain in unlimited amounts until their target is hit -- and the marginal seller -- by offering to lend in unlimited amounts.

But they cannot control interest rates in general, unless they agree to *lend* to anyone at a specific rate as well as to borrow at a specific rate. It is not enough to only offer to sell a security at a specific rate.

Now you could argue that by offering to lend to banks, they are, by proxy, offering to lend to anyone else. That assumes that banks are competitive, but more importantly, it ignores the fact that banks generally do not fund businesses (except for commercial real estate). For the most part, banks lend solely against real estate. The bank *loan* market and the bond markets are separate markets, as the recovery rates for bond defaults (40%) are too excessive for levered banks, and bank leverage levels preclude them from funding capital in which the borrower adds a lot of value to the collateral. The business model of using leverage limits them to lending against collateral whose value is relatively unchanged regardless of what the borrower does -- land, and to a much lesser degree, short term funding secured by inventories.

So what you have is a situation in which investor (e.g. credit market) interest rates are separate from bank cost of funds, the latter reflecting real estate funding costs but not productive investment funding costs -- leading to all sorts of arbitrage situations.

But long term credit market rates are set endogenously -- by the return on invention of new forms of capital, and the intermediate term rates are set by the diminishing returns of capital, so *everything* is endogenous at the arbitrage-free level, and any intervention allows for financial arbitrage, primarily in the area of real estate and the banking sector.

Maybe that arbitrage is "good" for the economy, because the right kind of people/sectors are receiving the windfalls. But then just use fiscal policy to give people windfalls. Perhaps in a crisis when the credit market rates are irrational it's good to intervene. But there are colossal moral hazard issues, and I think it's no accident that the "modern" era of monetarism coincided with the financialization of the major industrial economies as well as secular real estate bubbles.

"When the Fed Govt pays interest on its securities, does it deficit spend to do so?"

I'm not sure what you mean by that question. "deficit" is the residual of aggregate money spent net of aggregate money drained. But I argue that the increase in net-worth of households as a result of deficit spending is equal to the primary deficit net of money leaked to the foreign sector. This assumes that the government is not going to default, so that when the households bought the bond, they were correct in valuing the future interest payments. When those payments arrive, the households will then be no worse or better off if they bought equity instead of the bond. And because of the leakages to the foreign sector, it could be that large amounts of deficit spending don't actually increase the net-worth of households as much as you expect, either.


this it?

"When the Fed Govt pays interest on its securities, does it deficit spend to do so?"

Not if its running a primary surplus equal to or greater than the interest.

RSJ: "I don't believe that the treasury can set price and let quantity float.
That is another point of disagreement with the good folk at BillyBlog."

There is the "durable goods monopolists" problem; the monopolist is competing with the used goods market, the size of which depends on how much the monopolist sold in the previous period. This can create a "time consistency problem" for the monopolist's ability to maximise profits. He might want to make a promise to not sell too much in future periods, but that promise might not be credible.

But a central bank that is not trying to maximise profits would see this as less of a problem. For any given P (or Q) today, the Q (or P) will depend on expected future P (or Q).

If demand falls, and the monopolist wants to hold P constant, despite the fall in demand, the monopolist would have to buy back some of the goods (reduce the stock of money in circulation).

Bottom line: I don't see a problem with the Billyblog folks on this. It sounds standard (and right) to me.

RSJ: "The fact that Mosler doesn't understand this and refers to the government issuing money like an electricity company selling electricity is mindboggling."

The only problem with the metaphor is that electricity is a flow, and money a stock. Switch it to the government renting out cars, and it works.

Fiscal deficit = primary deficit + interest on debt

I always think of the interest rate as being equivalent to the line voltage, with the amount of reserves being equivalent to the amps. An electric utility could in theory choose to keep either a constant voltage or a constant amperage, but in practice only voltage works, since the consequent spikes in voltage would destroy anything connected to the system. Volcker's monetarist experiment was similar - the attempt to control money supply led to some spectacular "transformer explosions"...

Nick: Yes. A government running an endogenous currency regime does not care about profit, and could set P (interest rate) in the good they were a monopoly supplier of (Treasuries) by standing ready to sell in the primary market, and being active in the secondary market. This doesn't seem so different to what happens during open market operations, so I guess I disagree with RSJ there.

RSJ said nations who peg shouldn't need forex reserves, but this is incorrect. They need forex to defend the peg, and a pegged currency is an exogenous currency regime, not an endogenous one.

RSJ: The marginal cost to a bank of extending a loan is the cost of capital. FFR may be an input to it, but that isn't the same thing. Banks don't borrow from the Fed at FFR and then lend that money out. At a macro level, you get the marginal CoC equalling the marginal RoC (I guess) which seems like the right result. Credit markets SHOULD be different that the term structure on Treasuries.

My question about whether interest payments on debt were "fiscal" was me trying to understand how the Govt made those payments. If it just marks up the account, then it's Government spending just like any other. Whether this is expansionary (per MMT) or neutral (per your opportunity cost argument) is an interesting question. I don't think of the foreign sector "leakage" as being to the detriment of national households, I think it means you can just deficit spend even more before creating inflation.

Winterspeak: "RSJ said nations who peg shouldn't need forex reserves, but this is incorrect. They need forex to defend the peg, and a pegged currency is an exogenous currency regime, not an endogenous one."

I'm halfway between the two of you there.

If you peg the exchange rate, and there is an excess supply of your money, you need to reduce the stock of money, or the peg will collapse. In principle you could reduce the excess by selling bonds, but in practice it would be very hard (read "impossible") to fine-tune the peg if you didn't have at least a small stock of forex reserves to buy and sell.

But if you peg the exchange rate, sure, the exchange rate is exogenous, but the stock of money is endogenous. Would you call that "exogenous money" or "endogenous money"? Does it really matter, as long as you explain what it is that's exogenous/endogenous.

exogenous. a regime running a peg is obviously operationally constrained in a way a regime running without a peg is not.

OK. Makes sense. I would say that monetary policy is determined by a rule, rather than discretion.

"RSJ said nations who peg shouldn't need forex reserves, but this is incorrect."

I said no such thing.

" This doesn't seem so different to what happens during open market operations, so I guess I disagree with RSJ there."

Wow. Ok, riddle me this:

1. The government just announces to lower FF by 100 basis points, and it drops that hour. Maybe they spend 50 billion on an OMO to move the rate to exactly what they want, and reverse it a month later, but the rate *stays* at the lower level, even *after* the trade is reversed. The demand curve is vertical.

2. The government announces to intervene in the mortgage market, spends over 1 trillion dollars, and experts peer at correlations debating about whether rates fell by 20 basis points or 50 basis points. And you can't say with confidence that it was one or the other. The demand curve is horizontal.

Until you understand the difference between the money markets and the capital markets, you will not be able to explain the difference, and you will believe that control of price in one market is the same as control of price in the other.

Moreover, both of you seem to argue that there are separate markets for government versus private sector debt. All debt competes. Government debt is about 20% of the total stock. Therefore the government would need to buy debt faster than the private sector can sell it. Then they would need to start buying equity, etc. That should tip you off to the difference between capital and money markets.

Finally, it is illegal for the (U.S.) OMO to purchase non-government debt -- there is a loophole for "emergency loans". Basically, you can quickly retire all the government debt without moving capital market rates too much.

And then what will you do? Car rentals, huh?

RSJ: Sorry, I misquoted you, but my point was that countries that peg are not running endogenous currency systems (so I would not consider a pegger a "monopoly issuers of their own currency". Nick read my mind correctly though!

I also want to set aside the agency debt issue for just a moment as I think it's complicated for a number of reasons -- can we focus on Treasuries instead?

Right now, the Govt auctions, say 10 year Treasuries, so their price is whatever gets the market to clear. Instead, suppose the Govt just fixed the price ("3%") and let Q float. In this instance, an auction may fail, as there may be insufficient buyers at 3% to purchase the entire issuance. MMT would say that this is no big deal, as the Govt does not need to issue Treasuries to spend, and the spending can remain trapped in bank reserves.

Suppose the demand is much higher, and the auction sells out, but the price keeps rising in the secondary market (your point, I believe, and what Nick was getting to re a durable goods monopolist). The Fed could sell additional notes to lower the price back to 3%. Suppose demand then changed, and now the price falls too low. The Fed now buys notes back to raise the price.

In the case of Treasuries (and Treasuries alone) this seems very similar to ONO to me, and I don't think the Govt has any limits buying or selling Treasuries.

There is LOTS the govt can keep itself busy with once it retires all its debt and stops issuing new Treasuries ; )


You are still treating bonds as a consumption goods.

1. All bonds compete. There is not a separate return for government and private sector debt.

2. Anyone in the private sector can and will issue debt -- or re-finance debt -- when doing so will save them money

3. The interest rate charged is the long term growth rate of each firm's capital stock -- as the firm will re-invest out of earnings.

You are still operating in the world in which there is some special utility -- patriotism? -- that would cause someone to accept a lower return on the government instrument than on the private sector instrument, and you are assuming that the private sector will not issue more debt when the cost of doing so falls.

It's a view of the world in which money is an idiosyncratic consumption good with zero elasticity of substitution between different debt instruments, and no relationship between ability to pay and rate charged.

Moreover, the government would need to buy all securities in all markets in order to control the yield. It would need to do this simultaneously, and make it illegal for the private sector to issue more debt as well.

Finally, I still haven't heard of an explanation as to why the government was unable to control mortgage rates when it can control the MM rates so easily. It was afraid to even announce a target price, whereas it announces and keeps target MM yields easily. Can you see the difference? Can you explain it? There is a reason why it is illegal for the CB to purchase debt not guaranteed by the government, and that reason has nothing to do with the CB "losing money" -- understand that, and you'll understand why the government can't control yields.

RSJ: In an endogenous currency system, the Govt does not need to issue debt to spend. The US Govt could stop issuing Treasuries tomorrow and it would have no impact on its ability to deficit spend. It would drive rates to zero (given the mechanics of the interbank market) but in today's world of ZIRP + OIR I don't even think that would be a big deal operationally.

Bonds do compete on the demand side in that purchasers need to decide what do buy, but the US Govt is not competing on the supply side with (say) a corporate issuer in that a corporation that fails to borrow might go out of business, while it makes no difference to the Govt whether anyone buys its bonds or not so long as it does not mind a ZIRP environment.

This is the point I was trying to make when I said MMT does not see Gov bond issuing as an "offset" of any sort to deficit spending. It is merely a mechanism to maintain the FFR above zero given how the FFR is currently set (actually, how it was set before OIR).

I asked that we set agencies aside for just a moment, I am happy to get back to that, but I want to make sure that I'm clear on Treasuries first. So I will address your question as soon as we're square re: Treasuries.

My assertion is that the US Govt could, if it wanted, set the entire yield curve for Treasuries at whatever it wants it to be. And yes, this means it may have auctions where there are no buyers, but the Gov stands ready to sell at x%, and that the Govt, if needed, would buy/retire all outstanding Treasury debt. Basically, it would intervene just as it does in the ON IB market. The Govt does not need to sell Treasuries to deficit spend, as that deficit spending can remain trapped in reserves with no operational consequence.


OK, there are diminishing returns here, but you cannot make up your own definitions. If the government buys up all the treasuries -- which is what it would need to do -- then that is not the same as controlling the yield curve. Moreover, assets that are not sold have no price. The government cannot sell an asset to itself -- with the government as the sole buyer -- and declare that is has controlled the price. Yet this is what you are arguing.

The risk free rate is the (ex-ante) expected return from investing in broad basket of securities of a given maturity. In practice, we use government bonds as a benchmark for this rate as they are assumed (in most cases) to have zero default risk, but the fact that government securities are used to benchmark this rate should not fool you into thinking that if the government bought up all the government securities for X, that the risk free rate would become X.

In terms of the policy points you were raising, I think they only serve to add confusion. I am trying to describe constraints on the economy whereas you are arguing policy points. Let's agree first on what the constraints are, and leave the ideal policy response for another time.

And with respect (my last post was a bit snippy), I think there is still confusion about money markets versus capital markets, the interbank market versus credit markets, as well as confusion about the distinction between FFR and investor rates. All these are jumbled up so that you believe investors are receiving FFR for a zero day commitment, or that if the government were to stop selling debt, then capital market yields would fall to zero. Perhaps I am misreading your post, but keeping these concepts separate is a pre-requisite, whether you agree with me or not.

RSJ: I think I'm being meticulous about my definitions, but I agree we seem to be at cross purposes. I was talking about Treasuries (strictly) and did not think I was bring up policy points, just mechanical operational points.

I think we can let the Treasury issue rest here though: if the Govt offers a 10-year Treasury at 3% (say), then I would say the price of that Treasury is 3% whether or not anyone buys it. It is not the market clearing price, true, but if you want it you know exactly how much you have to pay. If you accept this, then yeah, the Govt can set the entire yield curve for Treasuries. If not, then not. I would add though that "market price" is a slippery concept in the ON IB market as that is so manipulated by the Government. I would not consider that to be a true market, and this may be why I'm very comfortable extending that to the rest of the Treasury yeild curve.

Moving on.

I agree that, although we use govt bonds as a proxy for the risk free rate it is not the risk free rate. I'm not sure where I claimed this was the case. In an endogenous currency regime, govt bonds need never default so there is no default risk, but there are others which I'm sure you know about.

If we had a zero yield curve on Treasuries, private sector credit markets would still need to, you know, factor in credit. Depending on the creditworthiness of the borrower, rates could be anywhere, and in this instance borrowers are competing to issue because they need the money. I don't think the Government really can, or should, interfere in this market (or, if it does, it's engaging in fiscal policy by essentially forcing loans that will not get paid back. This does happen, I believe, especially via the State owned banks in China, but that's a digression). Similarly, capital markets still need to value on DCF or whatever. They may no longer use short term Treasuries as their risk free discount rate, which is fine.

I'm not advocating a zero Treasury yield curve in this post (I can take both sides on that). I'm just saying that the Govt can do it, and I don't see why credit and capital markets can't continue more or less as they do today. Of course, I am also asserting that a standing offer price is the price even if there are no bids (which is me making up a definition, but a reasonable one I think) and pointing out the fact that the Govt does not issue debt to fund spending (which is unique to it in a endogenous currency regime).

So, a short note on agencies. Agencies are not pure Govt entities, and so they take on some of the private sector credit concerns, which cannot really be controlled by the Govt. This is why Govts ability to influence them is limited (but not zero. Without the agencies, mortgages would be more expensive today where I believe 90%+ mortgage debt in US is Govt backed). Moreover, Govt has a long history of training the market that it can and will control MM rates through active intervention, so there's an expectations factor there. I don't think it could ever get there with agencies unless it brings them on 100% as Govt depts, but I think it can get there with longer term Treasuries and might need to go through a training period first.


The issue here is that the the price of securities is not set by default risk -- default risk is a complicating factor, which is why you want to talk about risk-free rates, which removes this complication.

What you are left with is the "true" cost, or the risk-adjusted return, which is set by opportunity cost.

A growing economy will have a non-zero opportunity cost. A growing economy in which capital is endogenous will have a non-zero, endogenous opportunity cost. Depending on the economy, really depending on perceptions about the performance of the economy, your opportunities for obtaining a certain risk-adjusted return will differ. If the economy is booming, then you will earn a greater return, and if the economy is depressed, there may be no opportunity for a return. This has nothing to do with credit-worthiness.

And all government securities, as well as private sector securities, must meet those opportunity costs and the yields on those securities will converge to the opportunity costs.

So you must believe that opportunity costs -- possibilities for economic growth -- can be controlled by the government if you believe that the *market* price of government securities can be controlled.

This is why I don't agree that a "standing offer" is enough to set the price. The price -- if the risk-free security is allowed to trade -- will be the opportunity cost. If the risk free security does not trade, then debates about the price are moot.

And that opportunity cost is endogenous, depending on the rate of invention, of new markets opening up, of the growth rate of capital.

Now let's move to policy:

If the economy is such that we are Japan and there are few opportunities for growth and profit, then rates will be low regardless of what the government offer price is. In that case, if the offered yield is too high, there will be arbitrage -- everyone will buy the security, shifting resources away from the real economy and towards market participants. If the yield is too low, no one will buy, and no resources will be shifted.

Do you see an asymmetry? The results of such a policy -- a policy to sell unlimited amounts of a financial asset at a fixed price -- will be, on average, to enrich the financial sector at the expense of the real sector. Only if that price is always guaranteed to be below opportunity cost will no one be enriched, but then no one will hold the security either. Therefore the possible effects of such a policy go from no effect to harmful -- the results are never beneficial. The advantage of selling something for the market price is that no one is guaranteed to benefit from the sale, and arbitrage becomes more difficult.

In general, I think the goods market is where we should have arbitrage. The act of inventing new goods and creating new wants, as well as responding to shifts in wants and shifts in production costs creates the $100 bills that rain on the ground, and businessmen rush to vacuum them up. Economic activity is the vacuuming up process, and it should be restricted to providing goods and services to people.

In the financial markets, there should be no $100 bills that can be vacuumed up, and all the attempts of government to control rates -- whether this is mortgage interest deduction, or the 401K tax breaks, or the bank bailouts, or the lowered FFR, interest deduction on business debt, or your own proposal to sell risk-free assets at non-competitive rates -- that creates an environment in which the $100 bills are dropping in the financial market, rather than dropping in the goods market. The result is a financialization of the economy, with resources, particularly human capital, shifted towards the financial market rather than towards the satisfaction of wants in the goods markets. You should have a steady state of arbitrage in the goods markets (with new markets appearing as the arbitrage opportunities in the mature markets dissipate) and no arbitrage in the capital markets.

"Moreover, Govt has a long history of training the market that it can and will control MM rates through active intervention, so there's an expectations factor there."

The demand curve is not vertical in the MM market because of "training" but because of the flexibility of borrowers and lenders in these markets. When the port authorities were paying 10%+ annual interest costs for funding it was not because of "training" but because they needed to roll over short term debt to meet payroll. And when lenders today accept 0% for short term debt it is not because of "training" but because they need to park their working capital for a few weeks somewhere. The alternative for the port authority is to shut down, and the alternative for the lender is not make that profit in the first place.

The MM are for cash-flow management purposes, and just as households purchase checking accounts inelastically with the interest rate, so the MM markets are inelastic. The capital markets are different -- people do care about return, not about the transactional value of having a place to park money for a month. It has nothing to do with "training" but with the realities of what these markets provide: cash-flow management versus investment. Now there is a continuous spectrum in which cash-flow management becomes investment. There is not a sharp cut-off. That is why the best way to look at this is in terms of the arbitrage-response function. Eventually the port authority can shift its capital structure to go from being a borrower to a lender, and eventually investors can drop out of the capital market to exploit higher rates in the MM market. Eventually. So view the slope of the demand curve as being almost vertical for zero day funding and almost horizontal for perpetual (e.g. equity) funding, and the slope rotates as the time of the capital commitment increases. As the demand curve flattens, the ability of the government to control price by adjusting quantity rapidly diminishes, and simultaneously the economic damage in terms of windfalls supplied when the market price is not exactly equal to opportunity cost increases.

RSJ: This thread was originally about MMT and the Long Run Phillips Curve, a topic I have no interest in, although I did like your comments re: labor/capital etc. In the course of those comments, you represented the MMT view as believing that Govt debt is an offset to deficit spending, and this is simply not their view. Govt does not need to issue debt to fund or finance or offset anything. It only issues debt to drain reserve accounts and thus maintain an FFR above zero. With OIR it does not even need to do that (as you can plainly see today).

I disagree with your assertion that whether an economy is booming or collapsing has no impact on credit worthiness, as it clearly does. Banking (and credit) is pro-cyclical.

I'm comfortable with settling the question of whether or not the Govt can set the yield curve for its debt depending on what you mean by "price". You think you need a market clearing price to count, and I think that in this case, since the market does not have to clear, in the absence of a market clearing price an ask is good enough. We can agree to disagree quite happily.

There are a lot of assumptions embedded in your policy comments. You seem to assume that banks borrow from the Govt to fund loans (and they earn a spread between the low Govt rate and whatever they charge to borrowers). This is not true, banks do not lend out reserves, they lever on top of capital. Therefore, bank lending is capital constrained and therefore the marginal cost of a loan is the bank's marginal cost of capital.

You also seem to believe that there's a "loanable funds" market where money that moves into Treasuries is not available to be loaned out. This is also not true, there is no loanable funds market, and the S=I causality flows from investment creating savings, not the other way around.

It seems you also believe that shutting down the Treasury markets will be a windfall for the financial industry. I disagree, and all those Govt bond traders probably will as well!

In general, in the MMT view, the term structure of Govt liabilities (held as assets by the private sector) have very little to no impact on economic activity. There is no mechanism for them, as monetary base is disconnected to money supply. The quantity of Govt liabilities does matter very much though, and the mechanism is straightforward, and this is why MMT rejects interest rates in particular and monetary policy in general to the extent that it does, putting fiscal front and center.

Anyway, I've thrown even more assumptions at you. I'm not sure to what degree you understand MMT (irregardless of whether you agree with it or not) and I don't like assuming that someone disagrees with me just because they don't understand -- people can understand perfectly well and still think you're full of it! So, if I'm rehashing stuff you already know but have rejected, or if I'm inferring assumptions you don't actually hold I apologize. There's too much here to get through, way beyond the scope of this thread, so I'm signing off.

If you want to read more about MMT, I would recommend Mosler's mandatory readings above Billy Blog (so many words! so much posturing! where is the content?) If there's a place you could point me to for the mandatory readings of RSJ, I would appreciate it, although it's sounding like it's the textbook stuff which I'm familiar with.


It's a dead thread, so I'll comment freely.

"In the course of those comments, you represented the MMT view as believing that Govt debt is an offset to deficit spending, and this is simply not their view."

No, I didn't argue this. I used the phrase "offset government deficit spending with bond sales" because this is how our government operates -- there are laws requiring this. Changing these laws would take an act of Congress just as the creation of the central bank took an act of congress. It is just as radical of a change as debt repudiation, adoption or abandonment of the gold standard, or any other serious change. Until such a change is made, Treasury will be legally required to offset government spending with bond sales and such a constraint is just as binding as being on the gold standard. There is no meaningful difference between an external constraint and a self-imposed constraint, as honoring your foreign obligations is also a self-imposed constraint.

One annoying aspect about discussions with MMT proponents is the evangelistic nature of the dialogue, replete with disputations over the allowed language that can be used. ugh.

"I disagree with your assertion that whether an economy is booming or collapsing has no impact on credit worthiness, as it clearly does. "

I made no such assertion. You argued that there would be interest rate differentials due to credit-risk. I pointed out that even adjusting for credit-risk -- e.g., by looking at risk-free rates -- that still these rates are endogenous. How you parsed that into an assertion that "whether an economy is booming or collapsing has no impact on credit worthiness" is puzzling.

"You seem to assume that banks borrow from the Govt to fund loans (and they earn a spread between the low Govt rate and whatever they charge to borrowers)."


"You also seem to believe that there's a "loanable funds" market where money that moves into Treasuries is not available to be loaned out. "

Again, double plus no.

You believe in loanable funds, whereas I do not.

You will dispute that, and argue that MMT does not believe in loanable funds, but they do. That belief is at the heart of my disagreement vis-a-vis interest rate policy.

Loanable funds is the belief that there is a downward sloping interest rate price for bonds with respect to quantity. Now you have to be careful and adjust for credit risk and inflation expectations -- i.e. obviously a business that sells more bonds than it can service will see the yield rise. But the yield is not rising because there are more bonds, but because the expected payout per bond is changing. If the revenues of the business were to rise, it could increase increase the quantity of bonds sold without seeing the price fall. That is why you want to talk about real risk-free rates, to get rid of the credit adjustment and talk only about the underlying expected payout.

Anyone who argues that by removing government bonds from the market, you will make them more scarce and therefore drive up the price is making a loanable funds argument. They can say they don't believe in loanable funds all they want, but *if* they believe the demand curve is downward sloping with quantity, then they are claiming that loanable funds equilibrates supply and demand.

Decreasing the quantity of chocolate may cause the utility per gram of chocolate consumed to rise and therefore a price increase is justified. But you cannot argue by analogy that the same is true for financial assets -- that if I offer to sell you one bond that will yield $1 in real risk-adjusted return, that you will be willing to pay more per bond than if offered to sell you two such bonds. You must explain the source of the irrationality and cannot rely on diminishing utility -- there *may* be a reason for the downward sloping curve -- for example, a failure in the payments system can force a liquidity crisis in which people cannot borrow to buy the bond, and therefore the arbitrage function isn't effective enough to absorb an increase in quantity. Those types of arguments have weight in the short term funding markets, but not in the capital markets. In the capital markets -- which drive real investment -- a horizontal demand curve assumption is the only rational assumption to make.

And as long as you believe in a downward sloping demand curve, you are surrendering to a Say's Law economy, at least over the long run, and it doesn't much matter what you have to say about functional finance. You will be, as Nick once quipped "Wicksellians without the micro-foundations", and won't substantively challenge the neo-classical paradigm.

If you were to sum up my views in a nutshell, I would argue the following:

1. What equilibrates planned savings and planned investment is the accuracy of expected return. When the cost of capital -- the ex-ante expected return -- is less than the return on capital (the ex-post return), then investors get more then they paid for, you get an investment boom, and wage shares fall. When the cost of capital is greater than the return on capital, investors get less than they paid for, and you have liquidation, unemployment, and rising wage shares.

2. There are feedback effects, or persistence, from these expectations misses.

3. Counter-cyclical fiscal policy has real positive effects to the degree that it mutes the feedback effects of expectations misses, and it has real positive effects as long as the return on capital is less than the cost of capital. As the government spends more and more while draining less and less, profits will increase, but if they increase beyond the cost of capital, then capital holders are getting free money and wage shares are falling. When the fiscal spending reaches the point where returns exceed costs, then fiscal spending becomes harmful from a distributional view. Tax policy can be used to offset this.

4. The "best" interest rate is an accurate rate, not a "low" rate. Any deviation is harmful. Attempts to lower rates out of a mistaken belief that low yields equals low rentier profits will result in falling wage shares and rising rentier profits.

5. Attempts to manipulate the long term rates via adjusting bank marginal cost of funds disproportionately affect the real estate market, as banks cannot extend loans to non-real estate businesses in any meaningful quantity. Effectively you are hoping to set off a spending boom funded by increasing land prices, and are hoping that the falling land yields will drag down capital yields as well. Generally this does not succeed.


Thanks for your response. You are, of course, correct in saying that enabling the Treasury to run an overdraft at the Fed requires changing the law (which is why MMT stresses the "operational" as the limit is self-imposed), but I don't agree that this is as radical a move as creating a central bank. I believe other entities have accounts with overdraft provisions at the CB.

Let me repeat my understanding of your perspective, and you can tell me where I'm right (or wrong):
1. If the Govt offers an endless supply of 10-year tbill (zero default risk, but inflation risk etc.) for 3%, and none are bought, there is no price (since the market did not clear).
2. If the Govt offers a 10-year tbill at 3%, and that is bought, but then starts trading at 4%, then the price is 4% (market) not 3%.
3. If the market price is 4%, but the Govt wants it to be 3%, buying or selling additional bills will not get you there (because there is no downward sloping demand curve).

Therefore, when MMT talks about "setting the yield curve to zero" the only way they can really operationalize this is to eliminate the yield curve, cease debt issuance, and let the Treasury run an overdraft at the CB.

Do I understand you?

Re: whether the change is radical or not

The magnitude of the change is measured by the magnitude of the political shift needed to bring the change about, not by the number of lines of code that need to be changed, or the number of pages in operational handbooks that are re-written. Let's be honest about this -- you are advocating a change that strikes at the heart of our gold standard thinking. I support the change, too, but don't misrepresent the political situation. It is more serious than going off the gold standard, since we haven't done that in our minds. And of course, many of the economists are still stuck on barter -- you are way ahead, here.

About your questions,

There is only one yield curve -- the riskless yield curve. This is the ex-post return obtained from buying a broad basket of debt at a given maturity. Alternately, it is the opportunity cost of making a capital commitment for a given period of time. The rate charged to a specific borrower is this rate plus a credit premium. As the government has zero credit premium, it would pay this rate.

Whether or not the government actually participates in this market is a separate question -- the opportunity cost is there regardless, and that is what the debt would trade for, if it was allowed to trade.

When you are asking someone to buy a 10 year bond, it may be the case that this is the first 10 year bond that you are selling, so to you the bond "has no price", but that *does not* mean that the price is undefined, because the purchaser already has a basket of 10 year options to invest in. And your "first" bond must compete with these opportunities, and therefore your borrowing cost is defined by that opportunity cost. If you were to try to sell the bonds, you would need to pay this rate. If you do not sell any bonds, fine, your borrowing costs, if you decide to borrow, are still set by this rate, and the expected return exists independently of whether you borrow or not.

That is at the heart of our disagreements about price.

"Therefore, when MMT talks about "setting the yield curve to zero" the only way they can really operationalize this is to eliminate the yield curve, cease debt issuance, and let the Treasury run an overdraft at the CB.

No, the only way they can operationalize this is to tank the economy so that investors do not expect any positive nominal return from making a commitment of any maturity. Basically you have to drive expected nominal growth rates to zero. It is a foolish, dangerous goal.

Now I should say that there are all sorts of complications here. For example, some investment funds may have a requirement to hold government debt, and in this case they are captive buyers and would buy debt at any price -- until their charters were changed. You can argue that China is not after a return, and is also price-inelastic. But that can also change.

More importantly, the financial system has grown dependent on safe long term debt that it holds on the asset side of it balance sheet, and broker dealers need the ability to short safe debt in order to absorb risky debt. Money market funds are required to hold government debt. So any change in debt issuance policy would need to be accompanied by financial sector operational reforms, otherwise it would put the system into shock, raise borrowing costs, and put downward pressures on credit growth. Such a shock might end up hurting the economy and lowering returns.

RSJ: Thanks for your response. Let me press you on your opportunity cost point.

The opportunity cost of doing something is NOT doing the next best thing. You assert that the next best thing to buying a 10 year t-bill will be buying a broad basket of debt at that same maturity. Is that really the right comparison set? A broad basket of, say, AAA corporate paper has credit risk, Treasury debt is unique in that there is no chance it will not be honored. Cheques from the Treasury don't bounce.

Since that is its unique feature, wouldn't the opportunity cost be leaving the money in reserve accounts or FDIC savings accounts? That has a closer risk profile to Treasuries than any basket of long term private debt.

As you point out, there are also participants in the Treasury market who have currency/import/export agendas -- what would THEIR next best option be? It isn't obviously a diversified basket either.

I fully agree that the changes discussed above would require financial sector operational changes. And I 110% agree that letting the Treasury run an overdraft is operationally a small step from where we are, is a profound and radical change because of the cognitive shift it requires. The emperor has been fully naked since the 70s. This would be pointing it out.


You can diversify away credit risk, but not market (or inflation risk). The lack of credit-risk is *already* reflected in the price of treasuries -- the price is set to achieve a level of indifference between treasuries and assets that have credit risk, which is what allows those treasuries (or other assets) to be bought and sold in exchange for each other.

And that indifference price must be such that it measures the expected rate of default, so that the total return to holding treasuries is the same as the total return to holding risky assets -- so that there is only one return, and all assets converge to this return.

In the money markets, this is different -- you just want a safe place to park your operating surplus, and others need to fund their operating deficits. In that market, yes, the alternative to a government bond is a deposit account. But only in that market, and not the capital market. And the capital market is what determines investment decisions for aggregate demand.

By the way -- this concept that there is only one return, but a myriad of investments that yield this return, is why time preference does not play a role in setting interest rates. When you buy a 10 year bond, you are not promising to forego consumption for 10 years. You could sell the bond tomorrow. You may have borrowed to buy the bond, etc. The decision to forego consumption and invest is influenced by time preference, but the price of the security is determined by expected return and opportunity cost -- not the quantity of people desiring to save or invest.

As a result, the Japanese did not become very patient when the economy tanked, lowering yields, but the economy tanked, lowering returns and therefore yields on all assets, government or otherwise. Neither, during the boom period, were the Japanese very impatient, but the return on government as well as private sector assets was high because the economy was booming and there was an expectation of large returns to be had.

So these micro-level psychological factors, such as patience/impatience and risk tolerance or intolerance play an important role in determining the propensity to consume, but they do not play a role in determining yields. Yields are determined by expected return, and the economy is able to endogenously generate enough financial assets to fund investments that meet the expected return regardless of how much people decide to save or invest. In essence, any theory of endogenous money must incorporate the assumption of a horizontal demand curve, if it is to also be an arbitrage-free theory.

OK. So say the Govt sets the FFR at zero and stops issuing Treasuries. The yield curve is whatever it is. If someone wants to take on credit risk and make an investment decisions, they are free to do so in the capital markets as they do now. Else, they have money sit in deposit accounts of one form or another.

I think that that is operationally what "setting the yeild curve at zero" really means for MMT, and it is quite different from offering Treasuries at a variety of maturities at 0%

OK, first, you can diversify away credit risk -- credit risk is just a distraction here, as all that matters is expected return.

As to the specific proposal, I have two objections:

A. You are giving banks below market funding costs by setting FFR to zero. This is a harmful incentive, as the return on capital for someone starting a bank should be the same as the return on capital for someone starting a non-financial business, otherwise you will financialize the economy and make banks too powerful.

B. You are making the things that banks lend against -- in this case, land -- cheaper to finance than things that banks cannot lend against -- productive capital. This will hurt the economy, maybe with inflation, maybe with asset price inflation, and maybe with deflation (when the asset price inflation runs its course). One way or another, this is a harmful set of incentives. You can offset this effect by, for example, by rationing loans -- China is trying this -- but that's not robust, and is in any case open to corruption.

I think you can mitigate A and B by imposing a fee on each asset held by the bank. It's not as robust, from a compliance standpoint as making money have a non-zero cost, but it's the next best thing I can think of.

That can be used to lift their marginal cost of funds back to the cost of funds of the productive sector. So instead of charging a FFR of 4%, you can charge a 4% fee on all assets held by the bank. This would effectively undo the harmful effects of A and B. The fee would be adjusted to match whatever the long term risk free rates are -- and these can be calculated from an index of private sector bonds even if there are no treasuries outstanding.

Assuming that you do this, then I have no problem with it.

But I don't think this is what the MMT adherents really mean. From reading the essays of Bill, Randy Wray, Warren Mosler, and others, I think I can summarize the rationale for a zero FFR as follows:

1. The government, as the issuer of currency, *must* set a rate, and the "natural" rate is what it would be if the government did not intervene in the interbank market by selling bonds.
2. Any positive risk-free rate results in rentier profits -- e.g. no one should get a positive "risk-free" return.
3. Bank regulation should be used to prevent imprudent loans, rather than making marginal bank costs more expensive.

The problem I have with this rationale is that

1. The government is intervening in the interbank market by adding currency, so there is nothing "natural" by only intervening to add supply but not intervening to remove supply.

2. In a growing economy, there is always a positive risk-adjusted rate. For the economy as a whole, rentier profits are not determined by the rate of interest paid, but by debt growth. This is a simple fallacy of composition error, in that the interest paid by a business comes at the expense of dividends. Or the interest paid by a household comes at the expense of spending on goods. So in the aggregate economy, a "rentier" holding a broad portfolio of equity and bonds is not actually receiving more money due to the interest rate rising, they are receiving more money due to aggregate debt growth increasing. Now, if assets are fairly priced, then the growth rate of debt will be the rate of interest charged, but trying to lower the rate of interest charged is not going to necessarily succeed in lowering the growth rate of debt -- it could have the opposite effect, and is in any case harmful to the economy. Besides, there is such a thing as tax policy to deal with rentier profits.

3. This is similar to the argument that it is OK distribute loaded guns for free on demand, because an effective police force should be used to fight crime, rather than the availability of guns. Maybe, but in general you cannot give banks access to zero funding costs, and make lending against land cheaper than lending against productive capital without experiencing harmful effects. The system needs to be robust, and you shouldn't stack the deck against the regulators.

So I would turn this around and ask -- why have a zero FFR?

Winterspeak: "the Govt sets the FFR at zero and stops issuing Treasuries".

Is this something that MMTers think is a good idea? Hope not, that's an idiotic idea in practice.


You can diversify away credit risk until you can't.

The FFR does not set the banks cost of capital, or if it does, it is just one factor amongst many. Since CoC is their funding cost, we don't agree here.

I don't know why you keep raising agencies and land. It isn't something I brought up -- maybe it figured prominently earlier in the thread? It is true that land is very standard against mortgages, but if you're worried about funding disparity, certainly you can balance that out through a variety of mechanism (including the ones you suggest). Banks make collateralized loans to businesses as well though. Basically, the marginal RoE will equal the marginal CoC across all asset classes.

MMT would eliminate the interbank market completely, and have all reserve requirements funded at the discount window, uncollateralized, at 0%. This is what it ends up doing whenever the IB market breaks down anyway.

Their point on rentier profits is focused on the fiscal effect that Treasuries have. MMTs think the fiscal effect is more important than monetary mechanisms.

I understand your gun analogy, but don't think it is applicable in this situation. FFR is not equal to bank's funding cost. And we *certainly* need much better policing!

I'll respond to your question later.

AdamP: It is what they advocate, and it may well be an idiotic idea in practice.

Hmm, maybe this is a definitional issue. To me, when you diversify, you eliminate credit risk and are left with market risk. The whole market can and does move, and losing a dollar due to an interest rate movement is the same as losing a dollar due to a default. Expected return incorporates all these risks, without biasing one over another. It seems to me that you are trying to bias a portfolio movement due to default as being more costly than a portfolio movement due to interest rate shifts. As someone who has lost money via both methods, I can guarantee you that owning treasuries is not risk free, and that the pain of the loss is the same. All that matters is the expected return from whatever strategy you employ.

I agree that FFR is not *equal* to a bank's funding costs, but it is the dominant factor in setting this cost. I think it's pretty clear that banks enjoy below market funding costs.

"MMT would eliminate the interbank market completely, and have all reserve requirements funded at the discount window, uncollateralized, at 0%"

Am I the only one who sees the huge risks involved in this? And for what purpose would we be so reckless?

"Their point on rentier profits is focused on the fiscal effect that Treasuries have." I know and it's not a valid point. Actually they don't even define "rentier" -- it seems to be someone who holds a government bond and receives income from it, whereas someone receiving income from some other asset -- e.g. a landlord or dividend holder -- is not a rentier. Strange.

I think a key source of confusion is the cancellation of liabilities into Net Financial Assets -- they do this across the entire non-government sector, and so business liabilities cancel with household claims on businesses to give no net wealth benefit.

Effectively this is an argument -- or really a definition -- of wealth that assumes the impossibility of capital growth. Obviously if you make this assumption then the only wealth will be government assets and treasuries will have a fiscal effect. Part of this is a fixation on government assets -- they are a non-rival good that anyone would hold at all costs, presumably.

A better aggregation is to look only at household net worth, and in that case, Treasuries have zero fiscal effect. Fiscal spending has a fiscal effect, primarily as it allows business profits to increase. The advantage of this aggregation is it reflects what people see when they look at their portfolios. So when an MMT proponent looks into a household balance sheet, all the equity claims and private sector bonds are invisible -- all he sees is currency and treasuries. As a result, he sees a fiscal effect from treasuries. When anyone else looks in their bank account, they see everything, and expect all their assets to increase in value. Because they bought the treasuries at market prices, relinquishing some other interest paying asset in order to do so, they do not see any benefit from holding the treasury that is in excess of their opportunity costs, and they would be equally wealthy whether they held the treasury or some other asset instead. Thus there is no fiscal effects from treasuries -- it's just a misunderstanding on the part of MMT because they improperly measured the financial net worth of households.

I would be hesitant to lend in unlimited amounts to banks without collateral and at zero interest rates, all due to a foolish misunderstanding. Besides, you have tax policy to deal with distributional issues.

RSJ: Credit risk is when you don't get paid back, on time, in full. It is different from inflation risk, interest rate risk, political risk, etc. (although those are all real, and can certainly hurt the pocket book as you say!) There is a fashion these days to blur them, but I'm old fashioned about these definitions.

Also, FFR may be the dominant factor in a bank's funding cost until it isn't. I forget the terms Buffett got on his note to Goldman Sachs, but I'm pretty sure the FFR was about zero and the note was not. There's an old Fed paper on this where, IIRC, bank's CoC was much (3x) higher than the FFR. Anyway, you can look this up--I don't think they are at all close.

The MMT position points out that 1) market forces do not work on the liability side of bank balance sheets, and 2) when the interbank market breaks, it threatens a collapse in the payments system which has the CB step in and lend, uncollateralized, to everyone via the discount window. Since an actual bank collapse gets underwritten by the FDIC, and the FDIC is part of the same Govt as the CB, it makes sense to 1) acknowledge market forces don't work on the liability side of the balacne sheet, 2) ensure that nothing can threaten the payments system, and 3) use strict capital controls to manage risk (on the asset side). You can disagree with their conclusions, but I think their initial observations are correct.

In order for the above to work, you need to eliminate the interbank market, and the most natural place to set FFR then becomes zero. (You could maintain OIR, but why?) Also, if you implemented the Treasury overdraft etc., you would also eliminate interest rate risk as it would always be zero.

"Rentier" is clear to me. It means people getting Govt money for not doing anything. A landlord or dividend holder has invested in a property or a company -- that's something!

I disagree that this assumes wealth without capital growth. Wealth is in real goods and services, and those certainly benefit from capital, have real value, and provide solid collateral for private sector credit.

I also don't agree that, when MMT looks into a household, they ignore private sector credit. They do, and they ask (at a sector level) to what degree it's sustainable. This is more than the Fed did during the bubble, and it's more than the Fed's doing now.

The role of banks in MMT is to put private credit in first loss position before Govt credit. But it's all Govt money in the end.

Heh, I didn't realize that you were *that* old fashioned, because you are articulating a zero opportunity cost (e.g. zero sum) barter view of the economy, in which investing is like loaning grain to your uncle.

He gives you an IOU and you are paid back with corn. You must defer consumption and patiently wait for him to repay you. In that case, you would demand compensation for the consumption time shift as well as risk.

But in a non-barter world, the liability *is* the asset with which you purchase both other assets as well as goods and services. You don't "get" the government annuity, you buy it at competitive prices, and in order to buy it, you must first sell some other private sector asset first. "Lending" is nothing more than exchanging one person's debt for another's, and cash is just a mechanism that operates underneath to keep things liquid. "Borrowing" is what actually sets the total stock of wealth that can be held, and the "lenders" pick and choose what they prefer to hold in order to set relative borrowing costs.

I agree that this can be subtle -- but the existence of secondary markets that can make these claims liquid makes a qualitative difference, and all of a sudden you do not have any more or less wealth when you rotate out of the Verizon bond and into the government bond, but when you loan money to your uncle, you really have less money to spend -- because your uncle's note is not liquid, whereas the Verizon bond is.

In that case, loans create deposits is just a special case of liabilities creating assets generally, and the household sector is only able to purchase a government asset if it first shorts an equivalent asset that pays the same interest. "To lend" becomes the act of shuffling around whose debts you hold, and in the process, helping to set borrowing costs. "To borrow" is to create the financial assets that creditors hold. So a large qualitative shift occurs when there are secondary markets that can make financial claims liquid.

At the end of the day, you are only receiving an economic profit if the asset is mispriced, and when you are not profiting, the return you obtain is equal to the growth rate of the economy as whole, as this sets your opportunity cost. And everyone gets this return, on average, regardless of what paper they hold.

"It means people getting Govt money for not doing anything. A landlord or dividend holder has invested in a property or a company -- that's something!"

I don't know, I log into my account and click a few buttons, rotating out of a REIT and into a government asset, and the next month I rotate out of the government asset and into NSRGY. I don't *feel* like I am doing any more or less work, or that I am a different person when I do one or the other. I guess the "old-fashioned" view is that I am transformed from a landlord to a chocolate maker to a rentier and back again. And I guess I am doing something "productive" when I'm in chocolate maker mode, but not when I am in rentier mode.

As to the other points, my fingers can't type out all the objections now, but I do disagree with the initial observations. In particular, it cannot be the case that bank creditors are in a first loss position but that in the end government is on the hook for everything. It is either one or the other. There are 9 Trillion of financial sector credit market debt that is not FDIC insured. More than enough to absorb any losses without the government being on the hook for anything.

About the cost of fund observations, I'm familiar with some of this work, and it assumes the conclusion. The underlying assumption is that bank profits are equal to their funding costs, and therefore because banks make excessive profits, they must have higher funding costs. No need to carefully attack this line of reasoning.

But I guess I am also old fashioned in believing that the government can safeguard the payments system without bailing out bondholders, by addressing liquidity rather than solvency risks.

And you can do this by "lending freely against good collateral at penalty rates", as well as supplying FDIC insurance for the 7 Trillion in *depositors* but *not* for the 9 Trillion in bondholders. I.e. pay out only 40 cents on the dollar, and let the credit-markets take the hit for the other 60 cents. If you don't do this, then you really are supplying economic rents.

But you are arguing for "lending freely against *no* collateral at *zero* rates", and that in the end, government is on the hook for 100% of bank liabilities. If you do this, talk of strict capital controls are meaningless -- debts will always be rolled over a la Japan. Such a policy would create enormous rentier profits, misallocation of capital, and price instability.


I'm not saying there is zero opportunity cost. I'm just separating the risks that there are and being precise in my language. The rest of the world exists. But Govt notes promise a nominal payout, nothing more, nothing less.

When you buy a REIT (or whatever) you are making a credit risk decision. When you buy a t-bill you are not. It's the difference between saving and investing.

As for the CoC of banks -- this is public information and anyone can look this up. It isn't 0.25% though.

You and I agree that there are plenty of non-FDIC insured liabilities to soak up losses. The bailout of bondholders has been the #1 scandal in this Orgy of Scandal. I don't think Warren is 100% on point here either. But this fact is congruent with Warren's position that the liability side of the balance sheet does not enforce market discipline.

If you're going to let the bondholders take a bath, and I would, you need to be willing to massively recapitalize the private sector with vertical (govt) money. Obama's unwillingness to do this is what lets the banking sector assume a role of importance.


But the government is *selling* the payout at competitive prices. Why would you think that the buyer of the income stream is receiving a benefit in excess of any other competitively priced asset? Wouldn't they bid up the price of the income stream to its perceived value? That is the risk -- that you are overpaying or underpaying, and it is just as real for treasuries as for any other asset.

But the difference between investing and saving is that when you invest, you are dissaving and incurring a liability. No one purchasing a financial asset is investing, they are saving. Saving is the accumulation of financial assets and investing is the issuance of financial liabilities in order to increase the real stock of productive capacity.

For the CoC, the equity cost of capital is not .25 -- the issue is the cost of debt, which is artificially low. I am not saying is .25, but I am saying that it is too low. To check this, look at all the highly leverage pro-cyclical firms with a AAA rating. Doesn't that tell you that they have artificially low borrowing costs? If not, look at the P&L.

"But this fact is congruent with Warren's position that the liability side of the balance sheet does not enforce market discipline."

I don't see the logic here. Why not say that the asset side does not enforce market discipline? In fact, there is no market discipline, because of the government backing. That may be an argument for government banking, but it is not an argument for government guarantees of bank liabilities, or allowing the banks to have lower borrowing costs than the productive sector.

The reason why Warren is arguing this is that his banking model is such that assets will perform according to an "approved" credit model, and the government will absorb losses beyond that. I.e. -- make reality conform to the model. Liabilities will be free (except dividends). Or maybe .25. Now I can see why a banker would view such a business model as appealing, but I don't. And attempts to get people to support it by saying that this "squeezes rentiers" is at best a misrepresentation.

About the "best" solution to the crisis, it's hard to say exactly. I don't think government injection of funds would be necessary, as you can convert bondholder claims to equity claims and recapitalize that way. What would be necessary is a massive fiscal response, and of course CB lending at penalty rates against good collateral. But it's hard to say -- certainly the government could create new banks if too many failed, and then sell them off later. The individual banks themselves are not important, merely the existence of some banks, and in any case banks don't do much on the lending side anyways. They are important for their market-making B/D faces, which should be stripped off regardless.


I don't think the buyer of a Govt Treasury security is receiving anything in excess of what they have paid. They paid a market clearing price. I'm just pointing out that Treasuries, uniquely, don't carry credit risk (default risk) while all private debt does. And Treasuries carry plenty of other risks as you keep pointing out. It is impossible to synthesize this zero default risk condition in the private market.

In the MMT model (agree or not), saving is the act of not spending income. If your action cannot be booked as income by anyone else, then you are saving, otherwise it is an investment. If you buy a stock, someone else sells the stock and books that as income. If you put your money in a bank, or under your mattress, no one else books income.

I am not going to defend the veracity of credit rating firms re: CoC. I'm just pointing out that banks do not borrow at the FFR! The marginal cost of a loan a bank extends is the marginal cost to the bank of raising the equity (capital) they need to fund that next loan. IIRC it's about 6%.

OK, so we agree that bank liability holders did not care about credit quality at the bank. Switch to assets. When banks fall below the capital requirements, they are meant to be shut down by the FDIC. Whether or not they fall below capital requirements depends on the credit quality of their assets. It is the job of the FDIC to assess assets, and shut the bank down if the assets are bad. There was lousy Govt regulation of this as well -- we agree there too! -- but at least in part this is because the regulators don't believe they should be in this business, it should be left to "the market" and MMT disagrees. This is not an argument for Government banking, nor is it an argument for guaranteeing ALL bank liabilities. It's just an argument for an active FDIC who does not leave things to non-Govt backed liability holders.

I don't know why Warren wants Govt backing of all liability holders either. I would 100% back depositors, and everyone else is on their own. Warren would argue that a bank can always swap its non-FDIC liabilities for FDIC liabilities through brokered deposits, if they were willing to pay, but I don't agree with that.

The massive fiscal response is the key to managing credit bubbles after they pop. You have private credit contracting, and therefore high unemployment, lost real output, etc. Unless you are willing to replace private credit (horizontal) with vertical money, you are always going to be at the mercy of bankers who say "tighten up regulations on us and we can't make as many loans" -- and they are correct!

A CB penalty rate just drains the private sector of net financial assets, precisely at a time when it is short of net financial assets. That's why I actually don't support bank taxes, or even the FDIC "tax" which "pre-funds" the bailout account.

"If your action cannot be booked as income by anyone else, then you are saving, otherwise it is an investment. If you buy a stock, someone else sells the stock and books that as income. If you put your money in a bank, or under your mattress, no one else books income."

Let's think about what this means. NYSE has a daily dollar volume of 30 billion, so this means that people are booking 30 billion of income each day from equities? That's only true if you define "booking income" as something irrelevant to demand or investment. Perhaps you only mean new issues -- but those are frontrun by the financial sector, so if that is what you mean, then the non-financal sector is generating zero income by purchasing assets. So what do you really mean?

Just as government spending creates the income to buy government debt, and as loans create deposits, in the same way borrowers -- private sector liability issuers -- create the income to buy private sector debt. MMT understands the first two, but not the third -- or perhaps they also understand that all three are the exactly same thing. I get mixed messages on that. Perhaps you can clarify whether this is MMT orthodoxy. In my book, this is orthodoxy:

The issuance of liabilities creates the income to buy the original liabilities. A portion of that income goes to the purchase of capital goods, and this is investment in the NIPA sense. Another portion goes to augment the purchase of consumption goods, which would constitute household borrowing for consumption.

Admittedly this is a subtle point, because in order to see that this is the case, you have to look at a flow of overlapping transactions in which the deficit spenders in period 1 are generating income for "savers" to purchase assets in period 2. Then you need a financial sector that can operate via balance sheet expansion to buy the assets in period 1 and sell them to households in period 2 -- the financial sector absorbs any income flow mismatches between period 1 and period 2. Then take the limit as the number of actors grows and the periods overlap so that the distance between them goes to zero, or 3 days -- whatever the payment settlement period is.

That's the easiest way to see how "loans create deposits" or "dissaving creates assets to fund the dissaving". This is an emergent phenomena for an economy that can operate via balance sheet expansion, and has nothing to do with being on a gold standard or central bank operations. The latter only serve to prevent the payment crises that are endemic to this type of arrangement by keeping assets liquid -- e.g. taking over some of the role of banks when the latter fail, or alternately if you are pro-banking -- preventing the banks from failing.

"I'm just pointing out that banks do not borrow at the FFR! The marginal cost of a loan a bank extends is the marginal cost to the bank of raising the equity (capital) they need to fund that next loan. "

OK, with a leverage of 12:1, and an equity cost of 6%, that is a marginal cost of 0.5%, not 0.25%. With leverage, you cannot rely on the equity costs alone but must have an accurate cost of debt. If the debt was correctly priced, then increasing leverage would not result in increasing profits (on average), but would result only in increasing volatility of profits. But banks are able to earn higher profits than their cost of equity demands, and so they do not need to pay out all their profits to either equity holders or debt holders. Instead they pay out enormous bonuses and compensation to themselves. This is a *proof* that their cost of debt is too low. These excess profits are the definition of an economic rent. Any attempts to further lower the (already too low) cost of debt will increase the economic rent.

FFR comes in because it lowers the banks cost of debt -- but you are right, it is not the only factor. FFR, FDIC, lower leverage ratios -- these are all fees that we are free to impose, and there is nothing fundamental or endogenous about one or the other -- they are regulatory devices in order to ensure that the creation of credit is not costless. The fundamental, endogenous thing is 1) that the economy will be distorted if banks have lower costs than non-banks and 2) the non-bank capital costs are endogenous, so that lowering bank cost of capital will not lower non-bank cost of capital, but will instead destabilize the economy.

Therefore we are not "free" to adjust bank costs however we want, based on our favorite political economy rule of thumb -- we need to keep bank-costs in line with non-bank costs, and *this* is how you minimize economic rents.

"The massive fiscal response is the key to managing credit bubbles after they pop."

I agree that a massive fiscal response is required. But it is not the *only* thing that is required. A credit crisis does not happen because the government issues too few "Net Financial Assets" to the private sector -- that is not an economically meaningful metric. Over the business cycle, the government is only required to increase government liabilities in order to prevent deflation in the face of growing output. The crisis happened because assets were mispriced, and could not deliver the return expected of them. It is just as important that misallocators of capital suffer losses -- particularly the government protected ones -- as it is that demand be maintained. You need both.

The government's job is not to guarantee income (in the flow of fund sense), but demand (in the NIPA sense). And that means that less income goes to the financial markets (via defaults and restructuring) so that the level of income flowing to the goods markets is maintained. The government needs to not only support demand with a fiscal response, but simultanteously oversee defaults and restructuring that cause a decrease in household financial assets (e.g. the bond and equity holders of the bad debt), oftentimes forcing organizations into the restructuring, and imposing penalties on government protected institutions that misallocated capital.

In no case should the government guarantee that the same level of income flows to asset holders during the adjustment, only that aggregate demand in the goods market be maintained. In order to see this distinction, you need to recognize that income flowing into the financial markets is not the same as income flowing into the goods markets -- e.g. that those purchasing financial assets are not "investors" in the NIPA sense, but savers. Then the "excess savings" is reduced by having the income of financial asset holders reduced, while the government runs a blocking operation in the goods market. To the degree that the government just guarantees demand without imposing penalties on asset holders, then the fundamental imbalances are not addressed, and capital misallocation as well as unsustainable income patterns are allowed to persist.

" Rentier capitalism is a term used in Marxism and sociology which refers to a type of capitalism where a large amount of profit-income generated takes the form of property income, received as interest, rents, dividends, or capital gains.

The beneficiaries of this income are a property-owning social class who play no productive role in the economy themselves but who monopolise the access to physical assets, financial assets and technologies. They make money not from producing anything new themselves, but purely from their ownership of property (which provides a claim to a revenue stream) and dealing in that property.

Often the term rentier capitalism is used with the connotation that it is a form of parasitism or a decadent form of capitalism. "

RSJ: In MMT, "saving" is best defined as "not spending". When I sell a share, I get money in return from the person I sold it to. That person has spent money, and therefore is not "saving" even if they hope to cash in the share when they retire.

I think you're talking about the "S=I" identity here. The MMT interpretation is that investment creates saving (so, it is not the usual loanable funds causality where savings fund investment). But you may not be talking about this -- what do you mean by "private sector liability issuer"?

Coming back to CoC, I'm not sure the WACC is the right margin to be looking at. If a bank is capital constrained and needs to raise more in order to make additional loans, it needs more equity, and the marginal cost of raising that is 6% regardless of what the WACC will be once it's done levering up.

We are actually in agreement re: the correct response to credit crises. My point here is that the Govt wants to maintain private sector demand through private sector credit expansion, and therefore is making a distribution decision towards the financial sector. Since they are being advised by economists (who, like Nick Rowe, think that bank lending is reserve constrained etc. etc.) they will not just recapitalize the private sector as a whole through something straightforward like a payroll tax holiday funded by deficit spending.

I mean seriously, why is FDIC not unlimited? It's a joke.

a: MMT is mostly accounting and operations. It is not political.

MMT is Marxist

a: MMT is a lot of things, not all of them good, but Marxist is not one. Unless you feel double entry bookkeeping is Marxist too (which you might)

Definitely not Marxist. I've never met a Marxist who stressed the revolutionary role of accounting for laying bare the inherent contradictions in capitalism. Would make a good Monty Python sketch though.

you've both missed the point

its not about double entry bookkeeping - that's only a mask, and they admit themselves there's nothing mysterious about correct accounting

its about the ideological agenda

forcing the deficit through the banking system and forcing treasury rates to zero is Marxist

(they use accounting to illustrate it)

- euthanasia for the rentier


Absolutely agree that FDIC insurance should be unlimited. I think, in practice it is, as they claim to have fully re-imbursed everyone. But it should be officially unlimited.

"what do you mean by "private sector liability issuer"?"

LOL. Forgive my tortured prose. I mean "borrower", but wanted to be sure to include equity issuers in that group as well. Anyone in the private sector that incurs a liability, regardless of the repayment terms.

"In MMT, "saving" is best defined as "not spending". When I sell a share, I get money in return from the person I sold it to. That person has spent money, and therefore is not "saving" even if they hope to cash in the share when they retire."

Yes, that person has spent money -- to buy another share! You have no way of knowing that they spent money to purchase goods. Maybe they did, maybe they didn't. Either way, looking at "spending money" won't tell you. So that is not a good definition.

In other words, you can't assume that the balance sheet identity:

(1) change in liabilities = change in assets,

is the same as the NIPA identity:

(2) Savings = Investment

A better definition is to assume that for each agent,

money in = money out = money spent

And your choice is of whether that money was spent to purchase financial assets or goods. Now Nick will say that people will refuse to spend on financial assets or goods, and just hold cash under the mattress. I don't think that's what happens in a modern economy. But this depends on the institutional arrangements of the nation being studied. For us, it's enough to assume that all income is spent in the financial markets or goods markets. And any income spent on the purchase of financial assets can be viewed as "saving" in the balance sheet sense, regardless of whether those assets are treasury bonds, corporate bonds, etc.

But it is *not* the case that money spent to purchase financial assets reflects real investment demand. It could reflect ponzi borrowing with zero additional purchases of goods. It could reflect rising asset prices with zero increase of "real" asset purchases. It could reflect falling wage shares while goods purchases are constant. Usually, it will be some combination of all of the above.

In that case, additional government income boosts -- say via a tax cut -- can result in increased financial asset purchases, but *not* an increase in NIPA demand, or a lower multiplier than would be the case if less money was spent on financial assets.

A very simple model is to assume that a bubble economy is one which the proportion of income spent on financial asset purchases is too high relative to the proportion of income spent on the purchase of goods (whether those be capital or consumption goods). What is "too high"? If Y% of Income is spent on financial asset purchases, then it better be the case that the economy is growing at Y%. If the economy is growing at .6Y%, then P/E multiples are swelling and we are in an asset bubble -- most likely due to income inequality -- requiring ever greater government income support to maintain even a constant level of output -- i.e. Japan, the poster boy of a nation that suffered the greatest median wage share loss of any industrial country.


In that case, when the government boosted incomes, they found that those incomes ended up in the hands of the wealthy who used them to purchase more bonds, while consumption demand continued to stagnate. The distribution of income in Japan is what makes deficit spending ineffectual for boosting output. And Japan is nearing 3 decades of requiring constant *increases* of government income support to keep NGDP from falling below 1990 levels, as a high proportion of government spending is immediately rolled over into financial asset purchases and does not stimulate output.

The Dutch republic was stuck in this trap for 150 years.

Earlier on, Nick was musing about the "bubble" economy.


The bubble economy is possible because (1) does not equal (2). You can have a huge swelling of financial assets without a 1-1 increase in investment and output in the NIPA sense. The question is -- can you have a huge deflation of financial assets without a huge loss of output? Mosler says no, and we must supply households with all the assets they "desire". And if we do not do this, then the lack of wish fulfillment for wealthy households is what causes unemployment. I claim the latter is difficult, but possible, and requires simultaneous government purchase of output while also shrinking, rather than increasing, private sector assets.

The best way to do this is with tax policy as well as forced bankruptcies/restructurings. The only period in U.S. history where we experienced rising median wage shares outside of deflations was in the high tax era. As soon as that era ended, median wage shares began a continuous and rapid decline.


Whereas those nations that manage to recapture the bulk of their deficit spending via progressive taxation have avoided declining wage shares while maintaining demand and output growth.


So the goal of government should never be to prevent asset deflation, but to suppress the ratio of income spent on financial asset purchases back to a sustainable level, and doing this means that less government income support (e.g. deficit spending) is required.


Thanks for your clarification. I still don't think we're on the same page here.

Consider two scenarios.
1. Someone takes some income and uses it to buy a stock in their 401(k)
2. Someone takes some income and simply leaves it in the bank

In scenario 1, that person is investing, as the individual who sold the stock gets income from the transaction.
In scenario 2, that person is saving, as there is no entity who get income from that individuals action (or in this case, a non-action). The bank does not book its larger liability (deposit) as "income". I'm not sure if you would count putting money in a checking account as purchasing a financial asset.

The reason MMT defines "saving" as "not-spending" is because in common parlance, buying stocks in a 401(k) is considered saving for retirement. That's a fine concept, but does not make the distinction MMT needs for its interpretation of S=I. In a period, net, some income is not spent. That's your S.

Bubble economies aren't mysterious to me, and those graphs were pretty scary! I would love an experienced, insider look at Japan from someone who knows MMT. If you have any suggestions, I'm all open. I'm not sure to what degree Japan ever really used tax cuts, nor do I understand why they tax today. Goldman is selling some pitch that Japan's going to default any day now btw.

I'm not sure your characterization of Mosler/MMT is fair or accurate when you talk about income distribution. Mosler says you should deficit spend to reduce unemployment when the economy is under capacity, and inflation is not a concern.

Yes, not yet on the same page!

First, yes, I would count making a deposit in a checking account as purchasing a financial asset, although in practice, people only do this if they plan on spending the money in a short period of time. A checking account is really an inventory or buffer stock, and can safely be ignored. But I would view it as the purchase of a financial claim.

I am trying to make the distinction that an increase in financial assets (let's call this "financial savings") and the corresponding increase in financial liabilities (let's call this financial "dissaving") is not the same as the NIPA savings and investment. It does not map well to the the NIPA S = I. You are claiming that it does map well, and that your mapping consists of the instrument purchased. I am saying that all instruments (of a given maturity) trade at indifference prices and are substitutable, and therefore such a mapping cannot exist.

Now let's look at your transaction.

"1. Someone takes some income and uses it to buy a stock in their 401(k)
2. Someone takes some income and simply leaves it in the bank"

-- Did they get the income by selling some other financial asset first? Or did they get the income from the goods market (by selling labor or goods)?

"In scenario 1, that person is investing, as the individual who sold the stock gets income from the transaction."

Is the person who sold the stock using the income to buy another asset (in the financial markets) or to buy a consumption good, or an investment good? How do you know? The *instrument sold* will not tell you whether NIPA savings, investment, or consumption is occurring.

Moreover, you are associating the funding of an activity with the performance of the activity. This is a loanable funds argument -- e.g. the depositor causes the loan to occur, and therefore by making a deposit, I am "lending". In the same way, you argue that the buyer of the stock causes the investment (or consumption, or rollover) to occur, and therefore by buying the stock I am "investing".

Her is an alternate set of definitions:

Those who increase their stock of financial assets are saving, and those who decrease their stock of financial assets are dissaving -- or spending. I.e. the *net sellers* of financial claims are the dissavers, and the *net purchasers* of financial claims are the savers. The instrument bought or sold is irrelevent. And whether the net dissaving funds consumption or investment -- you still don't know. But at least this is a consistent definition.

"I'm not sure your characterization of Mosler/MMT is fair or accurate when you talk about income distribution. Mosler says you should deficit spend to reduce unemployment when the economy is under capacity, and inflation is not a concern."

RIght. In other words, no awareness that demand and deflation is a mechanism the economy uses to boost median wages back to levels where there is sufficient endogenous demand. You cannot ignore wage inequality and pretend that households are spending less for irrational reasons, rather than because their wages are too low. Any attempts to fight deflation and maintain demand -- which I support -- need to be met by equally vigorous attempts to boost median wage shares. If you only do one and not the other, then you are just guaranteeing the business sector that their inventories will clear, regardless of how little they pay their employees. Government will step in and make up for the demand shortfall, with no attempts to fix the underlying distributional problem.

Such a policy will only push wages lower, as deflation is the only mechanism that the market has to keep median wages from falling excessively. I am not arguing for deflation, but for understanding what purpose it achieves, and therefore if we prevent the market from righting itself, we need to also achieve the same purpose artificially. Mosler is trying to short-circuit this, and create a world in which there is zero connection between how much people are paid, and how much businesses sell.

RSJ: If I take out a loan, I increase the stock of (gross) financial assets and I have dissaved by taking on debt. This is contrary to your assertion. At a macro level, it is not possible for the sector to increase its net financial assets (although of course it can increase its real assets).

And I certainly am not making a loanable funds argument, as banks do not lend out deposits. Here's how MMT tackles the savings/investment/consumption NIPA identity:

In a period, an economy produces some quantity of real goods and services. It also pays itself a (nominal) wage for that activity. It consumes some quantity of its period output, leaving the rest available for the next period as inventory or capital goods.

The increase in inventory and capital goods available for next period is "investment". The quantity of this investment must equal the income earned (by producing the goods) but not spent (on consuming the goods in this period). That's "savings". Money in the bank is "dead money", all it does is leave room for output to be rolled over into next period.

So savings does not "direct" the investment, it simply leaves output unconsumed so it's available next period. There two must be equal, so S=I.

A purchase of any kind triggers income, so it cannot be "savings". Money in the bank does not trigger income -- certainly the bank does not think so! -- and thus counts as "saving". So, incidently does money under the mattress etc.

Re Mosler, I think he would agree that households are spending less because their wages are too low. He would say that the Govt is taxing too much, leaving households unable to purchase their own output, and not stepping in to buy the excess capacity that creates. Median wage is a pre-tax measure, but purchasing power is post-tax, and Mosler's suggestions absolutely increase that. Taxes render the private sector unable to afford its own output. Either the Government steps in to make up for this shortfall, or you have the AD issues we see.

And whether or not the Govt stepping in and making up for the demand shortfall increases or decreases the median wage depends entirely on what price the Govt pays for its purchases.

Finally, I agree with you that income distribution within the sector can impact that sector's total desire to increase net savings (and thus, demand). There are a whole host of political and microeconomic incentive issues that come up with income distribution though, Mosler does not talk about them a lot and I don't blame him! The basics of MMT are fundamental enough and not understood enough that extending the discussion to this can be left to MMTers who find it interesting.


which part of "how MMT tackles the savings/investment/consumption NIPA identity" do you suppose is different from the way standard macro tackles it?

Or at least, which part of what you describe do you suppose is different from the standard story?

AdamP: Standard macro, in my recollection, has a loanable funds market, where savings are doled out as investment.

But since banks don't lend out deposits, this is incorrect.

MMT interprets S=I without this market in the middle

" If I take out a loan, I increase the stock of (gross) financial assets and I have dissaved by taking on debt. This is contrary to your assertion. "

The person who takes out a loan to buy a car has his stock of financial assets reduced, as he has a new short position, therefore he is dissaving. That is my assertion. The person who takes out a loan to repay some other loan, or to buy another financial asset, did not increase or decrease his financial assets, and therefore contributed nothing to NIPA demand. Zero dissaving. That is also my assertion.

My assertion is that you need to look at each household's balance sheets -- at changes in net financial assets for each household, in order to determine how much that person contributed to saving or dissaving. Even then, you do not have a clean relationship between that and NIPA consumption or investment, but at least you net out the bulk of the income flows that are portfolio shifts.

Your contention is that you only need to look at the *instrument* bought or sold -- i.e. that the making a deposit in a bank is saving, but purchasing a mutual fund share is investing. Your own contribution to NIPA demand is exactly the decrease in your financial net worth, regardless of whether your deposits decreased, bond holdings decreased, etc, or some complex combination of the above.

There is no relationship between "income flows" and NIPA demand. Putting money into a bank is an income flow, just as putting money into your 401K. Leaving money (e.g. not doing anything) is not an income flow, neither in your 401K, not in your bank. There are many trillions in daily income flows, almost all in the financial markets, and very few of them do anything to contribute to NIPA demand. If you want to try to measure the latter, then you need to look at changes in the balance sheets of households, rather than trying to track specific instruments.


"standard macro" does not have savings doled out as investment, but there is a "supply curve" of savings, and a "demand curve" for investment, with the interest rate equilibrating between the two, so that the S = I.

The real problem with loanable funds is not that it assumes that savings "causes" investment, but that it treats financial assets as if they were consumption goods. It assumes two separate populations of producers and consumers, with an increasing marginal cost of production as well as a decreasing marginal utility of consumption.

But as soon as it possible to expand your balance sheet -- i.e. to borrow in order to lend, then you no longer have a supply curve that shifts with quantity. It *may* shift with quantity, but the direction of the shift is unknown:

If there is a surge in borrowing to buy financial assets, then that could result in falling real yields, so that interest rates decline with the quantity borrowed. If there is a surge in borrowing to buy capital goods, then that may also result in falling real yields with the quantity borrowed. If there is a surge in borrowing to buy consumption goods, then there may be an increase in profits (and real yields), so that the rate climbs with the quantity borrowed.

Whatever the rate of return happens to be, as long as there is an investment opportunity that meets or exceeds that rate, then it will get funded, regardless of how much people have 'saved'. Whether or not the funding of that investment lifts the rate of return demanded for subsequent investments will depend on how the investment changes the expected profit opportunities in the economy as a whole; there is not a finite stock of funds that is depleted, forcing yields to increase as the existing pool of funds runs out.

RSJ: A person who takes out a loan to buy a car has *increased* their balance sheet. They now have a new (real) asset -- the car (valued at something), plus a new (financial) liability -- the loan itself. I don't see this as saving.

A person who buys a car out of cash debits their cash account (asset) and now has a real asset (car). They are clearly (nominally) dissaving, although at a macro level no new (net) financial asset is being created. We're just shifting an asset from a buyer to a seller. NIPA, I believe, is at the sector level.

If you look at a household level at net financial asset change, it will tell you something about that household but nothing about the sector. At a sector level, there can be no change in net financial assets unless the Govt deficit spends.

The problem with the standard macro S=I are legion. I agree with your point that savings does not "cause" investment, the way the standard Macro story is told, but I think it is also true that the narrative embedded in the "loanable funds" model is that money saved gets loaned out for investment.


"A person who takes out a loan to buy a car has *increased* their balance sheet."

They have decreased their *financial* assets, exchanging them for real goods. They have dissaved. This is how people spend money, their stock of financial assets decreases and their stock of goods (whether for consumption or investment) goes up.

Really, this is a tautology -- and the contortions required to equate an increase in your net financial assets with dissaving is massive. An increase in your net financial assets is saving, and a decrease is dissaving -- this is simple stuff.

As a result, there is no difference between increasing your financial assets by increasing your holdings of treasuries or deposit accounts or private bonds. In all cases you are saving. And when you reduce your net financial assets, you are dissaving.

"If you look at a household level at net financial asset change, it will tell you something about that household but nothing about the sector. "

What exactly did you mean when you said that someone purchasing a private sector bond is investing but someone purchasing a government bond was saving? You were making a claim about a household, and in order to challenge that view, we have to stick with that household. Yes, the behavior of that household does not tell you anything about the sector, nation or world economy.

All I did was point out that the bond purchase or sale could just be a financial sector portfolio shift, or a balance sheet expansion, and at the end of the day, the instrument held did not matter -- what mattered was the change in financial assets held by that household, under the simple assumption that money in = money spent on either financial assets or goods.

All of this should be orthodoxy -- I think you just aren't connecting the dots because your intuition "buying a private sector bond is investing, but buying a government bond is what rentiers do" is clouding your ability to apply definitions consistently.

RSJ: Yes, we agree that the individual who borrowed money to buy the car has spent money, and reduced their (individual) NFA, but the individual who sold the car has credited their financial asset. At a sector level, of course net financial assets are unchanged, but financial assets have increased as we have a new asset (receivable/loan) and a new liability (deposit/deposit) in the sector. I have repeatedly said that NFA is unchanged (as of course it must be), but gross financial assets have increased (at a sector level) because someone has taken on a new loan.

This is not a contortion, it's very basic horizontal money (credit) expansion. Private sector credit extension creates no new net financial assets (it cannot) but does increase gross financial assets. And I do not count a sector taking on (gross) debt as "saving" because it clearly is not.

When someone purchased a private bond they increase private sector credit (by taking the other side of that trade) and make a credit decision. When someone purchases a Govt bond they are not making a credit decision (although, as you say, they are taking on all kinds of other risk) AND they are not changing the quantity of private sector credit, or the quantity of outstanding Govt liabilities.


OK, Good. I think we agree that dissaving is a decrease in the household's NFA, and saving is an increase in the household's NFA. Excellent progress so far.

"At a sector level, of course net financial assets are unchanged"


When looking at the spending decisions of a household, the NFA is of the household. It certainly changes.

When aggregating households, the economically relevant "sector" is the household, or personal sector. In that case, NFA will increase if the matching liability is outside this sector -- primarily issued by a business. This is what measures the real wealth of households. And for most questions, this is the appropriate sector to look at.

If you start aggregating further, you start to lose information and the numbers start to be less and less relevant to questions of consumption, investment, prices, and aggregate demand. Once your "sector" is the earth, then NFA never changes and you have zero relevance. Already by the time you aggregate to the "private" sector, you are already citing mostly meaningless numbers.

The MMT obsession of arbitrarily defining net worth as aggregated across the entire private is a huge blunder. It means that capital growth is never measured and is never reflected in household net worth. Household net worth is also not properly measured. All the claims of households on businesses cancel with the liabilities of the business and what you are left with is an economically insignificant residual, one in which the value of capital completely disappears. It is not as dumb as looking at the NFA of the entire earth, but it's almost as dumb. Then enormous importance is placed on this residual, and bad economics follows as a result.

To elaborate further, Net Financial Assets is just assets minus liabilities of the balance sheet you are looking at. Then you start to consolidate the balance sheets of various economic actors and you get an NFA of a larger group.

In the process of that balance sheet consolidation, you lose certain information, namely the assets that members of the group hold that are matched by liabilities in that group. So as you do this, your model loses the dynamics of spending shifts due to one subgroup borrowing from another subgroup. Whether that is important or not depends on how much you aggregate and whether these shifts are important economic drivers.

This is the exact equivalent of selecting a "representative agent". But even the simplest macro models realize that their representative agents cannot be *both* businesses and *households*, because if you do that, then you have no model of production or capital accumulation. Even the simplest infinite time horizon models operate at the level of representative households working for and holding claims on representative businesses, and interacting with the government and foreign sectors. Better models have more sectors and overlapping generations of households and businesses.

But when you just look at the consolidated balance sheet of the entire private sector, then there are no businesses or households, so you lose all information about capital growth and production -- that is *excessive* aggregation, and does not give a meaningful measure of household spending decisions or business investment decisions, since you can't measure any of these numbers by looking at the aggregate balance sheet of the private sector.

If you want that type of information, you need to limit your balance sheet consolidation to households -- at the very most. And it would be really better to disaggregate a bit further and consolidate households by income level, or occupation, and look at the NFA of each.

Then you can talk about sectoral balances between households and financial or non-financial businesses, or between wealthy and poor households, and get something that is both consistent from an accounting level as well as modeling something that is economically significant -- the process by which capital grows and earns a return.

But when all of that is aggregated away, then you start thinking that capital grows at an exogenous, government controlled rate, and that household net worth is exactly equal to the stock of government bonds and currency. And then you start believing that saving can only be defined as the accumulation of government liabilities, and that leads you to start making all of the harmful policy recommendations.


I disagree. The right way to carve up a sectors depends on the question you're trying to answer. You lose somethings when you consolidate households and businesses, but you also gain some things. The key (for MMT) being the difference between a currency issuer and a currency user. The confusion of the economics progression around this distinction is absolutely central, and I think it's perfectly reasonable to try and explain this by splitting the two into two difference sectors.

MMT goes even further, btw., and lumps foreign Government and state and local Governments in with households and businesses, which is why you sometimes see that sector called "non-Federal Govt". But it is crystal clear on who can issue currency, and who cannnot, something you see confused every day in the pages of macro text books and, by emanation, the pages of the NYTimes.

I also don't agree that this issuer/user distinction is less relevant, dumb, etc. The residual it leaves is thought unimportant by economists, but it is causing a 10% unemployment rate which, in my book, means it merits some importance. I think you put it best elsewhere when you said economics believes that banking is both utterly transparent and irrelevant, while at the same time being indispensable!

I don't know why you think this distinction destroys the value of capital. I don't think that and I'm very comfortable with the distinction. You may be missing the real vs nominal distinction (something else that I have found MMT is absolutely crystal clear on). Non-Fed NFA precisely equals Fed debt, but non-Fed NA is different, and that's where you have your real goods. No one claims that the private sector can't increase it's stock of real goods by itself, of course it can, but it cannot increase its stock of net financial assets.

But when all of that is aggregated away, then you start thinking that capital grows at an exogenous, government controlled rate, and that household net worth is exactly equal to the stock of government bonds and currency. And then you start believing that saving can only be defined as the accumulation of government liabilities, and that leads you to start making all of the harmful policy recommendations.

I dunno RSJ. I divide the world into issuer/user and I don't reach the conclusions you do.

I think we're done. Thanks for the good discussion! I learned a lot about the yield curve.

It was an enjoyable discussion, W.

In fairness, I used the word "dumb", not "stupid":)

That was maybe too over the top, but the point is that it is a residual. The actual decisions to spend or not spend, or the causes of unemployment, cannot be reduced to just looking at the quantity of government liabilities, as the decision makers (e.g. households and businesses) do not see private sector NFA, they see their own NFA, which necessarily includes changes to bank deposits, bonds, and equities -- none of which are government liabilities, or even move in the same direction as government liabilities.

Of course you can argue that if the government deficit spent to hire all of the unemployed, then there would be no unemployment, and from that you can conclude that the unemployment is "caused" by a lack of deficit spending. But I think this is misleading.

We had lower unemployment during the "30 glorious years" from 1950-1970 than we did in the period from 1980-2010. But the former period had government debt/GDP decreasing rapidly, and the latter period had government debt/GDP increasing, meaning that private sector NFA was low in the low unemployment period and higher in the higher period. So in order to really understands what drives unemployment, you can't focus on the residual, but need to disaggregate and look at the various NFAs of the households and businesses. Anyways, good discussion.

The comments to this entry are closed.

Search this site

  • Google

Blog powered by Typepad