Forewarning: This blog is generally about monetary theory. This post is going to be unusually loaded in political content, because this stuff is required to lay down the foundation of the framework that explains why using fiscal policy to regulate the price-level is a horrendous idea. Moreover, I do not have a background or any formal training in political science, so for all I know I’m repeating other people’s work: as usual, I make no claim of originality.

Democracy is, by nature, imperfect. Without a strong constitution, it’s a dictatorship of the 51%. Here are a couple of extreme examples that illustrate the shortcomings of a pure majority rule:

1. If 51% of the population has blue eyes while 49% has brown-eyes, the blue-eyed people could in theory vote to enslave the brown-eyed folks and take all their wealth
2. The 51% poorer can vote to take a huge share of wealth from the top 49%

In practice it’s difficult for a 51% majority to impose such outrageous will on 49% (because the oppressed would become violent and it would be too costly to keep them under control) but once you get in the sub-15% range you’re in business. The blacks, the jews, the unskilled unemployed (think minimum wage law), the gay, have all gotten screwed by a democratic majority.

As a minority, your vote basically doesn’t count because voting is a winner-takes-all game. Therefore when things get untenable, the only thing left to do is to with vote with your feet–and that’s assuming you’re even allowed to do that (think East Germany).

Politicians get away with retarded policies because there’s a great deal of friction which prevent folks from voting with their feet:

1. There needs to exist a better place than the one you’re in now and that place has to be willing to take you in.
2. There’s a big cost to moving, selling your house, moving your stuff, possibly having to learn a new language, uprooting the family, etc.

When politicians want to get away with even more retarded policies, they have to increase the friction for getting out by forcibly locking people in or through propaganda.

Enters Seasteading.

Seasteads are a mean to reduce these frictions to their practical minimum. Essentially the idea is to build floating cities outside the jurisdiction of the majority-controlled land where minorities could flourish, new political systems could be experimented with, and the liberty to vote with your feet would create ideal conditions for the betterment of political systems through competitive pressures.

1. The availability of alternative seasteads is only limited by the resources we devote to build them and the surface of the sea
2. The cost to move is basically that of towing your entire floating house from one city to another

The Seasteading Institude is working to make Seasteads a reality under the leadership of none other than Patri Friedman (Milton’s grandson). While watching the Seasteads evolve will provide very interesting data not to mention an opportunity–for a handful of people–to live free of the majority’s oppression, we don’t have to wait until they’re actually built to use them for the purpose of analysis: every political proposition should be evaluated in terms of how it would perform in a world with a large number of seasteads.

Let me try to describe this analysis in semi-formal terms:

Definitions: An competitive political environmentis a world where a large number of seasteads exist, new ones are built all the time, and they continuously experiment with every possible variation of all the possible political systems. In this world, moving between seasteads is costless. Some fraction of the political systems will manage to sustain for meaningful length of time without major modifications to the rulesets: we’ll call those stable political systems.

The “Seastead test” for a political system: what would happen to a seastead implementing this political ruleset in a competitive political environment?

(Yes, this is a vague question)

Now, this is still a monetary theory blog, so let’s use this to analyse what we’re interested in: deficit spending. The question we’re asking is: What would happen to a seastead in a competitive political environment if it were to issue debt to pay benefits to its residents?

To keep things separate and simple, we’ll assume no redistribution and no foreign bond purchases: for instance, there are 1,000 people on the seastead, all of them produce the same output, pay 10% of it in taxes, receive 20% of it in benefits, and “save” the remaining 10% in bonds.

Note that they don’t consume any more than if they were paying 20% in taxes and saving 0% in bonds so the cliché of “living beyond their means” doesn’t apply here. The only difference is purely financial.

Risk adjusted real yields on the government bonds must be in-line with those in the rest of the world so we’ll assume our seastead is small enough to be a pure price taker on that. As the debt grows larger (at the rate of 10% per year), the real interest cost as a fraction of tax revenue also gets bigger.

Let’s now assume that one of the residents decides to move to another seastead for an exogenous reason. Now all of the sudden, the 999 remaining residents are paying interest to themselves, plus 1/999th of the interest the leaver receives. Since we’re in a competitive political environment, there exists somewhere a stable seastead identical to this one except that it decided to keep a balanced budget and charge 20% taxes to its residents.

In that alternative seastead, residents get to spend 80% of their income plus get the remaining 20% in public services. In their current seastead, they also get to spend 80% of their income but only receive 20% minus the share of interest paid to the leaver in public service. Instantly, all 999 remaining residents move to that other seastead and the original seastead is left with tons of debt and nobody to pay taxes. A default is now the only option before the seastead can be salvaged and used again–this means that the system was unstable.

This is the part where MMT proponents sing their chorus: “If the bonds are denominated in domestic currency, solvency is never the issue, inflation is”. Fine, so the seastead inflates its currency into oblivion–it makes no difference: the bonds are worthless either way. MMT comes back again: “if inflation is too high, raise taxes”. Well, there’s no one left to pay any taxes. “Seize the houses of the residents, it doesn’t matter who owns them”. The houses have already been towed away.

“But your seastead example is not realistic, in a real country people and capital can’t just float away”. Yes in the real world things will happen much more slowly and much less visibly. People and capital will trickle out. More importantly we’ll never see the capital that would have come and never did, the investments that would have been made but never were. Any political system that takes advantage of the frictions-to-voting-with-your-feet which exist in the real world, but not in a competitive political environment, is oppressing its people and I reject it, not because it violates some economic welfare optimality condition, but simply on ethical grounds.

Of course if the agents in this example were “more” rational, they would have seen this coming from miles away and would have never agreed to purchase the bonds after some point. What point? The point where the value of the common parts of the seastead started to exceed the value of the debt, in other words, when the seastead government became negative equity. That’s because at that point, the cost of building a new seastead, or of renting an existing one, is going to be equivalent to servicing the interest on that debt. This leads to the following conclusion on fiscal policy:

A government should always maintain stable value for its public equity (value of public assets minus public debt)

This is not a balanced budget condition: if the seastead government invests into a wind turbine meaning it no longer needs to pay electricity to the private sector to power the public facility, it’s ok for it to issue debt up to the value of the turbine: the residents are indifferent between paying taxes to service the interest of that debt or paying taxes to pay an electricity bill in an alternative seastead.

Governments can, and should, issue debt when they make investments. Governments should never use debt to pay benefits.

I believe in (1) nominal price/wage rigidities, (2) NGDPL targeting and (3) the superiority of monetary policy over fiscal policy to achieve that objective (4) even at the zero-bound (well, assuming we nuke the zero-bound).

It’s likely that (1) makes me ineligible for membership in the Austrian family as well as the neo-classicals. (3) made me persona non grata with the MMT crowd (I’m now getting censored there..), and (4) probably means I’m not a new-Keynesian. While I’ve disagreed over central banking mechanics with the market monetarists in the past, I thought it was as a secondary issue, trumped by the large overlap of ideas. Scott Sumner might have just changed that with is attachment to quantity of money:

Suppose a central bank sees inflation rising too fast.  They might decide to raise interest rates.  But higher interest rates don’t lower inflation, they raise inflation.  That’s because as interest rates rise, velocity rises.  Higher velocity raises inflation, it doesn’t lower it.

[..]

Let me repeat, higher interest rates are inflationary.  [..]  Higher interest rates are inflationary.  Repeat 100 times.

There are two key overnight interest rates relevant to a central bank: the rate at which money is being lent risk-free (roughly speaking, the Fed Funds rate), and the rate being paid on money by the central bank: the IO(E)R. The spread between target rate and IOR is effectively what central bank is charging for its intermediation service.

When that spread goes up (keeping the lending rate constant), agents will attempt to reduce their use of the CB, for instance by trading private IOUs (which can still be denominated in the unit of account): money supply would go down, and velocity would go up presumably keeping NGDP roughly constant (assuming the target rate is kept fixed and IOR decreased).

On the other hand, if the target rate is raised (keeping the spread-to-IOR constant) beyond what’s expected, that’s purely deflationary (everything else remaining constant). To see this, freeze the inflation expectations for a moment; by raising the (nominal) target rate, you’ve just raised the real rate of return on money (and incidentally the real cost of borrowing money) which makes money more desirable to own–which is deflationary. The lower inflation expectations only contribute to make the real rate of return on money higher and so on.. What breaks this infinite loop is the expectation that at some point the CB will course-correct (otherwise you’d have a violent deflationary spiral).

It is true that quantity and interest rates are intertwined, but the more relevant variable is interest rates. That’s a bit wishy washy, so let me make a crisper statement:

The CB can always achieve any price-level target while keeping the money supply fixed (using the target rate and the IOR/target spread jointly)

In fact, I’m proposing that the CBs target NGDPL while keeping the money supply at a fixed fraction of GDP here.

Consider the following extreme example: assume the Fed announces tomorrow that IOR is -20% and Fed Funds target is -5%. The result will be: (a) a pickup in NGDP growth AND (b) a contraction of the monetary base. (It goes without saying physical currency would have to be exchanged against digital currency at an exchange rate that drops 20%/year as per IOR to prevent money-under-mattress)

At the zero-bound, or when IOR is left to float along with the target rate, varying the quantity of money does nothing to the price-level, as argued here and here.

I’m not sure what defines a market monetarist but in light in the above theoretical disagreements, and despite my admiration for Scott’s impact on the promotion of NGDPLT, I probably don’t qualify as one. Looks like I’m now left with no school of economic thought to belong to

Update: Scott continues to bury himself into incorrect monetary mechanics:

Under our pre-2008 system the monetary base earned no interest.

Correct.

Thus nominal rates became the opportunity cost of holding base money.

So far, so good.

Higher rates meant less demand for base money, which is inflationary.

Wrong. { Interest Rate Target, IOR, Monetary Base }: pick two, the market gets to set the third! Scott is implying that the central bank moves the IR up, IOR stays at 0%, and the monetary base stays the same! Higher rates means less demand for base money, which just means there is less base money.. M goes down, V goes up by roughly the same proportion leaving NGDP just about unaffected (unless V was so high that the amount of medium of exchange is now insufficient to carry out transactions–but we’re generally very far from this)

In any case, demand for base money is irrelevant to inflation, only demand for “safe assets” matters! (safe assets are assets whose prices are relatively stable in terms of the unit of account–including but not limited to base money)

Update #2: As per JP’s comment regarding the often cryptic nature of Sumner’s posts, I spent more time trying to understand what Scott could have meant. I’m now thinking he’s saying something along those lines:

Assume the CB fixes IOR and M and lets the interest rate float; if the market pushes the interest rate higher, then inflation is a likely result.

Given that CBs fix IOR and the interest rate and let M float, I was having a hard time registering what he meant, but I guess that’s what you get from folks who are so attached to the quantity theory.

My view on that clarified statement is that it doesn’t really matter anymore than “will the car fly off the road to the left or to the right if I lift my hands off the wheel”.

Just for closure, I’d say I generally believe this would come with inflation (argument is: folks are seeking more money to carry out transaction and willing to pay a higher opportunity cost). But you could also imagine a situation where agents fear a lack of liquidity and just hoard money (keep larger balances in their checking account for treasury purposes, even at the higher opportunity cost). This means the money is diverted away from it’s transaction function, and that would come with deflation (lower velocity).

I’m very cautious to avoid jumping to macroeconomic conclusions based on “data” because there are always so many factors moving around that can’t be controlled for, but that doesn’t mean I can’t quote someone smarter than me making a solid observation. This post by Scott Sumner rubs MMT’s nose into its poor treatment of monetary policy:

Here’s another problem with the Fiscal Theory of the Price Level.  When the Fed decided to target inflation at around 2%, they pretty much succeeded.  Raise your hand if you think Congress should be congratulated for bringing inflation down sharply in 1981-82 with the massive Reagan deficits, and then after 1990 holding inflation near 2% for decades.  Congress adeptly nudges deficits up and down to keep that darn inflation rate pegged right around 2%.  I don’t see any hands in the air.

…interest rates do. A great many econ-bloggers criticize the emphasis of central banks on interest rate targeting (which is a misnomer by the way, because interest rates are the control, not the objective).

Those same bloggers typically believe that quantity of money can always be varied to achieve any inflation objective. This relies on the existence of a demand function for money: there is no such thing as money is substantially identical to safe assets, i.e. assets whose money-price is relatively stable.

When a central bank creates money, it substitutes money for a safe asset (either a repo, or a treasury bond) which means that the quantity of safe assets is left untouched. In fact the quantify of safe asset isn’t even all that well defined as there isn’t a clean divide between safe assets and unsafe assets: money, repos, short term treasuries, longer term treasuries, senior CDO tranches, corporates bonds, all fall into the same continuum. This relates to what JP Koning writes regarding “moneyness”.

It’s not even clear that issuance of financial “safe assets” truly creates safe assets: for instance, if I mortgage my house, I create a (relatively) safe asset: the mortgage. But in the process, I’ve transformed a somewhat safe asset (the house) into a much less safe asset (the mortgaged house = the house minus the mortgage). On the other hand, building a brand new house does create an asset (and therefore to some extent it adds to the total quantity of safe assets).

Having said that, money isn’t exactly identical to the other safe assets: it is transactable (checks can be written, currency can be handed over, etc.). For that ability, users of money might be willing to forego some return when holding money rather than holding safe assets, some spread if you will. But once the need for transactable money is satiated, agents will not be willing to borrow anymore money from its issuer (the central bank) unless the rate paid (IOR/IOER) is equivalent to that paid on other safe assets.

In the old days, when IOR = IOER = 0%, quantity of money might have been related to interest rates as the holders of money had to forego the entire interest rate in order to get money: so when the Fed injected extra money, it couldn’t find borrowers unless it lowered its lending rate (effectively reducing the spread between lending rate and IOR, i.e. the opportunity cost of holding money relative to another safe asset).

Nowadays, with IOR = IOER = target rate, money’s yield is the same as other safe asset yield. So when the central bank adds extra money into the system nothing happens. Someone who was previously holding a safe asset yielding X% is now holding money yielding X%: he’s completely indifferent unless liquidity in the transactable asset was in short supply.

To drive the point home: creating money achieves nothing for the price level. Reducing interest rates does. This is how.

Scott Sumner has argued that by creating more base money today, the central bank is signaling that it will keep the base money in excess supply tomorrow:

DOB, I think where we disagree is the temporary/permanent distinction. When considering monetary policy options I believe the baseline assumption should be that changes in the base are permanent. In that case it doesn’t matter (very much) what the central bank buys. The important changes occur because the base is larger, and that raises rates via the hot potato effect.

I don’t buy that: the actions of central banks are fully reversible on a moment’s notice: the OMO desk can just pick up its phone, call primary dealers and drain the reserves. Market participants are fully aware of that.

A central bank announcement that it will keep rates at 0% until a return to target path (possibly instilling the fear of some higher-than-usual inflation in the future) might create said “hot potato effect today”. That’s only if the central bank manages to convince market participants that it will stick to its plan, however unpleasant it might feel when the time comes. John Cochrane puts it best:

The problem is deep. How can the Fed have the power to [..] commit itself to doing something in the future that it very much will not want to do when the time comes — that it has so far explicitly and loudly promised that it will not do, and has built up 30 years of reputation against doing — undershoot and cause inflation? Why can’t the Fed 2015 convene another Jackson Hole conference, bring out Charlie Plosser and Jeff Lacker’s friends to say “we’re heading to a repetition of the 1970s, it’s time to become new-Monetarists and promise that we’ll be looking at monetary aggregates not unemployment rates?” It’s so much hot air, and people know it.

I don’t think that problem is insurmountable though: if people understand NGDPLT and its benefits, it’s possible there won’t be as much emotions should we get a little bit of above-trend inflation. Also if the remedy works fast, we won’t need all that much catch-up inflation.

That said, I continue to believe negative rates are a much more robust solution than commitments to keep rates at 0% for various length of times as they do not rely on threats that the central banks will do something that it “very much will not want to do when the time comes”.

Note #1: QE’s effect comes from the fact that the assets that are purchased aren’t all that safe (longer term treasuries have meaningful interest rate risk, not to mention the government credit risk), not from the actual increase in MBase. The exact same effect could be achieved if the Fed traded interest rate swaps without touching MBase (ignoring govt credit issue for simplicity). Another equivalent way to achieve the same thing would be for the treasury to buy back its long term debt and fund the buybacks by issuing short term. All of these are a horrendous idea as they ruin the asset-liability match of the government, but that’s an entirely different subject.

Note #2: Bill Woolsey wrote some excellent posts on the hot potato effects here, here and here.

A couple of years ago, as I was trying to understand how money worked, I stumbled upon MMT (Modern Monetary Theory). It felt like an elegant and consistent way to think about the world, and as an open-minded guy who accepts that macroeconomics can sometimes be counter-intuitive, I wasn’t too bothered by the fact that it contradicted most of the common wisdom that was out there.

Still, I sought out criticisms of MMT (just in case I was missing something) but was comforted by the fact that the ones I found either did not understand MMT or were flawed and easily rebutted by MMT proponents. A lot of those who took the time to attack MMT seemed to come from an unorthodox background themselves, such the Austrian school. Maybe because the mainstream simply thought the flaws were so obvious they didn’t think they were worth addressing?

MMT’s problems are very hard to spot when you look at the world from the MMT perspective, so it took me a while to really wrap my head around them. It also made it very difficult to communicate with Warren Mosler (the more reasonable leader of the MMT movement) and we eventually had to agree to disagree when we discussed the subject in his blog’s comment section. I told him I’d have a separate post ready in “1 to 2 weeks“. That was 6 months ago

In my opinion, the deepest flaw in MMT is its lack of understanding of how interest rates work. Here are a few quotes of Warren Mosler (some of which are from conversations I had with him in his blog‘s comment section, some of which aren’t):

“It’s still all a sign that monetary policy is shooting blanks” (from here)

“When the fed pays interest it adds income and net financial assets. And it got permission to do it from congress.” (from here)

“Another central bank may have it backwards as lower rates turn out to be deflationary and slow things down via interest income channels?” (from here)

“I have no fear whatsoever of the Fed causing inflation. In fact, theory and evidence tells me their tools more likely work in reverse, due to the interest income channels. That’s because when they lower rates, they are working to remove net interest income from the private sector, and when they buy US Treasury securities (aka QE/ quantitative easing) they remove even more interest income from the economy. Remember that $79 billion in QE portfolio profits the Fed turned over to the Treasury last year? Those dollars would have otherwise remained in the economy.” (from It must be impossible for the Fed to create inflation)

So according to MMT, the central banks have very little control on inflation (“shooting blanks”). Or worse, they have it backward: they should raise rates if they want to prop up inflation since that generates income for the economy (“interest income channels”). Pushing this logic further, MMT concludes that the central banks should get out of the way altogether, and let fiscal policymakers take care of inflation:

“I’d leave rates at 0 permanently and adjust taxes accordingly” (from here)

“The Natural Rate of Interest is Zero“

So why did MMT go so wrong? In my opinion, the MMT guys look at the world as a public/private sector dichotomy: anything that doesn’t create “net financial assets” in the private sector basically doesn’t do anything. Since monetary open market operations are essentially an exchange of one financial asset for another (a deposit vs. a repo for the same amount), MMT concludes that it can’t possibly have any effect. Only fiscal operations which truly alter the balance of financial assets in the private sector can push the rate of inflation one way or the other. Of course, MMT proponents love to point out over and over that the stock of “net financial assets” held by the private (and external) sectors is equal to the public debt, “to the penny” (I think that might be why Brad Delong deemed MMT a tautology)

The DOBaroo

In this subsection of the post, we’re going to go over how a central bank can control the price level of the currency it issues, without altering the balance of “net financial assets” in the private sector or requiring assistance from the fiscal side. In fact, I don’t even need to assume that there is any government spending, taxes, etc.

Start from a world of bartering, and assume a central bank pops out of the ground and announces it will be issuing a new electronic currency, the DOBaroo, to facilitate transactions. The mechanics are as follows:

• Everyday the central bank will post a rate at which it will lend DOBaroos overnight
• Anyone can borrow as many DOBaroos as they want provided they post a generous amount of collateral (could be government bonds if they exist, but could also be real assets such as a house, a business, etc.  It doesn’t really matter as long as the loan-to-value ratio is sufficiently low to make credit a non-issue)
• The payment system enables the initial borrower of funds to transfer his DOBaroos to a third party: the saver
• Savers receive the overnight rate of interest posted by the central bank (we’ll assume it’s equal to the rate paid by the borrowers, so the central bank pays and receives exactly the same amount of interest, i.e. no impact on private “net financial assets”)
• The central bank announces that it will altruistically set the rate of interest such as to achieve a certain target for the price-level. For simplicity, we’ll assume 2% CPI inflation

The fundamental question is: can the central bank thwart any inflationary pressure and remain on target? The answer is yes, by setting the overnight nominal rate of interest sufficiently high. Even if inflation expectations were out of control at say, 200%, the central bank could post a nominal interest rate of 500%, so that, if you chose to save for a year, you end up with 6x as many DOBaroos as you started with but they’ll be worth 3x as little: in other words you doubled your real capital (100% real return). Unless return on capital in your economy is that high, holding DOBaroos was a more attractive proposition over that period than holding real capital, and therefore rational agents will do it, preventing the uncontrolled inflation from even being expected in the first place, or the need for the ultra-high nominal rates in the example.

Conversely, the central bank can always set a nominal rate sufficiently low (possibly negative) to prop up the rate of inflation when below target.

We’ve simplified a few things to keep things tractable, but none of the corners we’ve cut invalidate the validity of this theory in the real world. The DOBaroo example is almost verbatim from this theory section, and a more “real world”-like currency construction can be found here. The DOBaroo name is a reference to MMT’s flagship example: the UMKC Buckaroo.

Interest Income Channels

Finally, let’s go over the dreaded “interest income channels” argument whereby savers are encouraged to save even more when interest rates are low in order to maintain a certain level of future consumption, thereby pushing down aggregate demand. This argument leads Warren Mosler (and a few others) to conclude that lower rates are deflationary.

First of all, Warren’s got the causation backward. The saver isn’t saving more because real interest rates are low: real interest rates are low because the saver wants to save more. Is Warren saying that the demand for saving goes up as real return goes down? Ask yourself this: if real return on your savings was -100%, i.e. whatever you save gets vaporized after a year: how much would you save? Personally, I’d consume everything I’ve got as fast as I could.. I might even borrow to consume some more → There always exists real rate of return at which the savings market is in equilibrium.

In any case, the government should not attempt to distort/control real return on capital any more than the price of apples. Real rates are the the price mechanism by which we mutually agree to exchange our consumption through time and we use that price signal to decide how much of our output we produce for investment vs how much we produce for immediate consumption. If we all want to consume tomorrow but not today, real interest rates will be very low (at least in the short term, everything else remaining equal). That’s especially true if tomorrow we’ll all be old and there won’t be many young people to cater to our needs (high demand, low supply → high price of future consumption → low real return). Conversely, if we all want to consume today rather than tomorrow, then real interest rates will be high.

Central banks generally set nominal interest rates wherever they need to be to keep the real overnight return on money in line with the real risk-free overnight return on capital. Like the Higgs Boson, the latter is not observable but it’s still an important concept. A central bank can wiggle its target rate a bit higher or lower than theoretical target, especially if the market is questioning its credibility (as we’ve seen in the DOBaroo’s extreme example), but by and large, the rates a central bank sets are being imposed on it by its mandate and the market conditions.

Conclusion

Contrary to what MMT claims, interest rates can be used to regulate the price level, regardless of whether central bank operations affect the level of “net financial assets” in the private sector. So if monetary policy does the job, should we use fiscal policy for that purpose? In future posts, we’ll see that the answer is No. As we continue to pull on that thread, the entire fabric of the theory will fall apart. In the mean time, I welcome all comments with constructive counter-arguments.

She seems to be leaning towards something along the lines of MMT. I’ll be posting a criticism of MMT soon.

Since we’re watching Youtube videos, this fellow seems to think Hayek would have wanted “stability” of NGDP (go to the 2 minutes 12 seconds mark). I know very little about Austrians, so not sure whether stable means fixed growth path or 0% growth..