02 Mar 2015

## Potpourri

==> I am pretty sure this is checkmate for Scott’s attempt to reason from a price change regarding the ECB. Note, I’m not claiming I’ve won the whole tournament–just this particular game.

==> I break out my undergrad Public Choice text to argue that the work of Niskanen is at serious tension with (some of) what the new Niskanen Center is doing.

==> Someone help me out here: What the heck is Brad DeLong talking about, when he says such-and-such an interest rate cut would lead to only a 6% increase in the price of a long-duration asset? Can anybody reverse engineer his calculation so I can be sure of what he’s saying? I think he’s totally wrong but I’d feel more comfortable if I knew exactly how he got that number, before going on the attack.

#### 31 Responses to “Potpourri”

1. Zack says:

I was confused by that DeLong statement as well. John Taylor has said that according to the Taylor Rule, rates should have been about two percentage points higher from 2003-2005. Is DeLong getting that number by just multiplying 2% by three years? I would appreciate it if someone could clear this up.

2. E. Harding says:

Good catch on Sumner! Even I missed the “not”.

3. Jeff says:

It’s difficult to understand what DeLong is trying to say, but I’ll try to offer one interpretation, as a financial economist. In this context, duration means price sensitivity to interest rates. This is a common risk metric for fixed income securities, so how it applies to real estate (as DeLong wants to do) is indirect at best. It would only directly apply to real estate properties held for rental income purposes, not to a family’s own dwelling which provides service flow.

Dollar duration is defined as the change in the asset price caused by a small change in the relevant interest rate. The percentage duration is then the dollar duration divided by the asset price. Duration and maturity are closely correlated, but certainly aren’t the same. The higher the asset’s duration, the more sensitive it is to interest rate changes.

So DeLong is saying, I think, that with the interest rate 2% below where it “should” be, asset prices are 6% higher than they “should” be. He is getting the 2% from Taylor. To get 6%, then, in this context, we’d need a duration that would result in a 3% change in the asset value for every 1% change in the interest rate.

Mathematically, we have Duration as: dV/di*(1/V) where V is the value, and i is the interest rate. In this context, I think DeLong is saying that dV/V = 0.06, and di = 0.02. That implies a duration of (dV/V)*(1/di) = 0.06*(1/0.02) = 3. That is a very low duration, similar to a 3-year bond, not a 30-year bond. To get a “long-duration” result, we’d need a change of X*(1/.02) = 30, X = 0.6 or 60%. It seems in this case the DeLong is off by an order of magnitude.

It doesn’t matter much is DeLong meant “long maturity” or “long lived” rather than “long duration” in this sense, since for the purposes here they would be largely compatible results. That’s because assets that have a long maturity also have high durations (interest-rate sensitivity).

If this is not what DeLong is thinking, then I have no idea what model he’s using to get that 6%.

• Bob Murphy says:

Jeff, thanks, this is exactly what (I think) DeLong was trying to do–but I too couldn’t figure out how he would get 6% to pop out as the change in asset price.

I think what may be going on is that the “3 years” is relevant. I.e. he’s not saying assume a permanent reduction in interest rates of two points, but instead a 3-year reduction. Yet it’s not obvious how you would apply the standard formula in that case, and still I can’t get a 6% increase to pop out; on a 30-year asset I only get a 1% increase.

• Transformer says:

This is probably a bit simplistic but if you took out a 30 year mortgage with a rate that was discounted by 2% for the first 3 years then you could pay 6% extra for the house (plus a bit of compound interest) compared to the house you could buy without the discount and still pay the same amount back. The reduced interest payment would pay off the additional purchase price and then you’d be back on the same schedule from year 4 onwards as without the discount, assuming fixed payments on the loan throughout.

As most people take out loans fixed at the same rate for the full term of a mortgage its hard to see how this is relevant. If you could borrow at a rate at 2% below the true rate for 30 years you could afford to pay way more than 6% over the true price to buy a house.

4. Andrew_FL says:

Also on Scott’s post I found his response to my comment somewhat completely unintelligible:

“In my view expansionary monetary policy always has the same effect, rising NGDP, mostly because I define easy money in terms of NGDP growth.”

Did he really mean to say, loose monetary policy always has the effect of…loose monetary policy?

• Zack says:

That is a bizarre comment. I thought Sumner’s view was that the path of Nominal GDP itself is the measure of whether monetary policy is too tight or loose. It sounds like he’s simultaneously referring to rising NGDP as an “effect” of loose monetary policy and also the definition of loose monetary policy in the same sentence.

• Andrew_FL says:

His attempt to clarify “sort of” explains it, but not much. More or less he said he’d define policy by the price of NGDP futures contracts (That is, policy is *future* not current NGDP) but they don’t exist yet.

Which would technically make the cause and the effect separate things. It’s still either vacuous or wrong though. I’ll be charitable and assume it’s vacuous.

• Zack says:

Well that makes more sense, but in the same comment he follows by saying ” Unfortunately we lack that market, so I often look at actual NGDP as an indicator of the stance of monetary policy.” So his original statement still seems very strange to me. I’m guessing he just didn’t think that one through carefully before he wrote it.

5. Jeff says:

Bob,

Indeed, the duration formula does “reverse” at any point, although the price impact would converge to zero as the bond approached maturity. Without maturity, as in the case of real estate, there wouldn’t be a convergence.

Now, I can get you a 6% increase in real estate prices with a 2% reduction the mortgage rate if that’s what you want, but it means I get to play around with all the other parameters. So without context from DeLong, I think all one can say is that his claim is non-falsifiable (and thus also can’t be confirmed).

• Bob Murphy says:

Jeff wrote:

Now, I can get you a 6% increase in real estate prices with a 2% reduction the mortgage rate if that’s what you want,

OK sure Jeff, play around and tell me what you get. The only constraint is, use mortgage rates that actually existed in the 2000s. I think you’re going to find these are short mortgages…

• Jeff says:

Sorry, something was clearly lost in translation. What I was meaning was I can “get you” what you want, but not with any reasonable parameters. You know, the way 2+2=5? I’m not saying I can legitimately get a 6% price change in houses.

6. Jeff says:

Upon further reflection, I’m more confused about DeLong’s position. An interest rate that is 2% too low for three years doesn’t relate to any concept of duration other than maturity. So let’s think of something really simplistic.

If I get a \$200,000 mortgage for a rental property, financed at 5.5% for 30 years, and rent it out so that I’m even-Steven on the payments (meaning a rental payment per year of \$13,760), then the present value of the loan is \$200,000. If I crank that rate to 7.5% per annum, I have an immediate price drop of nearly \$40,000. Then, I reverse that after 3 years, and go back to 5.5%. I end up at \$188,000, which makes up \$12,000 loss over three years. That’s 6% on a \$200,000 mortgage. I can do this \$100,000 or \$300,000, as long as I set the rent such that it’s a par investment. So I can get that 6% over three years, but I think the methods I need to take to do it are singularly unrealistic.

7. konst says:

Off topic but can anyone explain something about MMt.

Ignoring exports and imports, G is government budget, T is taxes, S is savings, I is investments

then why is G-T=S-I ? Why is that assumed to be true?

• Andrew_FL says:

Hm if income is either spent on consumption, saved, or taxed, then:

Y = C + S + T

And of course if income is equal to expenditures on final goods:

Y = E = C + I + G

G isn’t actually government spending, by the way, it’s purchases of final goods, so strictly speaking since much government spending is on transfers, G – T is not the same thing as the deficit.

At any rate:

C + I + G = C + S + T

By combining the above two equations. It follows that

I + G = S + T

G – T = S – I

This is not the way the NIPA work though, because there I = S by definition.

• konst says:

I don’t see where you get this from Y = E = C + I + G. Is Y supposed to be some kind of GDP like measurement?

To me G – T has no basis in reality, i.e. the real world we all live in with scarcity, constraints, and common sense and GDP has no real relationship to reality either. What I don’t understand is that private savings and assets pre-existed and pre-dates government attempts to co-opt them so that when MMT says money comes into existence because the government spends it into existence, that doesn’t seem true. It seems most private wealth and a lot of savings are beyond the reach of government (at least currently) and that the government is slowly feeding, or ties to, on these private assets thereby destroying decades or hundreds of years of wealth created without the use of government spending.

Basically I’m saying I think some of the savings, and probably some investments, people use have no relationship to the government just that government feeds on those like a parasite does to it’s host. So G – T is separate from S – I, just a small part of S and I (the part created by government money) participate in the government’s exchange with it’s host.

• Andrew_FL says:

Konst, yes, Y is supposed to be GDP. I’m just explaining how MMTers get their accounting identities. The idea of Income being equal to expenditure is simple enough: one person’s expenditure is always someone else’s income. When you pay money for something, someone else receives money for it.

If we take the same accounting identities without the Government making any purchases or engaging in any taxation:

Y = C + S

and

Y = E = C + I

What the MMTers are actually implying with their accounting identities is that absent the government making purchases of final goods and engaging in taxation,

C + I = C + S

or

I = S

So there is nothing in their accounting identities that actually contradicts the idea that these concepts predate the government, regardless of what they claim.

It’s also worth realizing, though, that again, regardless of what they might claim, the above equations are for nominal GDP. It’s perfectly possible for spending in dollars to increase and income in goods to decrease (and thus the rate of wealth accumulation to be slowed) it just requires (price) inflation.

I would advise against thinking too much about what the neo-chartalists are trying to say, even on their own terms it makes no sense.

• konst says:

Having recently rewatched the Murphy vs Mossler debate I was wondering if MMT is really based on reality or fantasy. I’m inclined to think they can’t even fit their model to the real world however everyone think those “accounting identities” are true. Just wondering why they, meaning people and economists, generally accept them.

• guest says:

I think the main thing to keep in mind regarding MMT (from what little I’ve read about it) is that they are wrong to think that when the government spends printed money, that they are actually creating or spending wealth.

All the government does when it spends printed money is counterfeit money. It isn’t spending created wealth, it’s spending STOLEN wealth. Same with lending printed money.

It only seems to work because people have been mislead as to the marginal value of the printed money they have been given in “payment”.

I believe this is the heart of the problem with MMT.

• Transformer says:

Say everyone earns \$100 and there are 10 people in the economy.
The govt taxes everyone \$10 so total tax is \$100
It borrows \$10 from everyone so total G = \$200 (\$100 tax plus \$100 borrowed)

After tax everyone has \$90 left of their income (\$900 total)
Assume they spend \$60 each on consumer goods (\$600 total)
The rest is savings = \$30 each (\$300 total)
They have lent \$10 each to the govt (\$100 total) which leaves \$20 each (\$200 total) which they are assumed to invest.

so:
G-T = 200-100 = 100
S-I = 300-200 = 100

(the I bit is dodgy IMO, because they may just save some of the \$200 too, but MMT people assume that uninvested savings count as investment in inventory by businesses)

• Transformer says:

That was a reply to Konst BTW

• konst says:

So what you’re saying is first the government create a budget, let say \$5 trillion. Then the Treasury issues bonds (as a formality), then the central bank buys those bonds thus G. Then the government taxes people, entities, pets, etc. thus T. Ergo G – T = Y.

So why is S – I supposed to also equal Y? Obviously there’s something I’m missing. Maybe I should look into it when I have more time.

• Bob Murphy says:

Andrew_FL gave the accounting reply, konst, but the intuition is: if the private sector saves more than it invests, where does that savings go? It has to go outside the private sector, i.e. into government bonds. But the only way the government can soak up that extra private savings is by itself running a deficit.

(I’m not saying this is good stuff, just explaining the intuition of how MMTers argue that the government must run a deficit if we want the private sector to engage in their odd notion of “net” saving.)

• konst says:

Well the problem for me is that I have a kind of model of the economy that consists of all actions and all the social the interactions of everyone; let’s call that “life”. Obviously it’s crude and I don’t get like actual numeric values from it, I just kind of visualize it/run it in my head. The thing is Austrian economics fits it but MMT does not (at least what I understand of MMT). For MMTers, the government and the governments “subjects” seem to be their whole universe.

For me, the government, by their actions extracts wealth from “life” kind of like a parasite. I’m not trying to be funny, it’s just that what a “parasite” does fits what the government does to this model. In this model “accounting identities” don’t relate to what’s happening through “government operations” as MMTers like to say. So when MMTers say it’s just accounting identities I don’t buy that. What I see is the operations of the government as it relates to the “economy” extracts value from “life” which wouldn’t add up to the numbers MMTers present. I guess what I’m saying is the MMT model isn’t reality even though it’s presented as though it is.

I’m not sure if I’m explaining my point clearly.

• Andrew_FL says:

Konst, it’s possible for an accounting identity to be true but for the intuition behind the definitions used to create the identity to lead to wrong conclusions.

Always remember that accounting identities like these are always trying to split up the nominal spending stream, measured in dollars, in any variety of ways. Regardless of what the creators of such identities believe, they say nothing about the real economy, only the nominal spending stream ie the flow of money. The real economy is measured in stuff, not dollars.

To understand the flow of actual wealth, actual stuff, you need price theory, capital theory, etc. That’s the difference between Austrian economics and all these accounting identity schools.

• konst says:

Ok so what you’re saying is the accounting identities would hold regardless of whether the monetary operations of the government destroy the real economy or not.

• Andrew_FL says:

Bingo.

• guest says:

“The real economy is measured in stuff, not dollars.”

Nice. Very important.

I recommend this video to help understand this point:

The Birth of the Austrian School | Joseph T. Salerno