Secular stagnation, liquidity, and rent/price ratios

No answers here, only questions.

By "secular stagnation" I mean "declining equilibrium real interest rates".

Most explanations of secular stagnation say it is caused by a rising desire to save and/or a falling investment demand. Call this the "Saving/Investment Hypothesis".

But there are lots of different real interest rates. For example, the real interest rate on Bank of Canada currency is around minus 2%. (That currency pays 0% nominal interest, and the Bank of Canada targets 2% inflation). But people are willing to hold currency, despite that, because it is very liquid. But all assets differ in their liquidity. More liquid assets will have a lower real yield than less liquid assets. And if an asset becomes more liquid over time, its real yield will fall over time.

David Beckworth, JP Koning, and now Steve Williamson, say that falling real interest rates (on bonds and/or stocks) might be due to rising liquidity, (of bonds and/or stocks). Call this the "Liquidity Hypothesis".

One set of data, and two competing hypotheses to explain that data. So we need more data, to try to distinguish between them.

1. Would looking at data on rent/price ratios, for houses, or farmland, help us resolve this question? Because my random Googling suggests that rent/price ratios have generally been falling (prices rising faster than rents), and this is a global trend over the last couple of decades that is consistent with falling equilibrium real interest rates.

2. The US housing market is a recent exception to this general trend. Now the US is not the world, and the market for US farmland seems to be following the global trend of falling rent/price ratios. But we still want to try to explain the exceptions.

3. The equilibrium price P of an acre of land is the present value of the rents. If rents R are rising at rate g, then the implicit interest rate r is given by:

P = R/(r-g)  or r = g + R/P (where r and g are either both nominal or both real)

Is it plausible that falling rent/price ratios are due to rising expected growth rates of rents, rather than falling interest rates?

4. Is it plausible that houses and farmland have become more liquid over time, so that the Liquidity Hypothesis can explain falling rent/price ratios too?

Comments

[Mike Sproul]

The gold lease rate is a good proxy for the real interest rate on highly liquid assets.

[Nick Rowe]

Mike: please tell us more. What is "the gold lease rate"? Has it been falling over the last couple of decades?

[louis]

"Is it plausible that falling rent/price ratios are due to rising expected growth rates of rents, rather than falling interest rates?"
I'm having trouble with the question. If land were the only asset, we could puzzle over this question. But we have fixed income financial assets! Look at the path of interest rates on bonds, both government and corporate. Correct for expected inflation (and maybe changes in risk premiums). Then you know what has happened to r, and then you can infer changes in expected g for rents. Where is the puzzle?

"Is it plausible that houses and farmland have become more liquid over time, so that the Liquidity Hypothesis can explain falling rent/price ratios too?"
Certainly plausible: maybe houses and farmland have become more acceptable as collateral for loans, or technology has improved the ability to match buyers and sellers. That works for those individual assets. Houses get valued more like money if their properties become closer to those of money. But what are the aggregate effects? If everyone's portfolio of assets becomes more liquid, does the desire for liquidity (and the demand for the medium of exchange at a -2% real return) diminish?

[Jonathan]

"falling real interest rates (on bonds and/or stocks) might be due to rising liquidity, (of bonds and/or stocks)."

I think you mean to say a rising liquidity PREMIUM?

Incidentally, I think this is a good partial explanation for the fall or short interest rates in 2008, but can it explain the long-term trend? Also, is there really a long downward trend in REAL interest rates?

Okay, another nitpicky comment:
'By "secular stagnation" I mean "declining equilibrium real interest rates".'

Declining or low? It seems that if one expects a declining path of r, that's rather different than a constant r at a low level.

"4. Is it plausible that houses and farmland have become more liquid over time, so that the Liquidity Hypothesis can explain falling rent/price ratios too?"

I like this idea.

Regarding rent/price ratios, I've heard a few people (including Matt Rognlie in his recent Brookings paper, and also Joe Stiglitz I think) argue that a big part of the recent increase in capital relative to GDP (and possible in the capital share of income) is due to real estate, and mainly rising land prices. I wonder how this fits in.

I'm still quite uncertain about the secular stagnation / capital share nexus of ideas, but they seem important.

[Jonathan]

Wait, I retract my first comment (about liquidity premia). I realize that I misread your sentence, and that the argument is that in facts bonds and stocks have become MORE liquid over time, lowering their return relative to cash (i.e. lowering the liquidity premium of cash, or lowering the ILliquidity premium of these assets).

[Frank Restly]

Nick,

"David Beckworth, JP Koning, and now Steve Williamson, say that falling real interest rates (on bonds and/or stocks) might be due to rising liquidity, (of bonds and/or stocks). Call this the Liquidity Hypothesis."

Careful. An increase in the liquidity of bonds might explain rising bond prices. It does not (at least completely) explain falling real interest rates.

[Nick Rowe]

louis: if bonds are more liquid than land, then the rate of return/interest r on bonds will be lower than the rate of return/interest r on land. And if bonds have been getting more liquid over time, but land has not, then r on bonds will be falling relative to the r on land.

Jonathan: declining or low? Well, low relative to what they have been in the past. Declining to a level near zero, or at least below the growth rate in GDP.

Yep, I've been pushing for people to include land in stories/models of secular stagnation for some time. See the link to Stefan Hombug's paper on Piketty in this post, for example.

Frank: for a given coupon and redemption value, a rise in the price of a bond MEANS the interest rate on that bond has fallen. (Do not try to derail this comment thread.)

[Jonathan]

Incidentally, since you mention Eggertsson and Mehrotra in the post you link to above, they do discuss land towards the end of their paper (in a discussion of extensions, last section in the latest version). I think they recognize its importance.

[Frank Restly]

Nick,

"By secular stagnation I mean declining equilibrium real interest rates".

Does that also mean rising real equilibrium bond prices?

[Nick Rowe]

Frank: Yes. Stop now.

[Mike Sproul]

Nick:

I can't tell if you're pulling my leg, but check the history of the gold lease rate here:

http://www.kitco.com/lease.chart.html

It's been pretty constant over the decade, except for a spike around 2009.

Example of the GLR:

I borrow 100 oz of gold, and repay 100.5 oz in 1 year, so the GLR=.5%

[Nick Rowe]

Mike: I wasn't pulling your leg. I thought it might mean something like that, but I wasn't sure. Thanks.

See the blue line on the bottom chart that Mike gives us. So "0.5" on the vertical axis translate as 0.5% real interest rate measured in gold? That's low, and stable, except for the massive spike up in late 2008.

[Majromax]

Land and especially housing may not be the best counterfactual, since housing wealth has become more liquid. The house itself obviously has to remain there save at great expense, but more flexible mortgages and home-equity lines of credit allow homeowners to use that equity for leverage. The liquidity doesn't come from real-estate sales, but instead from its use for credit.

But we do have "capital" that has a fixed and low liquidity, in the form of human capital. How has the wage premium divided by tuition cost changed over the years for a 4-year degree? What about the wage premium for seniority/experience?

If we assume that wages are at a competitive equilibrium, then the wage of a worker with 5 years' experience divided by the wage of a new hire (perhaps post-probation?) corresponds to a "real labour interest rate." However, this experience cannot be sold independently of the worker, nor can future experience be borrowed against.

[Majromax]

@Nick Rowe:

> [The gold lease rate]'s low, and stable, except for the massive spike up in late 2008.

I'm not sure that I see "low and stable" from that graph. It looks like the spread between 1 and 12-month lease rates was nearly nonexistent from mid-2005 to mid-2007 at least, whereas now it's a full 30 basis points.

[JP Koning]

Nick, I think #4 is a possibility. See here.

"One set of data, and two competing hypotheses to explain that data. So we need more data, to try to distinguish between them."

The thing that worries me is that we may never be able to find sufficient data to answer your challenge. Every new series we turn to could very well be tainted by a liquidity premium, making it impossible to tease out the real from the nominal.

[Nick Rowe]

Majro and JP, I see you are on the same page. I had forgotten reading JP's good post on land liquidity. Memory depreciates.

The rent/cost ratio on human capital looks like another data set. But what we observe there is the average cost, not the marginal cost of human capital. We should expect Pk=MCk > ACk. And is human capital more liquid now too, in the sense of being easier to borrow against?

Majro: "It looks like the spread between 1 and 12-month lease rates was nearly nonexistent from mid-2005 to mid-2007 at least, whereas now it's a full 30 basis points."

I wonder why that should be?

[louis]

Nick: Sorry, I misunderstood the post.

Having read the Beckworth and Koning pieces you linked to, the argument looks different to me.

US Treasuries are and have long been a supremely liquid asset. Nominal yields on US treasury bonds have fallen dramatically over a few decades. Some of the decline is due to a fall in inflation expectations over that same period. The residual is the fall in the real rate of interest, which should reflect changes in the balance of desired saving and investment. What Beckworth and Koning are saying is that you cannot determine the change in the real rate because you cannot determine inflation expectations by looking at the spread between conventional Treasuries and TIPS, because there is a discount people apply to TIPS due to their relatively lower liquidity vs other Treasuries, and that discount varies over time. To believe this, I think you need to believe the order of magnitude of this "noise" effect is as large as the signal of inflation expectations.
To show that real rates have indeed fallen, and that the saving/investment hypothesis is true, all you need to do is show that inflation expectations have fallen by less than nominal rates.

[Nick Rowe]

louis: take an extreme case. Suppose Canadians started using government bonds exactly like currency. Since the Bank of Canada targets 2% inflation, and the nominal interest rate on currency is 0%, we would expect to see the nominal interest rate on government bonds fall to 0% too, so the real rate on government bonds becomes minus 2%, just like currency.

[louis]

Nick: I get that. As a side point, in that case the bank of Canada would need to sharply reduce the money supply to keep its inflation target.

The thing is, I don't think anyone is arguing that the increase in benchmark government bond prices is due to a decrease in their liquidity discount to cash, as opposed to a change in the time value of money. What the first two links in your post talk about is the changing liquidity discount of TIPS to benchmark Treasury bonds, not benchmarks to cash.

[JP Koning]

Louis, my post argued that the decline in TIPS yields (and government yields in general) could be due to a decrease in their liquidity dicount to cash, not variations in the real return on capital. IOW, I agree w/ what Nick describes at 5:41.

Separately, I argued in that same post that the changing liquidity discount on TIPS relative to straight Treasuries muddies popular measures of inflation expectations. But that doesn't have anything to do with the secular stagnation argument, it's mainly me getting grumpy about how liquidity premia in general create distortions across a wide range of popular indicators.

Nick, I see a pattern here. If bonds, housing, and labour are all getting more liquid (or at least suffer from varying liquidity), then we can't use them as clear indicators of the equilibrium real rate, at least not without serious adjustment. So what do we use? As you said, only questions here, not answers.

Mike, the gold lease rate isn't a clear indicator of real returns because holding gold through time involves significant storage costs. The gold lease rate could be falling (or even negative) because storage costs are rising, and those who own gold are growing increasingly desperate to offload those storage costs onto borrowers. They might even be willing to pay negative interest if storage costs get especially onerous.

[Avon Barksdale]

Nick,

"Is it plausible that falling rent/price ratios are due to rising expected growth rates of rents, rather than falling interest rates?"

No. This is a well tested observation in finance: falling dividend/price ratios (or rent/price) are not indicative of higher returns, they indicate lower returns. This has to do with time varying risk premia. I suggest "Discount Rates" by John Cochrane.

http://faculty.chicagobooth.edu/john.cochrane/research/papers/discount_rates_jf.pdf

Pay close attention to figure 2 in this paper and the discussion surrounding it.

[Adam P]

Nick, are we sure these hyothesis compete with one another?

Let's imagine the the consumption/saving decision with 3 assets, stocks, bonds and money. The expected returns are of course: stocks > bonds > money and, of course, the opposite for the relative liquidities. At whatever the beginning values for risk/returns and liquidities are, we have a saving demand. BTW, assume everyone likes more return to less and more liquidity to less.

Now, make bonds more liquid (financial inovation I guess).

For the same yield I buy more bonds but also I save more. This is because in my portfolio one part got more desirable but the other 2 didn't get any less desirable, so my asset portfolio is more desirable on the whole. Thus, before I was indifferent on the margin between an extra penny of saving or an extra penny of consumption, now I must want more consumption.

Also, I shift out of equities and money and into bonds. This is a bit like substitution and wealth effects, if the "wealth" effect dominates (meaning the increase in saving) then the prices of equities go up with the price of bonds.

So both hypothesis are true, though I guess you could say the root cause is the liquidity hypothesis. I've also assumed that there is no effect on the marginal product of capital, which may not make sense. but anyway....

And if both equities and bonds get more liquid, then it's a pure effect. More saving unambiguously.

[Nick Rowe]

Adam: good point. No I'm not sure. Especially in a general equilibrium setting, where assets on average become more liquid. But perhaps I should have rephrased it like this: Was it an increase in desired saving caused by increasing liquidity? Or was it an increase in desired saving/decrease in desired investment caused by something else?

Avon: we need to distinguish between rents per acre of land, and rents per dollar of land. If people get news that leads them to expect a faster future growth rate of rents per acre, the price of an acre of land would rise, given constant real interest rate, risk, and liquidity.

[baconbacon]

I am not sure how (or if) it applies to your question- but the liquidity of houses (in the US) as a form of collateral has been the lowest since the great depression if you are talking about actually having to use a house as collateral. Backlogs in forclosure activity are still high and somecourts have passed judgment preventing eviction due to statute of limitations issues.

If the liquidity of housing has increased in the US it has done so inthe face of 7 years of historically high costs of claiming that collateral.

[Roger Sparks]

Adam makes a good entry "Let's imagine the consumption/saving decision with 3 assets, stocks, bonds and money".

MMT advocates would say that a saving decision has been made by government if government runs a deficit. This savings event must occur because there is no way the money can disappear (except into bonds) unless it remits back to government through the tax mechanism (and we just agreed that a deficit occurred so the tax mechanism did not fully deploy).

So what are we thinking when we ponder if the cause of increased saving/lack of investment might be voluntary on the part of the private sector?

[Jussi]

"For the same yield I buy more bonds but also I save more. This is because in my portfolio one part got more desirable but the other 2 didn't get any less desirable, so my asset portfolio is more desirable on the whole. Thus, before I was indifferent on the margin between an extra penny of saving or an extra penny of consumption, now I must want more consumption."

Is this correct or just a morphed fallacy of composition? The price went up but why there would be more bonds available? There is a seller for every buyer for the existent bonds. In other words the return for the savings went down against the higher desire.

The price appreciation might mean more bonds in the end but there is a lot of moving parts.

[Avon Barksdale]

No, Nick, this is an important point. Price/dividends or price/rents forecast something. As an identity, high price/dividends ratios mean:

1) Investors expect dividends or rents to rise in the future.

2) Investors expect returns to be low in the future.

3) Investors expect prices to rise forever.

Remember, this is a theory independent statement, it’s an identity. It is now an empirical question to determine which of these three capture asset pricing. Well, it turns out that almost all the variation in price/dividend ratios comes from variation in expected excess returns. It does not come from expectations of higher cash-flows and it does not come from lower interest rates. High prices relative to dividends or rents reflect lower risk premia – lower expected excess returns. The last 30 years have taught us that most returns and price variation come from variation in risk premia, not variation in cash-flows and interest rates etc.

[Nick Rowe]

Jussi: Interpret what Adam said as shorthand for "For the same yield I desire to buy more bonds but also I desire to save more." He is talking about a shift in the curve. Whether the effect of that shift affects price, or quantity, or both, depends on the elasticity of the other curve.

Avon: yes, I get that identity.

"High prices relative to dividends or rents reflect lower risk premia..."

Do we know it's a risk premium rather than a liquidity premium? Or a lower marginal rate of time preference?

[Avon Barksdale]

"Do we know it's a risk premium rather than a liquidity premium? Or a lower marginal rate of time preference?"

That's why I suggested reading Cochrane's article Discount Rates.

Remember liquidity is about information sensitivity, the less information sensitive an asset is, the more liquid it is. Assets that are illiquid have an embedded option that most policy or business analysts leave out in reporting. For example, if you try to buy a McDonald's franchise, you will see that the discount rate on the operating cash flow is nearly 20%. Now McDonald's is not that risky. The reason for the high discount rate is that you can't sell a franchise to anyone you want and there are encumbrances on the sale. The reduction in liquidity, the information constraint set up by the corporation to protect itself, creates and option value. The appearance of the 20% discount is because people report the cash flow only, and not the value of the implicit option. If you reported the option value, you'd find a much lower discount rate on the operating cash flow. The low price for the cash flow results because you also have to give up a valuable option to buy the franchise.

[Matias Vernengo]

Correct me if I'm wrong, but the reason you can equate a low real rate of interest with secular stagnation, which presumably is simply slow real rate of GDP growth, is because there is a sense in which there is a rate of interest that would equilibrate full savings and investment at its full employment level. That's for the IS or savings glut hypothesis. Your alternative about the Canadian rate being negative, related to liquidity preference, is a policy choice one. But if any Wicksellian mechanism is at play, then both should be equalized and the former, the IS or natural one, should be the attractor of the monetary one. In this case the liquidity hypothesis and the IS would be two sides of the same story, wouldn't they? I have other problems with the IS story of course.

[Nick Rowe]

Matias: If the central bank targets a high enough inflation rate (or NGDP growth rate) there should be no problem with having a low/negative real rate for the IS and a high/positive nominal rate for the LM. In New Keynesian terms, the central bank is able to set an interest rate that is compatible with the natural rate without going below the ZLB.

The cruise control setting is normally an attractor for my car's speed. But not if the engine is misfiring badly going up a steep hill (or if the mechanic wired it up the wrong way around so it presses the gas down further if the speed increases). (Not the best metaphor, but it gets the idea that attractors only work withing limits, and stability is not assured.)

[Alex Godofsky]

Nick:

louis: take an extreme case. Suppose Canadians started using government bonds exactly like currency. Since the Bank of Canada targets 2% inflation, and the nominal interest rate on currency is 0%, we would expect to see the nominal interest rate on government bonds fall to 0% too, so the real rate on government bonds becomes minus 2%, just like currency.

Doesn't this mean that the "real interest rate" is nothing but the liquidity premium of currency over bonds? But that can't be right. Isn't there a problem with your hypothetical, in that if bonds are fully liquid then the BoC can't hit an inflation target (since OMOs are well and truly irrelevant)?

[Nick Rowe]

Alex: True, the BoC would need to use something other than OMOs.

There is not *one* real interest rate.

[Alex Godofsky]

re: real interest rates, I agree fully, I was just playing nice and pretending inflation exists and is a real thing since that seemed to be an assumption of your post

re: something other than OMOs

It would seem to have only two options:

1. Find a way to pay (or charge) interest on currency, and then do normal macro stabilization with that interest rate

2. Engage in "fiscal policy" by buying risky assets
2b. Do so specifically via something like an exchange rate target

[Too Much Fed]

What about a definition of liquidity?

[NoniMausa]

I Am Not An Economist, but isn't an increasing interest rate based on the expectation of increased prosperity of the population at large? Not necessarily the expectation of the borrower, but certainly of the lender.

If a large lender expects his borrowers to be earning tomorrow the same or less than they are earning today, and if the lender perceives that their effective earnings will also slowly decline, and perceives that government supports for their dependants (whether children, elderly, or unrelated indigent) are also being slowly eroded, then the lender would not be wise to extend large loans to them.

And of course this is exactly what current lenders see. Employers have, for at least a couple of decades, worked hard to ensure that employees incomes and benefits do not grow, and that fewer people do more work for the same or lower wages. Supports for child care are not lavish, though thankfully elder care and health care are better supported. But other services have been whittled away. So, for many Canadians prosperity is essentially flat. Seems to me flat interest rates must follow flat expectations.

In such an environment, lenders can better increase their return on capital by increasing service charges, cutting staff, and in other ways setting turnstiles in the flow of capital. I'm not sure what it would take in the way of world affairs to change this. Alien invasion, maybe?

[Neil]

It seems to me that you don't learn anything new about "declining equilibrium real interest rates" by looking at "rent/price ratios." Declining rent/price ratios are an artifact of lower interest rates, as the value of holding land is the spread between the rent on the land and the return on alternative investments. As the return on the alternatives decline (declining real interest rates), the price/rent ratio can also decline, assuming there's no change in the premium being demanded to hold real estate.

The US exception in residential real estate is likely a result of an increase in the perceived risk of holding real estate, so the premium demanded has risen.