Wednesday, February 27, 2019

Quick note on Interest on reserves and inflation expectations

I've been working on the follow up book to Shut Out.  Some version of this graph will probably be in it.  It shows that the deep drops in equity values didn't come from the disastrous September 2008 Fed meeting after the Lehman Brothers failure.  They came during the period when the Fed began paying interest on reserves.

The story in a nutshell is that the Fed had the target interest rate pegged at 2%, which was far too high at the time.  In order to maintain the peg, they would have had to sell every Treasury on their balance sheet.  So, in order to suck cash out of the economy and maintain their interest rate target, first they asked the Treasury to issue T-bills and deposit the proceeds at the Fed, to fund emergency loans they were making to panic-stricken financial firms so they wouldn't be inflationary.  Basically, the Treasury was selling T-bills so the Fed wouldn't have to.  Then, when the Treasury had deposited hundreds of billions of dollars at the Fed and was balking at borrowing more, the Fed began paying interest on reserves, so they could effectively borrow directly from the banks.

Now, the point of this was because they were afraid of creating inflation when they made emergency loans.

In this graph, I also have the implied expected 5 year inflation rate, from TIPS markets.  For much of the time they were sucking credit from the banking system, 5 year expected inflation was negative 1%-2%.

Saturday, February 16, 2019

Housing: Part 343 - New Homes and Vacancies During the Boom and Bust

There may be a similar post to this one back in the previous 342 housing posts, but something I happened upon today reminded me of it.  Price trends among sold homes, vacant homes for sale, and existing homes give a clue about what happened during the turn in housing.

I haven't shown it here, but the median existing home price didn't rise as much in 2004-2005 as the mean home price.  That is because expensive cities were getting more expensive, so the distribution of prices was becoming more skewed.  The rise of prices in the most expensive places caused the average to rise more than the median.

But, here we can see that the median new home was slightly declining in value relative to the median existing home, especially in 2005-2007.  That is because Americans were not adding more expensive homes to the housing stock.  They were adding less expensive homes to the housing stock, compared to past trends.  That is because the housing boom was facilitating a decline in housing expenditures the only way it could, by creating compositional shifts of population to less expensive places.  Having a building boom in less expensive cities is, in fact, the only way to reduce aggregate housing expenditures in a Closed Access context.  We can't have a building boom in the expensive cities, and rents aren't going to moderate if we slow down building.  A building boom is the only way to do it.  And, it was working.

Since the crisis, the median prices of new homes has moved much higher because we have used mortgage suppression to slow building down and to reduce ownership in low-tier markets.  Because building is the way to reduce housing expenditures, rental expense has remained level for homeowners while rent for non-owners has continued to take a larger portion of their incomes, since the crisis.

Also, note the measure of the median asking price compared to the median existing home price.  It started to rise in 2005.  This was during the mass exodus from the Closed Access cities.  At the same time, vacancies rose among non-rental homes, and inventory of homes for sale was also increasing, suggesting that sales were becoming more difficult to come by.  But, note that during that time, the average price of homes for sale was rising.  This suggests that the inventory was at the high end.  It also happens to be the case that during that time, rates of sales and prices were slowing more rapidly in high end markets within each metro area.

This continued to be the case through 2007 and 2008 when defaults started to rise.  That is because it wasn't low end borrowers defaulting that caused vacancies to rise.  It was a change in sentiment at the top end.  The top end fell first.

The median asking price of units for sale has remained elevated because of the mortgage suppression.  Today there aren't as many sales at the low end, and many low-end households are sort of grandfathered into their units, and can't readily sell and buy into another unit, either because credit is tight or because they lost equity in the crisis.

Friday, February 15, 2019

January 2019 CPI Inflation

Non-shelter inflation came in relatively close to the Fed target this month, preventing non-shelter Core CPI inflation from declining too far as the hot January 2018 figure dropped off the back end.  Core non-shelter inflation fell from 1.5% to 1.4%.  Shelter inflation is holding up at about 3.2%.

So, we continue along at low rates of non-shelter inflation that aren't disruptive, in and of themselves, but if they decline, will probably find accommodation to be tardy because of the supply-heightened shelter inflation.  The same story that has been the case for several years, really.

The inverted Eurodollar futures yield curve between now and 2021 and the leveling off of mortgage lending and home sales suggest we are moving in that direction, but of course some indicators continue to be strong.

Thursday, February 7, 2019

Upside Down CAPM: Part 9 - The mystery of long term returns

Timothy Taylor has a post up about long term returns.

There is this:
In real terms, the "safe" rate doesn't look all that safe.
Indeed, if you look at the "risky" assets like housing and corporate stock, but focus on moving averages over any given ten-year period rather than annual returns, the returns on the "risky" assets actually look rather stable.

May I suggest the upside down CAPM model?  "Risky" assets earn a relatively stable *expected* return, which is whipsawed by real shocks to cash flows.  Over longer time frames, the shocks tend to wash out, and the expected return approximates the realized return.  (Mainly here I'm talking about equities.) "Riskless" assets have an expected return that is more volatile, and reflects a discount from the stable expected return on at-risk capital, which shifts with sentiment and on-the-ground reality.  They have more stable short term cash flows, but long term returns that can fluctuate.

There is basically a risk arbitrage between volatile cash flows and the expected return on stable cash flows.

He discusses the r>g issue.  Upside down CAPM says that (at-risk) r is relatively stable.  When g is higher, then r and g tend to converge, and risk-free r rises with g and converges with at-risk r.  If we're worried about r>g, upside down CAPM says to increase g.  Mostly, that can be achieved with something like NGDP level targeting that minimizes nominal income volatility, reducing the discount that must be taken to avoid it.  I predict that under NGDP level targeting, debt levels would decline, real long term interest rates would rise, and average income growth would rise.

Wednesday, February 6, 2019

Upside Down CAPM: Part 8 - Deficit Spending isn't stimulative or inflationary

Modern Monetary Theory (MMT) - not to be confused with market monetarism (MM) - has been a popular topic lately.  I have some thoughts on the matter, which I will lay out here.  I ask for generosity from the reader, and for corrections in the comments if I declare something here that is demonstrably wrong.  I don't have a deep understanding of MMT, and this isn't meant to be a critique of it, but the main issues that seem to form the core of MMT thought are related to some ideas that have been floating around in my head that probably aren't good for much more than embarrassing me, but I want to air them out.

As I have mentioned in some previous "Upside Down CAPM" posts, I think it is best to think of safe debt as a service provided from the borrower to the lender.  The service of delayed consumption - low risk saving.  This is the primary motivation for the aggregate use of debt in developed economies.

Considering this, I think it is best to think of public debt as a service the federal government is particularly capable of providing.  Since it can provide the safest form of deferred consumption, it gets to "sell" it at the highest price (bonds with the lowest yields).  This is wholly separate from the question of budgets and spending.  So, it is best to think of government deficits as two separate acts.  First, the act of taxing and spending.  Second, a debt transaction.

So, in this framework, all spending is funded by taxes.  Whether it is stimulative, inflationary, etc., stems from the spending itself, funded through taxes.  When that happens, capital (in both real and nominal terms) is transferred from private to public hands, affecting aggregate decisions about investment, spending, etc.

Now, if the government decides to engage in deficit financing, there is a second act.  This is purely nominal.  When it sells Treasuries, it simply creates offsetting accounts - an asset account in the private domain and a liability account in the public domain.  The creation of these accounts is purely nominal.  No real capital shifts as a result of this accounting.

In the aggregate, this is no different than imposing a tax.  Within the private sector, it is a decision to delay the distribution of that tax.  But, in the aggregate, the real capital was removed from the private sector when the spending was triggered.  If the government taxes a different individual in the future to pay back the bondholder or just defaults on the bond, the first order effect is the same.  The accounts are simply erased, and just as when the Treasury bond was issued, there is no aggregate effect on the use of real capital.

Ricardian equivalence is usually referenced here as a source of stimulus or lack thereof.  The idea is that the creation of those accounts affects the private sector's notion of its own wealth.  If it fully internalizes the cost of future taxes, then the issuance of the bonds isn't stimulative.  If it doesn't, then the bonds are stimulative, because they trigger new spending from this perceived wealth.

But, I think Ricardian equivalence is not particularly relevant.  The private sector, in the aggregate, can't spend those Treasuries.  It might be able to use them as collateral for private borrowing, which then can stimulate spending.  But, then the spending is coming from the growth of the money supply, which is under the control of the central bank.  The central bank will be managing its own targets regardless of deficit management, so any inflationary or stimulative effects from that will be offset in the natural course of monetary policy management.  Whether any spending is facilitated by the existence of treasury bonds, other assets, or simply growth in base money, is not particularly important to the question of whether public spending or borrowing is either stimulative or inflationary.

There is the issue of foreign savers.  In that case, the distinction is that they are outside the domestic tax base, so the consequences of future taxation are more complicated than simply a redistribution within a stable aggregate.  In that case, the first order effect of a default would benefit the domestic balance sheet.  But, still, it seems to me that the margin on which the effect of the debt rests is whether the interest rate is lower than the domestic income growth rate, so that the eventual tax will be paid with fewer dollars, relative to national production.

As long as long term income growth is higher than the rate of interest paid on the bonds, this process is beneficial because of the public ability to profit by selling deferred consumption.  The benefit doesn't come from the deficit itself, but from the government's ability to provide this service better than the next best provider.

In terms of thinking of public spending on the margin, that spending is useful or not useful, regardless of whether it is funded by taxes or bonds.  Practically speaking, some public spending might provide a very high return, and much public spending doesn't provide a return at all.  That's not the point of some spending.  Most of the growth in income isn't the result of public spending at all.  The ability of the government to gain from providing the service of deferred consumption is unrelated to the benefits or lack thereof of public spending.  And, even the ability of this service to lower deadweight loss by shifting taxes to wealthier future taxpayers is only partially related to the spending it funds.  The income growth that reduces that deadweight loss can come from effective public spending.  It could also come from regulatory decisions that aren't related to spending, or it could come from private sector innovations that have little to do with public spending.  If an unknown Mongolian tinkerer invents a perpetual motion machine next year, the entire globe will eventually become much richer as a result, and we will have benefitted for having facilitated deferred consumption purely because of the positive shock created by the Mongolian tinkerer.

The upshot of this is that deficit spending should have little to do with cyclical considerations, except to the extent that an economy with either cyclical fluctuations or secular malaise will be correlated with a high demand for safe assets.  But, it is much better for everyone if there is more demand for making risky investments, in which case, it would be more likely that income growth would be high and Treasury rates would be high, and the budget deficit would naturally be falling because of rising incomes, as it was in the late 1990s.

Whether it is a Keynesian or an MMT framework, the idea that funding spending with bonds versus taxes can be stimulative or inflationary seems questionable to me.  And, the idea that spending, in general, is stimulative or inflationary seems questionable.  The devil is in the details.  Spending should be done because that specific spending is useful, regardless of cyclical matters, and cyclical stimulus should come from monetary policy.  It is probably useful for some developed nations like the US to maintain a significant amount of public debt, but not as a cyclical governor, rather as a public service to risk-averse savers.  But, at the same time, fiscal policy should aim to reduce the risk-aversion that leads to the demand for that service.

Certainly, if something like this is beneficial, it should be done during economic downturns, but there is no reason this should be treated as a cyclical governor.  There is no reason to leave these hundred dollar bills on the floor during economic expansions.  It is just a double-entry accounting entry.  It isn't expansionary in and of itself except to the extent that it lowers deadweight loss, and that is something we should always aim for.  So, hypothetically, the proper level of public debt is the level that maximizes the value of this service, which has mostly to do with the ability to pay the interest from future income.  This is little different than the process a private firm would use to arrive at a target capital base, where generally the level of debt is the level that markets will fund without creating default risk that increases the credit spread that the firm faces.  Obviously, the failure of a nation is much more significant than the failure of a firm, so the limit should be set where default risk is highly unlikely.  Yet, that might be a relatively high level.

Sunday, February 3, 2019

January 2019 Yield Curve Update

I have discussed how there is a sort of mental accounting problem with the yield curve model.  The zero-slope is treated as a constant, when, in fact, meaningful inversion happens at low yields when the 10 year yield is as much as 1% higher than the fed funds rate, and at higher yields, the inversion has to become fairly steep to become meaningful.

During the past two months, the curve has become meaningfully inverted.  Here, in the Eurodollar futures market, the upward bias of the longer term yields is clear.  What is important is that forward rates in the 2-3 year time frame are inverted.  I suspect those 2021 Eurodollar contracts will close at rates much closer to zero.

Here is the plot of the Fed Funds Rate against the 10 year Treasury, shown with the adjusted inversion levels.  From this point, a normalized yield curve is highly unlikely to develop without lowering the Fed Funds Rate.  Expect the 10 year yield to be below 2% by the time that process is finished.

Friday, February 1, 2019

Housing: Part 342(A) - Building Homes Helps

Quick follow up to the earlier post.

I should have added in the numbers for the US, for more perspective.  Here, I have also added Atlanta, which I would point to as an example of a city that was growing and has generous local building policies, but doesn't have as strong of an income trend as Austin.

I apologize, though.  This first graph is getting a little messy with all this new data.  The messier lines are for housing permits per thousand workers, on the left scale. Per capita income is the right scale.

Austin = Blue
Seattle = Red
San Francisco = Green
US = Black
Atlanta = Purple

Points I would make.
  • Ample building is one reason why Austin and Atlanta have per capita income levels near the national average.  Mobility is a key equalizing factor.
  • Notice that home prices in Austin have trended higher since the crisis compared to the national average.  Atlanta has trended lower.  So there are three basic trends here:
  • San Francisco (and Seattle to a lesser extent) had high incomes and high home prices during the boom, and have continued on those trends since the crisis.  Building levels have recovered to their boom levels in both cities (low in SF and moderate in Seattle).
  • Austin had average local incomes and home prices during the boom and both incomes and prices have risen since the crisis.  Building has recovered to its strong boom level.
  • Atlanta had average incomes and prices during the boom and lower income and home prices since the crisis.  Building was strong during the boom and has been weak since the crisis.
The takeaway from these trends is that the housing boom was facilitating aspirational mobility.  But, since there is a supply limit in most of our most prosperous cities, in those cities, aspirational mobility can only lead to a bidding war on prosperity and a segregation by class and income.  Homes could only be built where it is legal to build them.  Incomes were high where housing obstruction causes that segregation to happen but building was high where it isn't so obstructed.  Seattle is in the middle between Austin and San Francisco, and where it ends up will be determined by supply.

There has been much misplaced kvetching about overinvestment in housing during the boom.  Notice how, of these cities, Austin is the only one that shows a real reaction in building permits.  That's because it has great economic promise and it doesn't restrict housing growth.  Imagine if Seattle and San Francisco had supply reactions like Austin did!  Imagine if housing permits could have doubled in those cities between 2002 and 2005.  The deep and undeniable irony is that if those cities, and others like them, had local housing supply that was responsive to demand like Austin's is, there never would have been a national moral panic about overinvestment in housing.

By the way, also note that, even with all of that building in Austin, even as late as 2006 and 2007, there was no price collapse there!  I can't tell you how much the literature on this topic is tainted with the presumption that the price collapse in non-bubble cities is the result of overbuilding.

Now, look at Atlanta.  There was little change in the rate of building in Atlanta during the boom.  Atlanta had a sort of steady-state rate of growth - a strong rate of in-migration - that continued during the boom.  And, then the bottom dropped out, and it didn't recover.  Building is still very slow in Atlanta, local incomes have dipped, and home prices dropped below the national trend.

That is because the moral panic that began in 2008 led to a public imposition on entry-level and low income home buyers.  The sort of buyers that used to be able to move to Atlanta aspirationally.  The borrower-based lending constraints that have been in place since 2008 that have kept first time homebuyer rates low and have reduced homeownership among young households and households with low incomes killed the Atlanta housing market in a way that high-flying Austin had the escape velocity to outrun.  Within each metro area, a comparison of building rates and prices between high tier housing markets and low tier markets is similar to the comparison between Austin and Atlanta.

As a result of all of this, building pre-2008 was strongest where incomes were moderate and building post-2008 is strongest where incomes are high.  Pre-2008 was defined by constrained supply and post-2008 is defined by constrained demand.  The ideal would be a housing market similar to Austin and Atlanta in 2005, that is relatively unconstrained by either supply or demand, and Americans are in fairly universal agreement that 2005 Austin and Atlanta is to be avoided at all costs.  And all costs is exactly what we're getting.

Housing: Part 342 - Building homes helps

I pulled up these charts today while responding to an e-mail, and they seem worth sharing.

This is Austin, Seattle, and San Francisco.  They each are cities with high demand for population growth and strong income growth.  They really make a nice example of how housing supply works.  When there is high demand for living in a city, it can block growth, like San Francisco, which may actually increase local incomes because of the obstructions to competition in the local labor force, but those higher incomes are generally claimed by higher housing costs.  And, the pressure is especially strong on households with lower incomes, who end up moving away at a high rate.

It can grow tepidly, as Seattle has done, which is enough to minimize the migration, so it stops the worst of the outcomes of blocked housing supply.  But, costs still move up a bit.

Or, it can grow boldly like Austin.  Austin brings in migration and offers strong income growth and the willingness to share it.  Not only are all sorts of Americans moving to Austin, but they get to keep more of their incomes when they get there because housing is more affordable.

There are rumblings of Closed Access policies in Austin, but so far their housing policy has been commendable.  If those who oppose change ever do get the winning hand in Austin, two things are certain:

1) Austin will become more expensive and will offer economic opportunity that is only accessible to richer households.

2) Those who support Closed Access policies will declare that supply and demand is an oversimplified model and building more homes in Austin couldn't possibly cure the cost problem. (It looks like they have already started.)

Follow up post.