Full steam ahead on 3 cylinders

JOLTS data continues to look strong.  Growth in the major categories continues to be strong (the slopes of the weighted moving averages), and the rates are basically at full employment levels.  Because a labor force that skews older will have lower turnover, hires and quits are topping out slightly lower than in the previous recovery and job openings are slightly higher.  It would be tempting to blame this on structural issues (read: my political hobby horses), but these small changes in rates are about where a back-of-the-envelope estimate of demographic effects would suggest.  Older workers, in aggregate, have much lower unemployment rates and much longer unemployment duration, which means dramatically lower employment churn.

But, the New York Fed's Household Debt and Credit Report showed no growth in housing debt still in the first quarter, in spite of some signs that mortgages have begun to grow.

Until mortgages show sustained growth, annual growth in bank credit probably will be limited to less than 10%.  If that is where we stay, it seems like we could have decent GDP growth.  I suspect that non-shelter inflation will remain tame if that is the case, but I am not sure about that.  I don't see how long term interest rates would be able to rise in that scenario, though.  It's strange that FOMC members still talk about low interest rates as if they are stimulative, when clearly the asset class that is unable to gain traction is housing and clearly interest rates are not the constraining factor there.  I am afraid that if mortgages don't move, long term interest rates will collapse, and the Fed will interpret that as stimulative, giving them more confidence about raising interest rates on reserves.

What a strange economy this is.  Housing, education and health care have been eating up our productive capacity for some time.  The details differ for each sector, but each is giving us a declining share of real output while we keep spending the same or more.

There is so much talk about inequality, usually revolving around confiscation and redistribution, and a shared feeling among progressives that those who balk at that are simply heartless.  But, what good would redistribution do us in the face of these problems?  We basically have pushed ourselves into a banana republic scenario, where those who own residential real estate capture huge excess rents because of barriers to entry.  Why would we solve that problem with redistribution?  All the redistribution in the world isn't going to build any new houses.  In fact, it will make the problem worse if it discourages what little new building we have now.

Sometimes I feel very grateful that we have somehow overcome our terrible intuitions about emergence for so long that so many of us are able to enjoy the fruits of abundance that come from the modern global marketplace.  But, then I look at issues like this.  There is such a widespread distrust of prices and market participants, and a willingness to impose disorder.  We console ourselves that even many economists as late as the 1970s supported price controls, and that policy norms have come a long way.  But, how is the current housing situation any different?  There aren't price controls in terms of forcing homes to sell at target prices, but there basically is a consensus among the public and among economists that if home prices reach some general level, we have to correct them, even at the expense of causing the worst economic downturn in nearly a century.  At least the economists in the 1970s were hurting the economy by accident.  They weren't imposing price controls knowing that they were creating significant dislocations.  To simply suggest that we let home prices float, mortgage credit flow at the discretion of private lenders and borrowers, and housing supply to expand at least enough to end 35 years of housing inflation is, as far as I can tell, generally considered to be an extreme opinion.

And, it also seems to be widely assumed that economic rents are the result of an elitist cabal directing Washington from smoke-filled rooms.  But, the economic rents going to current home buyers are the result of the collapse of the mortgage market.  The consensus has the whole story so completely bass ackwards.  I'm pretty sure BlackRock Blackstone wasn't secretly directing the Fed to tank the banks so they could spend the next decade buying up cheap real estate.  The rents have come from the demands of the consensus.  A robust mortgage market would cause the rents to be bid away.  It's certainly not the financial sector that is demanding that we tighten up lending.

And, look at education.  In that case, it's not so much a case of capitalists pocketing extra profits.  It's just a matter of bloated post-secondary administrations, over-consumption, and primary and secondary schools that are insulated from competitive pressures to be more responsive and efficient.  So, there are millions of middle class wage earners in that sector who are doing work that appears to be useful, but is really only necessary because of inefficient providers and subsidized demand.  At the end of the process, the Federal government shuffles students through colleges with little regard for outcomes, so we end up subsidizing those who will become the highest wage earners in the economy and saddling future low wage earners with debilitating, pointless debt.  I suppose redistribution can alleviate some of these problems, but is it really a heartless, extremist position to address the obvious problems at their roots?  After the decades (centuries!) of damage that publicly disbursed education has inflicted on marginal communities, how can it be that the "social justice" label is most closely affiliated with people who want to prevent parents in those communities from exercising choice?

And, also, with health care.  We have certificates of need requirements for hospitals, longstanding severe barriers to entry for physicians, state level committees that must approve insurance prices and prices that basically don't exist in the actual delivery of the service because of decades of federal meddling, and such a disconnect between consumers and providers that there is massive over-consumption of absurdly expensive health services that no rational consumer would demand.  These are obvious problems that we all know about.  Yet the debates, again, are about redistribution.

Time and again, we arbitrarily hamstring supply with public policy and our public solutions are to subsidize demand.  I think I can safely say that in the next election, no major candidate will be running ads pointing out how insane it is that existing hospitals get to decide if a hospital can open up or add new services.  That problem, along with many of these supply side problems, is a clear case where the problem is that there aren't enough markets or markets aren't trusted enough.  We all walk through this health system watching thousands or tens of thousands of dollars get spent where reasonable outcomes could be achieved for a small fraction of that cost, but it is basically illegal for anyone to provide those services to us in a sane manner.  The solution to this problem isn't to tax the doctor so that a poor and sick person can spend those pointless tens of thousands of dollars.

The inhumane position is to burn political bridges in battles over redistribution in the face of these preposterous supply problems.

Mortgages, QE, and recovery

I have been watching for strong growth in mortgage lending as a signal of stronger real and nominal economic growth and interest rate increases.  There have been some signs of strength, but, so far, April real estate loans on the balance sheets of commercial banks has been weak, after a positive trend change through the winter.  (March was boosted by a technical change in the data, but still appeared to show a continuation of acceleration, which has disappeared in April.)  It could be that the MBS market is recovering, and a high proportion of net new mortgages are not staying on bank balance sheets.  Less than a quarter of residential mortgages are on bank balance sheets.  This is an easy weekly number to check, but it is incomplete.

The number for 2015 1Q housing debt in Tuesday morning's Household Debt and Credit Report may shed some light on this.  The Mortgage Bankers Association is showing strong Q1 activity, in both single family and multi-unit housing.

Here is a graph of the marginal change in Commercial & Industrial Loans and Closed End Residential Real Estate Loans on Commercial Bank balance sheets since the end of 2007.  I think we can see some of the idiosyncrasies of this recession in this data.  Notice that C&I Loans have grown at a very steady rate throughout the recovery, regardless of monetary policy.  I believe that is still the case, and that C&I Loans will continue to expand with little effect from early rounds of interest rate increases.

The real estate loans behaved differently.  Keep in mind that they should have been increasing at a rate slightly higher than C&I Loans.  I have previously wondered about the decline in bank lending concurrent with QEs.  But, on further reflection, I think what we are seeing there is that during QEs, cash was making its way into the real estate asset class. All-cash investors were buying homes from overleveraged households.  This process was accelerated during QEs.  As QEs were ended or tapered, this activity declined or grew at a slower rate, and new mortgage originations once again would begin to push mortgages outstanding higher.  This happened as QE3 tapered, but mortgages at the banks leveled off at the end of 2014.  A new rise in these loans is one signal I am looking at.  This recovery is mostly about getting capital back into real estate.

Over the past year or so, the Fed has been testing out its Reverse Repo program, which is a roundabout way of pulling reserves back out of the banks and selling treasuries back into the private market.  Reverse Repos on the Fed balance sheet increased by about $150 billion.  In effect, they temporarily sold $150 billion of treasuries back to private investors.  Do any readers have any thoughts on this?  Could this program have had a disinflationary effect that might explain some of the nominal pullback of the past several months?

In addition to the several mortgage indicators I am watching, I noticed another piece of bank balance sheets that appears to have systematic behavior - the portion of bank assets held in securities in bank credit, which is mostly treasuries and agency securities.  Banks buy more of these low risk securities during stressful times and decrease their relative holdings during expansions.  The peak in these securities tends to come within a few months of the first interest rate hike.  (In the graph, the proportion of bank assets in these securities is inverted, to help show the pattern.)  This seems intuitively reasonable, as both the rise in risk free interest rates and the decline in low risk securities on bank balance sheets would both reflect a higher relative demand for riskier investments.

In the close-up version of the graph, we can see more recent behavior.  These low risk securities peaked as QE2 was implemented and began to decrease, only to recover to higher levels after QE2 was terminated.  Then, it peaked and decreased with QE3, but with a much weaker decline.  And, then, again, as QE3 was tapered, bank holdings of treasuries and MBS began to rise again.  But, there just now are the slightest indications of a decline.  Maybe this is not really a separate indicator.  The marginal new bank assets replacing these assets will probably be real estate loans.

Update: The NY Fed Household Debt numbers are out.  Housing debt was up about $40 billion flat for the quarter.  Technically growth and acceleration, but still in rounding error territory.

April 2015 Employment Report

Today was an interesting day.  I'm not sure what to make of it.  Stocks moved up on the employment report, but interest rates moved down.  In the interest rate market, I wonder if there are two separate groups of concerns about the zero lower bound.  One concern is that the economy never achieves escape velocity.  Another concern is that the economy achieves escape velocity but the Fed puts the brakes on too soon and too sharply.  So, a day like today with a lukewarm employment report might increase the escape velocity concern but decrease the Fed brakes concern.  It hit a sweet spot where the economy looks like it continues to recover, but not so strongly to move the Fed to act.  And, the continued movement through time without a major breakdown would continue to reduce uncertainty in general.  So, we might have seen a continued convergence of the mode, median, and mean forward rate expectations today, as variances in forward expectations continue to decrease, with a the general movement in expectations being slightly lower.  The reduced uncertainty would have a positive effect on equities, whose main concern would be revenue stability going forward.  How's that for a just so story?

The slope and the expected date of the first increase have been generally moving in the same directions since the QE3 taper began, but earlier in the week, they diverged.  Forward rates had been rising because of renewed strength in the expected slope.  This week, the slope continued to rise, but rates were pulled back down because the expected date of the first increase moved back from September into October.  Friday's movement after the employment report was pretty stable regarding the date of the first rate hike.  Most of the pullback in rates today was from a slight reversal in the slope expectations.

Regarding the employment report, itself.  I think it was a bit of an aberration.  Looking at durations, we continue to see steady trends downward in long-duration unemployment.  But, very short term unemployment saw a strong rise.  This is very likely a statistical outlier.  We have several weeks of freshly declining unemployment claims.  It could be that that won't hit the unemployment figures until next month.  We should see short term unemployment drop by about 300,000 next month.  Next month will have to be an outlier for unemployment to come in above 5.2%.  That will be an interesting print heading into the June Fed meeting.  Surely we don't have to worry about a rate hike in June any more, but expectations for a September hike would firm up if August unemployment could realistically be at 5.0%.

Next is the comparison of insured unemployment and total unemployment.  We should expect the blue line to trend back down toward the normal relationship.  This month was a big jump away from that trend.

Finally, flows data.  The first graph shows the 6 individual flows that I follow.  The next graph shows the weighted moving average of the net movement between each pair of flows.

Net flows from Not-in-Labor-Force to Employment remain strong, in the same range as the strong labor markets of 2005 and 2006.

The net flows from Employment to Unemployment appear to have weakened somewhat, but this looks like it is mostly due to three months of Employment to Unemployment flows at the top end of the trend range, which are not confirmed by recent trends in insured unemployment.  Unemployment to Employment has not been particularly weak relative to trend.

So, employment looks generally positive to me.

On the other hand, This week's H.8 report now makes 3 consecutive weeks with flat levels of closed end real estate loans at commercial banks, compared to March.  So, while I think the employment trend will eventually move the Fed to raise rates, probably in September, I am still waiting for mortgage growth to confirm an expansion of nominal economic activity.  Home prices are so far out of equilibrium that a functional credit market should see mortgage growth well over 10%.  The fact that we aren't seeing that yet is bothersome.  The grapevine is telling me that, at least in Phoenix, real estate is a sellers market, so I'm not sure where things are headed.

Equities always earn more than treasuries.

Source: Shiller Annual Data

In liquid, transparent markets, we can track total returns of a number of assets.  Here is a comparison of real total returns of stocks, 10 year bonds, and 1 year notes in the US since 1950.  This combines capital gains or losses in a given year with dividend or coupon payments, adjusted for inflation.  And, this gives the typical range of returns.  The higher the returns you get, the more volatility you have to accept.  This is clear, visually, in the graph, and the chart of averages and standard deviations confirms it.


Corporate returns are based on Nonfinancial Corporations, Net Worth at
Historic Cost, Market Value of Equities, Profit after Tax (w/o CC & inv. adj.),
Credit Market Instruments, and Interest Paid.

But what if we just look at income?  For equities, this would be profit after tax, for bonds and debt this would be the yield, or an estimate of effective yield from interest expense/total debt outstanding.  Here is what it looks like in nominal terms.

The nominal comparison is uninformative.  Debt is generally nominal, so the face value remains fixed and the interest payment includes an inflation premium.  Equity is basically a real security, because the earning power of firms in the aggregate, over time, will rise with inflation.

As I have pointed out before, returns on equities are surprisingly level over time.  Here we can see some temporary fluctuations in temporal income levels through the business cycle, but over time, equities provide real returns of around 8%, which includes the 6-7% in immediate earnings plus growth expectations that tend to be slightly higher than expected inflation, providing a small capital gain premium on top of immediate earnings.  Some of the high equity returns in the 1970s must be related to macro-uncertainty due to high inflation.  And, because equity values were so low, leverage in market value terms was very high, so this could justify a somewhat higher return on equity.  But, that is a bit of a circular argument, and I don't think it can explain such an aberration.  Maybe a combination of the tax effect of high inflation, uncertainty, and the de facto higher leverage, cumulatively, can explain it.  Please let me know in the comments if there has been any academic writing on that topic.  The extremely low level of equity capital in the 1970s and the high returns to equity is a mystery to me, and I think it is an issue that is underappreciated.

We tend to think of "Wall Street" in terms of share prices, but earnings is where value comes from.  If asset prices rise relative to earnings, that is simply a transfer between future asset holders and past asset holders.  A higher price means a lower forward yield.  As we can see in the chart above, it would be quite accurate to say that equity buyers never had it better than during the Carter administration, and equity holders saw returns on investment cut by 2/3 after Reagan took office.

But, back to the original topic.  In order to really compare these asset classes, we need to adjust for inflation to report these returns in real terms.  This is trickier than it seems.  For bonds, it's easy enough.  Subtract inflation from total interest.  While the bonds were earning that interest, the face value of the principal was declining due to inflation.

At first glance, equity returns are already in real terms.  And they basically look like a real return in the graph, except for the aberration in the 1970s.  The "principal" of firms in aggregate should basically rise with inflation, so there is no need to earn an extra inflation premium.  But, net profit has had interest expense subtracted from it.  And, that interest expense had both a real component and an inflation premium component.  The inflation component was a premium paid to creditors to make up for the fact that their principal is being eaten away by inflation.  In effect, in real terms, the inflation premium of interest expense is a capital purchase by equity holders to capture a larger portion of the firm's assets.

So, while the inflation premium needs to be subtracted from the nominal yield on corporate debt in order to arrive at a real return, it actually needs to be added to net profit to arrive at the real return to equity holders.  While the adjustment to debt returns can be estimated simply by subtracting inflation from the yield, in order to make this adjustment to equity, we need to separate interest expense into an inflation expense and a real expense, and add the inflation-related interest dollars back to net profits.  For a firm with no debt, this inflation premium effect would have no effect on real equity returns.

I am a little surprised by the outcome here.  I expected equities to always out earn debt.  They do basically always out earn treasuries.  (This would especially be true of short term treasuries.)  But, since the mid-1980s, corporate debt has captured a higher premium, so that, in terms of earnings, corporate bonds earned a higher return than equities for most of the past 30 years.

Now, I could get more complicated.  To be thorough, I would need to account for defaults, inflows & outflows.  And, we should probably credit equity holders for the added benefit of future growth on their eventual returns.  But, what originally drew me into this post was thinking about international capital flows.  Ricardo Hausmann and Federico Sturzenegger made an argument that is similar to one that I have made, that the trade deficit is basically financed by the large amount of foreign equity that Americans own.  Foreigners own assets in the US with a higher historic cost, but they consistently earn less.  This is a perfectly sustainable situation.  It is mutually beneficial.  In fact, I think it is a great example of a beautiful emergent economic order.

I see pushback on their argument that, in part, argues that Americans can't expect to earn excess returns without taking on excess risk.  But that argument is based on total returns (the first graph above).  The measures of international capital flows that fund the trade deficit are based on earnings, not on total returns.  As we can see, it is perfectly normal for equity to out earn debt consistently over time.  If we marked international assets to market and accounted for capital gains and losses in the international capital flows, we would probably get a graph like the first one above.  American investors are taking on more risk.  It just doesn't show up in flows, partly because measuring it would be impractical.

I will probably get into that topic on a future post.

In the meantime, I think this issue of earnings stability is interesting.  The fact of the matter is that aggregate earnings to equity are not particularly unstable.  So, why are equities so volatile?

This is usually presented as an issue of fickleness - of greed and fear and irrational herds.  (God, how I wish in finance we could get rid of attribution error.)  But, think about bonds.  Cash flows are stable.  The only uncertainty with a bond is the discount rate, which, depending on how one frames the issue, means that the market value of the bond can change over time, or alternatively, that, on a nominal bond, the cash flows will not have the same present value that we presumed they would when we bought the bond.  Bond prices move inversely to yields.

But, equity holders own the residual risk.  So, while they also have some risk associated with the discount rate, most of their risk is in the cash flows themselves.  First, from a change in earnings in real time, and second from a change in expected growth rates of future earnings.  So, when the "yield" in equities moves down (actual earnings decline) the prices of those equities move down with it, and the uncertainty of a business cycle shock will probably mean that growth expectations also decline, further pulling the price down.

With bonds, prices and cash flows naturally stabilize total returns, but with equities, prices and cash flows create a mutual effect on total returns.  This is a bit ironic, because liquidity clearly has a strong positive effect on the value of an individual corporate equity.  Yet, the transparent markets that create that liquidity create cyclical market-wide volatility in total returns that justifies a risk premium in equities that would presumably seem unnecessary if we viewed equities simply through an earnings framework, without the volatility in principal that is made visible by those same liquid markets.

PS (added): Here is a table with average real returns and standard deviations for the four series.  If we base equity returns on market price, then returns and volatility follow a typical pattern.  But, average equity returns based on historic cost are similar to returns based on market cap, but with much lower volatility.  Based on historic cost, equities earn more than corporate bonds with less risk and they earn so much more than treasuries that the small amount of extra risk is not material.  The risk in equities comes from the strong force of changing expectations more than from changing yields.

Housing Tax Policy, A Series: Part 31 - The Market Response to a Housing Shortage

I have noted how there has been a longstanding housing shortage, which only partly abated in the housing boom, and may have had a significant part in rising home prices, including rising price/rent ratios.

Here is a graph comparing different measures of Price/Rent.  Price/Rent increased the most where rent inflation was highest - in the 10 cities that are in the Case-Shiller 10 city index.  In that post, I walked through a valuation summary where homes are like an inflation protected bond, but where the inflation adjustment comes from rent inflation instead of, say, general CPI inflation.

(edit:  As pure speculation, I wonder if the lower Price/Rent level from the Flow of Funds data is a product of new homes.  The Case-Shiller data and rent inflation data should both basically track the price and rent of an individual home.  But, the Flow of Funds data is an aggregate number for the values and rents of all homes, which would include new homes.  Since most new building happens in the suburbs, where there are fewer limits to building, most new building happens where rents are lower, and where rent inflation is also lower.  This should have the effect of lowering the relative price/rent over time of the data series that includes marginal new additions to the housing stock.  Think of the potential utility we are losing as a society because we have pushed new building to the places where it is least demanded.  Trillions of dollars worth of potential value, unrealized.)

I think it is important to think about home values in terms of an all cash purchase, because thinking in terms of mortgage funding just complicates matters, homes have an intrinsic value regardless of funding sources, and there are many ways in which intrinsic values correlate with mortgage interest expenses in a way that confuses causation.  So, in my valuation exercise, I simply considered home values on their own terms.  And, I demonstrated how home values could have risen in the way that they did because of a combination of three factors: (1) lower long term real interest rates, (2) rising rents increasing home prices before any change in Price/Rent, and (3) an increase in Price/Rent from this concept of homes as real bonds with a rent inflation adjustment factor.

Today I want to think about this same issue, but from the point of view of a renter who is a potential leveraged buyer.  If we are in a steady state, without a housing shortage, in 2003, we might imagine a context where expected inflation is 2%, expected rent inflation is 2%, and 30 year mortgage rates are 6%.  This basically describes 2003, except that rent inflation was higher because of the shortage of housing, especially in major cities.

In this steady state context, the real interest rate on the mortgage is 4%.  As I have pointed out, home prices tend to move over time so that imputed net rental income follows the same pattern as real mortgage interest rates.  So, the renter would be facing a decision to purchase a real asset with a return of approximately 4%.  To the extent that they could not fund the entire purchase, they could borrow the unfunded portion at that same approximate rate, 4% in real terms.  So, the main decision for the marginal buyer is whether they want their portfolio to include a large asset with a fairly safe 4% real return.  (For this exercise, we can ignore idiosyncratic factors about the value of the house, such as the benefit of ownership or the length of time the household intends to remain in the property.  Most households are not the hypothetical marginal household.)

Now, let's tweak this scenario so that it is more like the actual 2003.  National rent inflation was persistently around 3%.  In large cities, rent inflation was more like 3.5%.  The nominal rate on mortgages was still 6%.  For mortgage lenders, the required rate of return was not dependent on rent inflation.  Mortgages just needed to provide a return relative to general inflation expectations.  (Also, even though rising Price/Rents would have been adding potential valuation risk to mortgages, expected rent inflation would have reduced risk similarly by raising expectations of future home prices.)  So, mortgages would have demanded a 6% rate - 4% real plus a 2% inflation premium.  But, in this context, if Price/Rent remained the same, the home would be expected to provide a 4% real return from net rental income plus a 3.5% return from expected rent inflation, for a total return of 7.5%.

So, without the housing shortage, our marginal homebuyer had been looking at a 6% total return.  Now, unless home prices change, they would expect a 7.5% return on their home.

So far, this is basically the same exercise I did before.  We would expect home Price/Rent ratios to be bid up to a level where their total returns would only be 6%, which would be the price level that eliminated arbitrage profits on residential real estate.  Marginal households that would buy homes with expected real returns of more than 4% (plus 2% inflation) would bid up home prices until that was the expected return.  That would be true for a household that was purchasing the house in an all-cash transaction.

But, what if home prices are sticky, and for some period of time, mortgage rates are 6% while returns to home ownership are 7%?  In that context, a household that would require more than a 4% real return would be incentivized into home ownership because of the arbitrage opportunity.  But, these tactical households wouldn't necessarily want to tie up their net worth in a 7% investment.  What they would want to do is utilize a mortgage that cost only 6% to leverage a position on the real estate that is returning 7%.  Leverage doesn't boost the returns of a household buying a home in a normal market with equilibrium prices.  But, leverage does boost the returns of an arbitrage position.

A surge of households buying homes with very low down payments would be a sign of a supply shortage and sticky prices.  And, except for the devastating dislocation created by a monetary shock, where households were being foreclosed on and were downsizing, rent inflation has continued to rise.  It is now above 3% again, even as core inflation has slowed.  So, housing speculators would have had good reason to continue to expect continued price appreciation, since nothing has been done to cure the housing shortage, and the public policies that created so many capital losses in the housing market have only made the shortage much worse.

It's easy to write off these sorts of models based on rational expectations, because they seem to rely on that old strawman, homo economicus.  Of course, though, this is frequently how markets behave.  It is very difficult to earn excess profits in the stock market - in other words, they are bid to non-arbitrage prices.  But, you wouldn't know that by reading Yahoo Finance message boards, or even tracking buy-side analysts.  But, even if home buyers didn't look at it with the model I outline above, they were certainly looking at the same pieces of the puzzle.  They were looking at rents that were rising more than 3% per year, with no lack of available renters, with mortgage rates at 6%, and alternative investments like 20 year TIPS bonds selling at 2% yields.

The counterfactual could have been an economy where limits to building were minimized - no rent control and fewer demands on private multi-unit residential developers in large cities, fewer zoning restrictions, etc.  In that context, I think all three influences on home prices would have been decreased.  Lower rents from the extra supply would have reduced both the basis for home valuations and the effect of expected rent inflation on price/rent.  And, in the end, after many trillions of dollars of new homes would have been built, the added vehicles for investment would have helped soak up the large amount of global savings, so that, maybe, even real interest rates themselves would have bumped higher, pulling price/rents down even more.

The irony is that since none of this happened, once we get past all of these self-imposed demand shocks, it will be the housing speculators who will have been proven right.  For those who have been able to hold their properties through the crisis, they will earn solid total returns on those properties.  And, in the meantime, the public policies which enjoy nearly universal support and which created this crisis have pushed the balance sheets of millions of households into disarray, likely causing more inequality and middle class stagnation than any of the policies at the center of the arguments fueled by those topics.

Employment Preview and Real Estate Loans

Continuing claims of unemployment insurance have resumed a strong declining trend.  So, I thought it would be useful to revisit some of my employment graphs.

Here is the graph that compares insured unemployment to total unemployment.  There is a typical shape of this relationship over business cycles and over time.  The unemployment rate has remained elevated as insured unemployment has tested all-time lows, partly due to a contingent of very long term unemployed (with unemployment durations of over 2 years) and partly due to a persistently elevated level of uninsured unemployment among shorter durations.  For a while, there was a pretty linear trend of about 0.05% of the unusual long term unemployment declining each month.  It seemed like this had possibly leveled out.  But, as a close-up view of this graph shows, there was one divergence in July 2014 of about 0.3%, where the expected unemployment rate went down and the reported unemployment rate went up.  In the months before and after that, reported unemployment has actually followed very closely with the unemployment we would expect from continued claims.  If we add 0.3% to the modeled rate this month, that gives us an expected unemployment rate of 5.3%.

(On each of those graphs, the April insured unemployed level is as reported, and the 5.3% unemployment rate is manually input to show where this relationship would fall for April if unemployment comes in at 5.3%.)

Unemployment at the longer durations has been declining at a healthy pace over the past couple of months.  This pairs nicely with the recent decline in continued claims as a strong statement about the breadth of strength in the labor market.

In addition, here is the model of unusual long term unemployment which confirms the trend that has continued in the insured vs. total unemployment relationship.  This continues to converge back to historical norms as long term unemployment declines.

Here we can also see the persistent unemployment level at lower durations.  Insured unemployment is near all-time lows, but unemployment duration at shorter durations is still elevated enough to inflate the expected long term unemployment levels.  The expected level of long term unemployment should be down to 0.5% by now compared to recent recoveries, but it is still at 1%.  And, on top of that, there are another 0.6-0.7% of workers at very long durations.  It appears that the very long term unemployment has a shorter "half-life" than the persistent short term unemployment.  Some of the increased durations in short term unemployment are demographic in nature, so this may bottom out in the 0.7-0.8% range.  But, those same demographic trends should pull down employment turnover in general, so that the unemployment rate should still be capable of reaching 4% or less if the recovery is allowed to age.

Just as a clarification, I don't consider a strong labor market to be inflationary.  I think a strong labor market tends to coincide with strong real economic growth and this is related to rising real interest rates and rising real wages.  These trends tend to make the Fed pro-cyclical, since the neutral rate rises while the Fed's target rate stays in place.  It is that lag in Fed rate setting policy that causes strong labor markets to appear to be inflationary.  If there is any truth to this conjecture, this is another reason why using target interest rates as the policy tool is not optimal.  In any case, I think that falling unemployment and related rising wages (which come from reduced frictions in the labor market, not from some sort of wage inflation) will be interpreted as inflationary, and will cause the Fed to raise the target rate.  I don't think the exact target of the Fed Funds rate is that important right now.  If the housing market wasn't hobbled, the neutral short term rate would already be well above the current Fed Funds rate.  If mortgages don't expand, there won't be any inflation.  If they fail to expand, the rate at which short rates top out will be lower.  If they do expand, then the peak rates will be higher, both from inflation and from real expansion.

Source

The weekly indicator of closed end real estate loans at commercial banks has taken a breather for a couple of weeks.  (This shows seasonally adjusted <blue> and not seasonally adjusted <red>.  Seasonal adjustments are in flux right now, so both series have a monthly cycle right now.)  The April levels are about even with March levels.  It would be nice to see a good jump over the next week or two.

Interest Rates after the 1Q 2015 GDP report

There was interesting movement in yesterday morning's markets after the (mostly expected) disappointing 1Q GDP report.  Equities fell slightly as trading began, but bond yields jumped.  Considering the current context of a tentative recovery in economic production and interest rates, we might have expected interest rates to decline as a result of a disappointing economic report.

I have previously speculated that the non-normal distribution of interest rate expectations resulting from the zero lower bound may be creating odd behaviors in the forward rate curve, and I think this may be an example.

This first chart describes a context where there is a negative skew in forward rate expectations that is truncated by the zero lower bound.  So, the mode expected rates are higher than the median expected rates.  Normally, the mean expected rate, which we might suspect to be near the market rate, would be even lower, but the zero lower bound causes a distortion in the distribution of expectations which pushes the mean higher.

As uncertainty declines, which we would expect over time as we approach a given date, the variance in the distribution would decline, and the mode, mean, and median would converge.  This is one factor that I think is in play in the current forward rate market.

Now, if one were to take a naïve position on forward rates to take advantage of this adjustment over time, this would be basically a position that earns profits by taking on long-tail negative risks.  If some very negative economic development comes to pass, rates will decline sharply.  But, as long as the economy progresses somewhat reasonably, these profits will accrue.  It's the proverbial nickel in front of the steamroller.

I am using growth in the mortgage market as a signal of economic expansion, so I am taking this position with the idea that the mortgage signal gives me insight into the likelihood of failure.  The hope is to capture the gains from the trend but to exit the position if the mortgage signal weakens.

This next graph is a very simple example of the influences we might see on rates.  The orange line is a hypothetical forward yield curve, based on rates rising in September and climbing at a somewhat conservative rate of 35 basis points per quarter.  The blue line is a bimodal distribution, one set of investors expecting the orange line and one set of investors expecting us to remain at the zero lower bound.  Here, I have set the expectation of persistent ZLB at 50%.  I'm sure that is higher than the market expectation, but for now I'm not too concerned with getting the numbers exactly right.

In the next graph, the blue line remains, and it compares to today's opening forward rates (the purple line).  The red line is the high mark in the forward rate market after the GDP report today.  The dark blue line is a hypothetical yield curve with adjusted expectations.  The expected date of the first rate hike didn't seem to move today.  But, the GDP report would, counterintuitively, have had two countervailing effects on forward rate expectations.

First, the disappointing numbers would have reduced the expected slope of rate hikes.  But, second, the fact that the report wasn't catastrophic would have reduced the variance in expected outcomes.  In this simple model, that would mean that the percentage of investors with the ZLB expectation would fall.  If the change in rates after the report reflects no change in the expected slope, then a reduction in the ZLB percentage of 3% would create today's rate changes.  If the change in rates after the report reflects a decline of 2 basis points per quarter in the slope, then a decline of 8% in the proportion of investors expecting a persistent ZLB would pull us up to the later market rates.

So, my point is that as time passes and we avoid a complete meltdown, rates should move naturally up from the blue line to some version of the orange line.  I also think there is another factor at work here, which is that the lack of nominal real estate investments is creating an oversupply of savings in the long term bond market, and this is dragging down forward rates in general, especially at the long end of the curve.  So, right now, the orange line probably reflects a lower slope (and, there is also probably a smaller probability of remaining at the ZLB).  But, in addition to seeing the median and mean expected rates move up toward the orange line as the ZLB expectations decline, we should also see the orange line move up, too, as savings finds new outlets in the real estate sector.  This could happen through new building or mortgage growth, and it is clearly a convergence that will happen in a functional recovery from this point.  Returns to real estate are far outside the historical norm compared to bonds.

Since each forward rate is the product of its own set of forward expectations, there is a chance that news of persistent growth in new home building or mortgage expansion could cause fairly large sudden shifts in this direction.  If recovery persists and mortgages expand, an eventual rise of around 2% or so at the long end of the yield curve seems like a reasonable expectation.

Eurodollar Futures and Mortgage Expansion

As the FOMC finishes their meeting, forward interest rates are at a fork in the road.  Since the end of 2008, when short term rates went to the zero lower bound, the expected date of the eventual rate hike has generally remained stable during episodes of QE and has moved forward in time when QE was off.  When QE3 began the expected date was around June 2015.  There has been some volatility in that expectation, but when QE3 ended, the expected escape date was still around June 2015.  This is some pretty solid evidence of QE3's success, but I wish it had been continued just a little bit longer.  Because, by ending it with some time remaining to the expected rate rise, the Fed has left open the possibility that economic stagnation would begin to push the date back again.

In fact, that is what has happened.  We can see in the first graph that through the course of 2014, the expected date of the rate hike moved in a fairly tight range around June 2015.  The second graph shows the date of the expected rate rise, both in terms of the number of quarters from 2012 and the number of quarters from today's date as we move through time.  As we can see, since the beginning of 2015, the expected date of the first rate rise has been moving forward in time on a 1:1 basis.  We are no closer to the date of the first rate hike now than we were in mid-February.

I think regulatory and market developments in the mortgage market are more important now than FOMC forecasts are.  If mortgages expand, rates will rise.  If they don't, rates will not rise, and if the Fed tries to raise them too much, I don't think they will be able to rise by much.

The slope of the yield curve peaked in late 2013 at about 35 basis points per quarter, and has since declined back to about 18 bp per quarter.  Previous rate hike slopes during the Great Moderation have been in the 50 to 75 bp range, so this is very low.  Again, I think this is highly dependent on mortgage expansion.  If the Fed begins raising rates without an expansion in mortgages, the yield curve will flatten.  If they begin raising rates and mortgages are expanding, then we are highly unlikely to see such a low rate of increases, no matter what the FOMC says they plan to do.

Here are four yield curves over time.  At the beginning of QE3, expectations had become very low - not much different than today's.  Then, by fall 2013, long term rates had rebounded significantly and the yield curve slope had steepened.  But, by the beginning of 2015, the slope had declined and long term rates had fallen to a level even below the pre-QE3 level.  Since January, the slope has declined slightly and the date of the first hike has moved back a quarter.  Rates at the long end have remained stable.

One possibility for the extremely low long term rates and slope is that, because of the zero lower bound, expectations are not normally distributed.  There may be a negative skew, so that the mode expected forward rates are higher than the median.  Normally the mean expected rate would be lower than the median in that case, but the zero lower bound truncates the values at the negative end of the curve, which has the opposite effect.  Regarding the slope, there could be a bifurcation of expectations, divided between rates rising at something more like the past rates of 50 bp per quarter or more, and another set of investors who don't expect a rise at all.

Or, another way to look at this is that there is simply a broad disequilibrium throughout the bond market, especially in the long term market, because the collapse in the real estate market has left a hole of $20 trillion worth of US real estate that evaporated, and the combination of a hobbled mortgage market and frictions in the own-for-rental market, that prevent capital expansion in that asset class from expanding enough to capture the large amount of low risk capital in US financial markets.

Mortgage expansion will solve both of these distortions, so I expect mortgage expansion to lead to an increased slope and a higher level for long term rates.  The FOMC announcement will bring some short term volatility, but rates forward enough to have a reduced exposure to the initial rate hike decision should be more exposed to mortgage developments than to transitive FOMC policy statements.

While the government continues to harass mortgage lenders, which is understandable given the consensus view that this is all their fault, mortgage originations appear to be on the rise.  This seems to be showing up in the Fed's H.8 reports, and Flow of Funds looks like mortgage levels are finally ready to rise.  But, this isn't a done deal.  There are even some more regulatory hurdles coming later this year.  And, it seems like there will be a lot of pressure to undermine housing markets if we start to see the 10-15% annual price increases that must be around the corner if mortgage markets begin to function.

In the meantime, I am positioned, tentatively, for this transition, but I consider tomorrow's FOMC announcement to be mostly a source of troublesome volatility until permanent factors become better established.

Housing Tax Policy, A Series: Part 30 - Secular Stagnation, a Housing Bubble, and Loose Monetary Policy can't all be true.

This graph shows the long term decline in real housing expenditures.  Contrary to popular narratives about the housing boom, there is no indication here of overbuilding.  On the contrary, as a proportion of total consumption, households have been bidding up the price of a dwindling housing stock.

This topic is a little bit tricky because of the position of the owner-occupier as both a consumer and a supplier.  It is tempting to think of rising home prices as a product of high demand.  But, if we think about this carefully, the housing "bubble" narrative is actually saying that households were being induced into being housing suppliers.  When a household takes on a new mortgage and buys a new home, they are increasing supply by investing capital into a property which will provide additional real supply over time, as measured in inflation-adjusted rent.  It is not possible to say whether they are increasing demand.  The best window into housing demand that we have comes from the measures of nominal and real rent and imputed rent.

A context with stable nominal expenditures and declining real expenditures suggests a relatively stable demand and a shortage of supply.  If we truly had an oversupply of housing, then real and nominal housing expenditures would have looked more like this:

Maybe the BEA messed up the imputations of owner-occupier rent, and nominal housing expenses were actually going up.  As a double check on that, here is an estimate of housing expenses with no imputations.  This is an estimate of cash expenses for all households, including tenant rent and owner-occupier gross interest expense, direct expenses, subsidies and transfers.  This was actually declining during the boom.

The trend in cash expenses is much flatter if we only include the real portion of mortgage interest.  As I have described before, the inflation premium portion of mortgage interest is really a purchase of home equity by the borrower from the lender, when considered in real terms.  So, much of the decline in cash expenses is the product of the decline in expected inflation, which reduced nominal mortgage payments.  That is a kind of imputation, so I have not included that adjustment here.  But, even if I did, it basically gets the housing expenses trend just back up to a flat line.

Without a rising level of nominal housing expenditures, we can't explain the housing boom as a combination of both rising demand and rising supply.  But, what if the measure of rent inflation is somehow wrong because of the difficulties of estimating rent imputations?  What if housing expenditures looked like my counterfactual graph, above, with level nominal spending, and an oversupply of homes which meant that real housing consumption was actually increasing?  That seems like the outcome we would have to see if there was a "bubble" in home construction and home ownership, but no increase in nominal home expenditures.

First, here is a graph of core inflation over time, separated between Shelter inflation and non-Shelter Core inflation.  (I have used a static weight for Shelter Inflation equal to 40% of core, which I think is a good approximation of the shelter weight over time.)  First, without talking about counterfactuals, just looking at Core minus Shelter Inflation over time, we can see that shelter inflation has been high since 1995.  Excluding the shelter inflation, core minus shelter inflation has moved between 1% and 2% for 20 years.  I believe that this reflects supply constraints, and has been pushing monetary policy too low, to the extent that the little inflation we have seen has been interpreted as a demand phenomenon.

But, what if we apply our housing counterfactual?  What if real housing consumption has been rising, and this has been mismeasured as inflation?  In that case, Shelter inflation should have been running below Core Inflation.  In this graph, I have adjusted Shelter inflation so that it is slightly below Core inflation instead of slightly above it, which is the inflation level that would correspond to our housing oversupply counterfactual.  In the counterfactual, Core Inflation has been running between 0% and 1.5% since 1997.

So, unless someone can unearth data that points to rising nominal housing expenditures, relative to total consumption, it seems to me that there are two possible narratives to choose from:

1) There has been a decades long shortage of housing, combined with relatively tight monetary policy that has contained non-shelter inflation generally below the 2% target rate.

or

2) There was a housing bubble, which created an oversupply of housing in the late 1990s and early 2000s.  The flood of money into the housing market created an increase in the real housing stock, relative to incomes.  This means that RGDP has been running at much higher levels than previously thought and inflation has been lower than previously thought.

If we believe the data, narrative number 1 is the correct narrative.  But, if we are going to believe narrative number 2, then the narrative includes neither secular stagnation nor loose monetary policy before 2006.  There is an awful lot of data that would have to be massaged in order to tell a story that includes an oversupply of housing, secular stagnation, and loose monetary policy.  That's the story I keep seeing in the papers.  The data tells the opposite story.  (In fact, secular stagnation just might be coming from the constraints we have placed on housing supply.)

Housing Tax Policy, A Series: Part 29 - Down Payments: More dogs not barking.

What do you think a graph of down payments in the 2000s would look like?  It would be plummeting, right?  That's how the banks pulled marginal households into the market, right?

Here's a graph from the FHFA of down payments on conventional mortgages on all homes, nationwide.  (This is 100 minus the Loan to Price value in the source data.)  From 2000 to 2005, down payments became substantially larger.

And, notice when they suddenly dropped to much lower levels:  2006 and 2007 - after house prices had leveled off and began to fall.  The period where I have been arguing that the Fed was already creating a liquidity shortage.  Then down payment levels recovered in 2008.  Of course, by 2007, and especially by 2008, mortgage originations were very low.  In fact, conventional loan originations had begun to fall by 2004 and 2005.

But, this is just conventional mortgages, right?  Wasn't the real damage in subprime?  Here is a table from the Demyanyk and Van Hemert paper that I referenced the other day.  This lists several characteristics of subprime loans.  This comes from loan level data from CoreLogic.  Combined Loan to Value ratios on these loans rose from 79.4% in 2001 to 85.9% in 2006.  This includes all loans, not just the first lien mortgage.  Here we see a mirror image of the downpayments in the conventional loans, with a reduction in down payments as a percentage of the home price of 6.5%.  That is something.

But, take a look at the average loan size.  It grew from $126,000 in 2001 to $212,000 in 2006.  That means that in dollar terms, in 2001 the average down payment was $32,690 and in 2006 it was $34,799.  In dollar terms, down payments rose slightly, even among subprime loans.

Further, let's think of a hypothetical household moving from renting to owning.  Using data from the Flow of Funds report and the BEA, for total owner-occupied real estate value and total imputed rent, the national average price to (annual) rent ratio was 15.6 in 2001 and 20.0 in 2006.  (This is more conservative than the rise implied by the Case-Shiller indexes.)  So, an average family in a home with $848/month rent could have purchased that home in 2001 for $158,690 - the average price for homes purchased with subprime loans that year.  But, by 2006, rent on that home would have risen to $979/month and it would have cost $234,902.  Even if they used a subprime mortgage, the average down payment for the average family went from $32,690 to $33,121.

Certainly averages hide some details about the distribution of mortgages.  But, it seems unlikely that broad based changes in down payments could have been a causal factor in creating unsustainably high home prices.

This also confirms another pattern I have been seeing, which is that rent inflation has been high.  That same hypothetical house was fetching more of the median family's income in 2006 than it had in 2001, even though real median household income had risen slightly.  Households were not purchasing larger homes (or, more accurately, homes which would fetch higher rents).  The average home purchased with a subprime loan in 2006 had rent that was a slightly larger portion of median family income, but over that time, subprime loans had grown from 7.6% to 23.5% of total mortgage originations.  Using Debt-to-Income, Mortgage Rates, and Average Loan Size from the Demyanyk and Van Hemert table, I estimate that the average income of subprime loan borrowers increased by about 35% from 2001 to 2006.  The average subprime loan borrower in 2006 was moving into a home with a lower rent/income value than the average subprime loan borrower had in 2001.

Demyanyk and Van Hemert also show that during the 2000s the premium borrowers had to pay via higher interest rates for having a low down payment was increasing as the boom progressed.

There certainly are some mysteries to be uncovered regarding the explosion of subprime lending in the 2000s, but the evidence suggests that, even among subprime borrowers, the trend from 2001 to 2006 was of higher income households making relatively larger down payments on homes with lower imputed rents relative to their incomes.  And that understates the downsizing they were engaging in, because rent inflation meant the same rent was getting less home.  Higher prices were not coming from lower income households making smaller down payments on larger homes.