Thursday, January 29, 2015

Housing Tax Policy, A Series: Part 4 - Real Interest Rates are important

David Beckworth had an interesting post recently, where he pushed back on the Secular Stagnation thesis.  He included this chart:

He attempts to fully pull out both the inflation expectations and the term premium for inflation uncertainty, which has the effect of flattening the 10 year yield over time compared to the yield that only adjusts for inflation expectations.  I'm not sure I'm completely on board, though.  Any time you do a nominal time series that includes the 1970s, that big inflationary hump swamps everything else.  So, everything looks flat compared to the inflation.  But, the real interest rate line here still looks to me like it has a bit of a downward slope.  Even less than a 1% change in a measure that is at such a low level can be significant.  The cyclical noise makes it difficult to see.

He also includes this graph, comparing the 10 year real yield to the output gap.  At this time frame, I agree that there isn't much of a downward slope until 2008.  So, it may be true that there isn't a long term downtrend in real risk free rates.  But, I think there are some interesting patterns here, and I think we can start here for a check on the importance of real risk free interest rates in housing markets.

My attempt at a real risk free 10 year interest rate only goes back to 1969, and it doesn't quite match David's, but it has the same basic shape.  I can get back to 1949 with the output gap.

There appears to be a relationship between the output gap, real rates, and housing returns (these graphs reflect owner-occupant returns only).  This first graph is the comparison of annual numbers.  The output gap and interest rates are very noisy.  So, in the second graph, I compare housing returns to the centered 10 year moving average interest rate and output gap.  This makes the co-movement more clear.

The disconnect between bond yields and home returns after 2007
is due to the very high excess returns going to home owners
currently because access to mortgage credit is so shut down.
It is interesting how much more stable real returns to real estate are than 10 year treasuries (2% has been added to the real treasury yield for comparison in the graphs).  The steep drop in home yields in the 2000s was much less volatile than typical bond movements.  From a volatility point of view, homes should have lower returns compared to long term bonds, based on these numbers.  The factors that lead to this apparent return on investment are part of what I want to think about in this series of posts.

But, here, I think we can see that home prices really do react to long term real interest rates.

Note that in the past 60 years, we did have previous periods where long term rates rose and returns on houses rose with them.  Real home price/rent ratios would have fallen.

I will submit this as yet one more piece of evidence that the 2000s weren't some aberration in history regarding home prices.  This was a normal market.  The new aspect was the tight monetary policy and relatively low inflation that failed to buffer the real capital losses of home owners, which eventually became ruinously tight.  It looks to me like the rate increases in the 1950's coincided with a building boom coming out of the 1940's.  Going into the 1950's it looks like there was a housing surplus because of the swoon in population during the Great Depression and World War II.  As the parents of the baby boomers moved into single family homes, helped by New Deal home ownership initiatives, owner-occupied home building boomed, buffering the drop in Price/Rent.  And again, between 1990 and 1993, according to the Case-Shiller 10 city Index, nominal home prices fell by 8 1/2%.  From 1989 to 1996 there was a 26% drop in Price/Rent.  From March 2006- March 2007, home prices leveled off in a typical way.  A Fed doing its job in 2007 makes this a very different graph.  We have learned all the wrong lessons from this.

I did wonder if this relationship was partly an artifact of BEA imputations.  Capital consumption appears to be generally a function of real estate values.  I think this would be distortionary, because changes in real estate values that are related to interest rate fluctuations would be related to the underlying property value.  The added nominal price would not be a depreciating asset.  I think it would be more accurate to make the capital consumption adjustment proportional to rent instead of price.  In the relationship I outline above, the BEA treatment of capital consumption could cause returns on homes to fluctuate more widely because, as home prices increase, the capital consumption adjustment would also increase, leading to an inflated volatility of measured returns.  So, as a test, I replaced the capital consumption figure used by the BEA with a capital consumption estimate that is a stable 22.5% of rent.

This doesn't change the return profile significantly.  But, it does narrow the range of returns to home ownership a little bit.  In this graph, I have included the returns to equity, the returns to debt + equity, and the returns to debt + equity with the stabilized capital consumption adjustment.  In addition to seeing the small effect of the stable capital consumption this graph also shows how debt captured a much larger portion of housing returns, first from the high nominal interest rates of the 1970s and 1980s, and then because of the mortgage interest deduction after 1996.  (Remember, part of that exchange between debt and equity is a swap of cash income for deferred capital gains, so the sum of the two as the total return on the home is fairly straightforward, but the division of the return between them is complicated.)

The last graph gives more detail to the relationship of major housing tax policies and some housing market measures.  The trends in these measures seem to point to definite effects on the market as a result of these policies.  (Changes in the CRA in 1994 may also be significant, although it appears to have triggered a rise in homeownership rates, but not prices.)  But, we might have expected more generous tax policies to pull down the rate of return on home ownership, and the graphs above don't show an obvious decline in home returns over time, relative to real risk free rates.  But, a 10% increase in home prices from tax incentives would only pull net returns down by around 1/2%.  So, it wouldn't take much of a change in the measured return to homes to create a significant de facto transfer to home owners through tax policy.  I will probably revisit some version of this graph as I think through these factors.

Wednesday, January 28, 2015

Housing Tax Policy, A Series: Part 3 - Total Returns to Capital is Conceptually Important.

There is so much to consider here, I'm kind of getting strangled in the weeds at the BEA.  I hope to have some follow-up posts soon.  In the meantime, I want to float an idea that I haven't seen addressed elsewhere.

In macro-economic discussions, profits seem to be frequently treated as the return on capital.  But, interest is just as important when looking at returns to capital.  I touched on that here.  Even in finance, we tend to start with the risk free rate and then add an equity premium to that, which, also leads to a viewpoint where debt is a step to take to get to profit.

I have gotten into the habit of looking at total unleveraged return as the baseline value.  Then within that total return, there is a swap where equity holders exchange fixed income to debt holders for a premium.  So, we start with a total return, and the equity premium becomes a fixed income discount.  I like this framing better for several reasons.  One reason is that total return appears to be more stable than equity return, and this framing helps to see how the changing risk free rate is mostly a product of debt holders adjusting the discount they are willing to pay to avoid manageable volatility.

Homes are no different than corporate assets, and the same behavior seems to be in place with them.

The red line in this first graph is the Rental Income to Persons after Capital Consumption Adjustment.  (All of these graphs are for owner-occupiers only.)  This red line is similar to the income I have discussed in earlier posts.  This is real income accruing to households that isn't showing up as compensation, because it represents the profit they gain from being home owners.  The red line is the important line for that analysis.  (edit: Although, even there, the total return is important in the long run.)  But, for analyzing the value of the homes themselves, we should look at Rental Income to Persons plus Mortgage Interest.  This is the total return on the asset.  And, it is much more stable over time.

Here it is shown as a portion of GDP.  Here we can see that the dip in Rental Income in the 1980s was the result of high interest mortgages.  But, this still doesn't quite give us the right comparison, because spending on imputed rent and home ownership rates change over time.  So, we need to look at these returns to real estate capital as a portion of real estate market values.

And, as with corporate returns, here also we find that the total return level is much more stable than the "profit" level.  Here we can see that returns to real estate have remained within a range of 2.5% to 4% of market value for nearly a century.  I think this will be an important reference point for my analysis in this series of posts.

A Difference Between Real Estate and Corporate Assets

There is an important difference between real estate and corporate assets regarding this debt-equity trade.  The main factor in this trade on corporate assets is a risk premium.  When the risk premium is high, more of the return is retained by equity as profit.  When the risk premium is low (interest rates are high, equity premiums are low), more of the return goes to debt as interest.

But, with real estate, both the house and the mortgage are relatively low risk assets.  They both tend to have relatively stable income returns (either through interest or through rent).  Both of those income streams are much more stable than the net income stream to corporate equity.  So the trade here isn't a risk swap.  It's an inflation swap.

But, this gets really complicated, and I haven't completely figured out how this affects reported incomes and production.  The BEA doesn't account for capital gains on existing assets because they are not the result of production.  But, when homeowners engage in this trade, especially in a high inflation context like the 1970s & early 1980s, they are exchanging a cash expense for a capital gain.  That is how net returns on housing (Rental Income) could actually dip into negative territory.  Homeowners were earning returns on their homes that were about as high as they have always been.  But, when inflation was 8%, part of their interest expense was an exchange of an 8% cash interest expense which they paid to the bank, in exchange for ownership of the home, which would provide them with an 8% unrealized capital gain.

With the increasing level of spending on housing and the high inflation of the 1970s, this has been a significant effect on the way incomes are reported.  Maybe it just amounts to a fairly arbitrary exchange between income reported as financial corporate profits vs. capital income to persons.  But, I wonder if stated national incomes are somehow being distorted.  I wonder if there could be some distortion related to this that is coming through the capital consumption adjustment.  I haven't fully wrapped my head around this problem, because it's a conceptually complex exchange through time, mixing changing asset values and income.

Monday, January 26, 2015

Housing Tax Policy, A Series: Part 2 - Outline of the model

<Update:  These estimates are bound to change some more as I continue to review data and concepts.  I am purposefully kind of thinking aloud.  So please check on follow-up posts in the series if you are interested in these estimates.>

Here is the very simple model I am using in excel.

Colored cells are user input.  The model is simple, but I think the implications will be interesting and useful.  I'll walk through it briefly:

Tax Rates: The landlord is my baseline, so these represent tax exemptions for owner-occupiers.  Tax policy, especially in real estate, is complex, so I am open to boot-on-the-ground input on how steep the de facto taxes are for the marginal landlord.  These seem like a good starting point, partly based on some commenter input from the last post.

Rent Adjustments:  Property Tax, Insurance, and Maintenance apply to net rental income for all owners.  Property Tax would be more accurately portrayed as a portion of home price, but putting it here helps me to avoid circularity in the model.  This means that I need to be careful with this measure if I test values much higher than current levels, because I lose some of the feedback loop between taxes and home values by putting it into the formula here.  The Owner Benefit premium is the added value of having control over one's homestead.  The data might steer this one way or another, but I am starting in the ballpark of 30%.  This is in the range of a control premium in a corporate ownership context.  That is a slightly different context, obviously.  Actually, I suspect this premium is very household-dependent.  It is clearly nearly zero for, say, a college student, but could be much higher than 30% for a very high income middle class family.  This is a factor I might try to think through as I move through the issue.

Equity % is the equity percentage of the owner.

I am simply estimating the mortgage benefit based on the aggregate effect.  This would not be very accurate for individual households, since the usefulness of this tax deduction would vary with the size of the mortgage and the tax characteristics of the household.  But, I think it is a decent estimate for aggregate effects.  The Mortgage Interest Benefit comes from the leverage level and the Mortgage Benefit Factor, which will be 0 before 1986 and .06 after 1986.

I am pegging rent at $1,000 per month.  Rent is the baseline figure here, so Price to Rent will be the home price measure that will be most relevant.  Where I use home prices as a comparative measure, they will tend to be adjusted for rent inflation.

Owner Net Rent Value is the net adjusted rent income, either imputed or paid, of the homeowner, after accounting for all the adjustments above.

Capital Gains Usage reflects the ability of owner-occupiers to use the capital gains tax exemption.  I have it set at 25% before 1996 and 75% after 1996.  But, again, boots-on-the-ground input here is welcome.

Owner risk premium is the added return on investment that owner-occupiers would require because their ownership cannot be diversified and would tend to have some correlation with personal risk attributes, such as the local employment market.

The Capital Gains Exemption Premium is a product of the capital gains exemption, the usage of that exemption, and future inflation.

The intrinsic home value (Home Price) is based on a discount rate that includes the real rate, the owner risk premium, and the capital gains exemption premium (which is negative).  I can ignore inflation here because of my assumptions about home price appreciation and inflation expectations.  I have based intrinsic value on the DCF value of 100 years of rent.

The mortgage rate is the sum of the real rate and the inflation rate.

Here is the update of my basic estimates of the subsidies, before digging into the conceptual details of how they might work through the housing market, based on comments from part 1:

The Cap Gains levels are affected by other factors, and I might tend to find
that they are even lower in many contexts where the other specs change.

I am interested in how these factors might influence home owning decisions and home prices.  I will begin to review these issues in the next post.

Thursday, January 22, 2015

Housing Tax Policy, A Series: Part 1 - I was wrong. The mortgage deduction is just the tip of the iceberg.

I've been yammering on about housing policy lately.  I referenced this study that claimed mortgage tax deductions amount to nearly $100 billion annually.  Additionally, the capital gains exemption for owner-occupiers is worth probably even more.  But, I have been missing the full implication of the largest tax benefit - the fact that owner-occupiers don't have to claim imputed rental income as taxable income.  I can't believe that I didn't think about what that means.  I'm going to model this out, because I think this is going to be significant.  What I am finding is that this is very complicated, and the magnitudes of many of the implications dwarf the mortgage interest deduction.

To begin, we basically have three tax advantages to home ownership.

1) The mortgage tax deduction.  This became more important after the Tax Reform Act of 1986.  Before that, all personal interest expenses were deductible above a minimum level.  After that, mortgages carried a tax advantage to other forms of debt.  The effect on the housing market is limited, however, because the marginal tax savings only kick in for a household once itemized tax deductions exceed the standard deduction.  Thus, most of the potential tax savings is unused.  The study mentioned above references work that puts the effect on home prices of 3%-6%.  In my models here, I am using 3.5%.  That looks about right to me. This amounts to an annual transfer from renters to owner-occupiers of about 0.5% of GDP.  I won't get into this here, but if we broke out households more finely, we would find that within the homeowner group, most of that transfer would be going to upper-middle class households who have mortgages large enough to capture the benefit.  So, an unfortunate redistribution of income, but not a significant distortion in the housing market.

2) The capital gains tax exemption.  This was significantly strengthened in the Taxpayer Relief Act of 1997, which made the availability of exemptions on real estate capital gains much more universal.  I should have concentrated on this more in my previous posts.  I am surprised this isn't more of a focus in general.  Here is a comparison of home prices and homeownership rates in the US.  Note that while homeownership rates seem to have reacted to the 1994 passage of the Community Reinvestment Act, price appreciation kicks into gear in 1997 along with the capital gains tax exemption.

I have been defending home prices in the 2000s as being a reasonable reflection of the effect of low real interest rates.  I might have to walk a little of that back.  (I don't know.  This is going to be really complicated before I'm done.)  But, in general, this does still point to my general intuition that while home prices do react to changing demand from more universal access, most of the change in home prices is a reflection of the changing intrinsic value.

In terms of its effect on the housing market, if we assume a 20% capital gains tax rate and 75% utilization of the exemption, the value of a home increases by about 10% at today's interest rates.  Since owner-occupier rent represents about 14% of GDP, this amounts to an annual transfer of about 1.4% of GDP from renters to owners.

Keep in mind, in aggregate, the capital gains taxes on real estate are largely a tax on inflation effects.  My preference is to avoid this distortion by eliminating all capital gains taxes.  The effect here is to create a dislocation between owners and renters.  And the effect is very sensitive to inflation.  So, even though this effect may have been very strong, the low inflation environment we have been in has greatly diminished this distortion.  The 10% effect on home values is based on 2% inflation expectations.  In a 4% inflation environment, the capital gains exemption would create a 20% price distortion.

3) The tax exemption of  the rent income itself.  As far as I know, this has always been the policy.  This is surely a reason why homeownership rates were nearly 50% even before the mortgage deduction, capital gains exemption, and low income homeownership programs were in place.  In effect, this is the opposite issue that two-earner families face.  If a second parent joins the labor force, there is a tax penalty that arises from the fact that much of the household value that had been provided by a full time homemaker is now filtered through additional household cash income, which is taxable, to payments for child and home care services, which are frequently not tax deductible.  Homeownership is sort of the capital version of labor's household production.  Owning a house allows a household to consume its own production in a way that doesn't trigger a public transaction.

This is a case where it is important to separate forms of ownership (debt and equity) from the returns to the asset itself.  We need to start here with a comparison between a landlord who owns an unleveraged house and an owner-occupier that owns an unleveraged house.  When we keep focused in this way, this is a very simple concept.  Let's assume an income tax rate of 25%.  If a landlord has, say, a $300,000 house that has a market rental rate of $1,000, then his post-tax monthly return is $750 on that property.  Since an owner-occupier doesn't pay that tax, she would be willing to pay 25% more for the same property.  At those different prices, both home owners - the landlord and the owner-occupier - would be earning the same after-tax return on investment.

Now, I know, there are costs to owning the home that need to be deducted from the rent, etc.  I will make these adjustments as I work through this issue.  But, I think the value of this factor is pretty clear.  And, unlike the mortgage and capital gains tax factors, utilization does not mitigate this factor.  For most households, if imputed rental income was taxable, their taxes would rise with the first marginal extra dollar of income.  This factor is nearly 100% utilized.

If we assume that all homes are fully owned with no leverage, a model of home values that accounts for costs of ownership, with a 25% income tax rate, ascribes a 20% increase in the home value due to this factor.  That amounts to an annual transfer from renters to owners of about 2.9% of GDP.

The relationship between the income exemption and the mortgage tax deduction

This is where the mortgage tax deduction comes into play, and before I have thought through it this way, but I hadn't fully appreciated the value of the income tax exemption separate from the mortgage deduction.  A cash homeowner receives the full advantage of the income exemption.  But, if we assume that a homeowner is fully leveraged 100% on a home, with a perpetual interest only loan in a zero inflation environment, we can imagine that they might have $1,000 in monthly income that is untaxed, but also $1,000 in monthly interest expense that is not deductible.  They would have no net benefit from homeownership.  (Differences between rent expense and interest expense for actual homeowners are effectively a product of inflation and duration exposures they are taking due to the different characteristics of the mortgage and the home.)

So, in an economy where household real estate is leveraged to roughly 50% in the aggregate, this means that owner-occupiers claim roughly 50% of the income exemption benefit - which would account for a 10% increase in intrinsic home values and a 1.4% of GDP annual transfer from renters to owner-occupiers.  The mortgage tax deduction is, effectively, a way for leveraged home owners to also capture this benefit.  But, since the utilization of the mortgage deduction is so low, leveraged home owners only capture about 1/3 of the benefit.  (As stated above, the mortgage deduction might increase home values by 3.5%, compared to my estimate of an additional 10% increase that homeowners would see if they had 100% equity instead of 50%.)

Here is a summary of the effects.  (Keep in mind that these are broad estimates, which interact with one another and also change slightly in scale as interest rates and other assumptions change):
The capital gain exemption is proportional to expected inflation.
(assumed to be 2% here)

<**Edit: Vivian Darkbloom has talked me down a bit in the comments.  After considering her input, I would revise these numbers to: 3% + 5% + 7% = 15% of home values and 0.4% + 0.7% + 1.0% = 2.2% of GDP.  The Cap Gains numbers may still be high, but for a number of reasons having to do with the relationships of the factors in the model, the effect on the modeled price is higher than the effect we might estimate simply by accounting for this year's expected real estate gains.  This factor may turn out to be lower in many contexts.>

<Edit #2:  These estimates are bound to change some more as I continue to review data and concepts.  I am purposefully kind of thinking aloud.  So please check on follow-up posts in the series if you are interested in these estimates.>

Over the next few posts, I expect to show that the actual increase in home values may be much less than 23.5%.  But I expect to demonstrate that this transfer to owner-occupiers exists regardless of the total effect on nominal home prices.  The conclusions I have reached so far have been pretty shocking to me, and I may continue to be shocked as I work through the models to complete the remaining posts, so my conclusions may differ from what I currently expect.  (I am proceeding as I work this out, so you might see my errors as I proceed.  And, this will get complex enough that I might end up reaching the limits of my analytical tools.)  At the point where I am in my analysis now, I believe that if these factors are not manifest in higher nominal market prices for homes, then the potential increase in home values coming out of these policies is a sort of risk-adjusted deadweight loss on our economy.

Tuesday, January 20, 2015

From the comments, on housing. A stream of consciousness primer on my view

I have had an extended discussion with "baconbacon" in the comments of this post.  He has used this URL to post extended comments.  I decided that responses to his latest post might be worth putting into a new post.  And this is it.

On Conceptualizing Homes as a Financial Instrument


I will summarize Kevin's position here- ... with very low nominal rates of interest it makes sense to borrow large amounts (to lever up) in exchange for the cash flow (in this case the implied rent)... This is a major point of disagreement for me.  By treating home ownership as an all-cash investment he eliminates the roles of banks, and banks and homeowners have very different risk profiles, and it is this risk profile where his analysis falls short.
There is a distinction to make here, which I think is very important when thinking about homes in the recent unusual context.  My point about thinking in terms of cash is that there are two separate figures with any asset - the intrinsic value and the market price.  We expect most markets to be relatively efficient so that most of the time, we might reasonably assume that these are the same price.

When valuing the intrinsic value of a stock, an analyst restricts her value to the characteristics of the firm.  As a first order effect, we don't change our estimated value of a stock depending on whether we are investing with cash or on margin.  Homes are no different.  Their market values are independent of the financing source of any individual buyer.

Now, there is some interdependence between the two, but it is subtle, and it is important to keep the conceptual effects of demand side credit issues separate.  In equities, the financial condition of an individual buyer would affect her personal risk aversion, and would change her optimal allocation of capital.  And, cyclical and monetary factors might influence aggregate risk premiums and capital flows, which would change the relative values of equities in different markets.

So, when I say we should conceive of homes as all-cash investments, I am not saying that banks have no effect on home valuations.  But, as a first step, homes have an intrinsic value based on future cash flows, as a stand-alone, unlevered asset.

This is a difficult point for me to make, because it is so common for home price fluctuations to be conceived of through the mortgage market's effect on demand.  It seems that when interest rates decline, everyone describes its effect on home prices as coming through demand, because households will be able to afford more house with the same mortgage payment.  I think this adds needless confusion.  It is also not a realistic portrayal of the market.  Average LTV generally hovers around 50%, and many home owners have substantial equity.  All-cash buyers are part of a significant portion of transactions.

When real interest rates decline, home values go up, mortgage or no mortgage, because far-future cash flows have a higher present value.  In fact, homes as a long-term financial security have very long durations - longer than the duration of a 30 year mortgage - so when real interest rates decline, the breakeven mortgage/rent ratio actually goes up.  (So, a period like the 2000's with very low long term real interest rates can look like a bubble if you view the mortgage/rent ratio as a conceptual constant.)  On the other hand, when the inflation premium (expected inflation) declines, it doesn't affect the intrinsic value of a home, but it does do exactly what the conventional narrative says.  It lowers the monthly payment, and lowers the barrier to mortgage qualification, decreasing obstacles to demand for home ownership.  This lowers excess returns available to owner-occupiers, so it can cause home prices to rise.

These factors, always moving up and down stochastically, look enough like the mortgage demand story that our minds find easy to conceptualize, so that's the narrative.  But, you have to detangle these effects to explain why Price-to-Rent ratios were rising in the late 1970's when mortgage rates were well over 10% and why they were rising sharply in 2012-2013 with no new mortgage credit and when many buyers were all-cash.  Real rates have a tremendous effect on intrinsic home values, especially when they are very low.

Now, banks come into play.  But, if we look at it this way, we might see intrinsic value as the starting point, and, due to the constraint we impose on real estate - that it is usually purchased as a whole unit, and not as partial shares - issues of financing usually reflect a discount from intrinsic value.  The lack of access to owner-occupancy, due to financing obstacles, limits demand and usually pushes home prices below intrinsic value.  Or, stated differently, there are usually excess returns to home ownership.  (Note, much of this comes about from tax advantages we give owner-occupiers.)

This is where I usually get the response that home ownership is not financially beneficial compared to renting.  While it is impossible to argue this definitively, because of the complexity and long life of the asset, I will make two points here.  First, there are many non-financial benefits to owner-occupancy, including the elimination of several principal-agent problems, cultural status markers, personal control, etc.  These are part of the equation, though they are difficult to quantify.  But, also, because there is a tendency to conceive of home ownership as a full mortgage vs. rent decision, we tend to compare mortgage payments to rent payments in our mental estimates of cost.  30 year fixed mortgage payments can actually serve as a reasonable proxy for the relative value of ownership.  But, the problem is that mortgage payments are inflation protected but rental payments are not.  So, at the outset of a purchase decision, the mortgage payment will always be higher than the breakeven rent payment.  At high inflation rates, it will be much higher.  The value to home ownership comes when 30 years later, you make your last $2,000 mortgage payment on a house that by then can fetch a $4,000 rent.  It is very hard to fully intuitively appreciate this huge benefit to owning.

So, for the first time in the modern era, in the 2000's, real interest rates were really low (high intrinsic values) and inflation expectations were really low (unencumbered demand).  Market prices were high, and, for the first time, nearing intrinsic values.

"Baconbacon" writes:
Now you can tell a different story with the data shown- but it basically requires the claim that either a substantial portion of the population (3-5% percent of households) went from being a poor mortgage risk to a good one
But 3-5% of households did become better mortgage risks.  Inflation premiums dropped by several percentage points between 1980 and 2003.  A reasonable monthly mortgage payment in 2003, even at very high nominal home prices, was much lower than in 1980 or 1990.  That makes a real difference.  Lower real rates meant that home prices should be higher, and lower inflation premiums meant that more households could handle the upfront cash flow demands of home ownership.

"Baconbacon" later commented that:
I think you are underselling the costs here. You need a liquidity premium for owning vs renting, you need a discussion of upfront costs (closing costs, inspection, realtor costs) and insurance to start to get a more accurate picture of owning vs renting.
But, again, I think it is important to see the distinctions we need to make in estimating market values.  Liquidity premiums, closing costs, etc. are important for deciding individually what real estate purchasing decisions we should make, just like they are important in deciding whether we should invest in small cap stocks, annuities, or bonds.  But, when we are looking at the level of the stock market, these are all simply inputs into aggregate risk premiums.  Those risk premiums are basically an aggregation of the countless personal judgment calls that make a market.

We can get an estimate of all of these aggregate personal judgments as we look at home prices over time, and use this empirical data to analyze the current market and estimate how the aggregate risk premiums that arise from these uncountable individual judgments are changing.  If we look at housing this way, we don't need anything unusual to describe the boom market of the 2000's.  Lower real rates and inflation expectations, compared to the 1980s and 1990s, can explain the price appreciation.  The current market can only be explained by a very high level of excess profit above levels we would estimate from long-standing risk premiums, which clearly is coming from a broken down mortgage credit market.

This is where viewing home values as a rent vs. own decision, with a "breakeven" price is problematic.  The apparent "breakeven" price does not easily account for the present values of those far-future cash flows.  It is much better to value the homes as unlevered assets.  There is no breakeven level.  There is a current nominal asset price, and there is some series of expected future cash flows from rent (or implied rent).  The price is a product of changing discount rates on those future cash flows.  Those discount rates are always positive.  And, when they get very low, as they were in the 2000's, the nominal price of a very long duration financial asset skyrockets.

Many observers looking at homes through the "breakeven" point of view claim that they are rarely a quantitatively profitable investment.  And, they also tend to move easily into claiming that buyers in the 2000's were highly irrational.  But, these observers are simply refusing to develop any coherent model of home values, and are projecting this lack of a model onto the market.  In some circumstances, like the 2000s, the poorly specified simple heuristics that worked most of the time, in place of a coherent model, become less accurate estimates of actual models, and so they blame markets for being wrong instead of blaming their pseudo-models.

You may argue that I can't expect markets to be reasonable if most of the market participants are using pseudo-models.  But, I can.  That's what markets do.  Have you read Yahoo message boards?  Markets usually tend toward efficiency in spite of our individual errors.  (Housing could be strong-form IMH, in which a market has high potential profits because of regulatory issues.  So, my analysis could be correct.  I could take long housing positions when I see mortgages start to increase.  And, I could end up with losses because Americans are so universally wrong that the Fed is pressured into kneecapping the economy again in their Quixotic battle against "bubbles".)

This skeptical view of home values would be similar to saying that we can't model prices from US large cap stocks, because small caps have earned higher returns that seem to be persistent.  Investors in large caps are not breaking even, compared to small cap investors, much as investors in homes supposedly aren't breaking even compared to renters.  In stocks, we recognize that there are factors which create different levels of required returns among asset classes.  Homes are no different.  The long term pattern of discount rates of future implied rents is the appropriate breakeven rate, just like the equity risk premium is.  It is for each investor to decide how they should allocate, but the risk premiums are what they are.  They represent the unseen.

On the Banking Bubble

Next, baconbacon gives a great description of issues in banking.  I am generally in agreement with this, and I found the description interesting.  I do believe that there was a AAA-rated security bubble, I just don't think this had more than a minor effect on home values.  Housing is an efficient enough market that I would expect mitigating factors to counter some of the excess demand that might have come out of the AAA bubble.  Of course, I can't prove this.  But, I can look at the housing market and find relatively reasonable specifications that justify the price levels.

Banks operate in a market with several important distortions.  There is a static regulator regime that gives value to AAA-rated securities, so there is a demand among banks for those securities which can fluctuate dangerously, without a robust set of mitigating responses.  Where a market is like a rainforest, a regulated institution with unmoving values regulated by capital controls, Federal deposit insurance, etc. is like an agricultural monoculture.  So, there are AAA-rated securities, which the banks were buying, based on some of the mortgages, which do appear to be clearly mis-priced in hindsight.  And it seems that some mortgages were issued carelessly because the AAA-rated securities they would fuel had excess value in that monoculture.  (Although, new research even casts doubt on this. <HT: TC>)

This problem extended beyond the commercial banks.  I suspect this is partly because, even though the shadow banks didn't have FDIC, they did have overnight repos.  Both of these legal factors create a large liability on balance sheets which is immune from pricing signals.  I think overnight repos are an accounting fiction that should be dropped along with FDIC.

I doubt that I would be able to convince anyone who is currently skeptical that the banking bubble didn't have much influence on home prices.  I haven't really tried very hard.  I could be wrong.  But, I'm not in politics.  I'm in finance.  So, I'll place my bets.  And, I could very well see tremendous gains even if I'm completely wrong about this.  That's probably why finance guys are so infuriating.

I hope this isn't getting repetitious for IW readers.  Thanks to baconbacon for posting thorough, thoughtful replies.

Monday, January 19, 2015

Another thought on the distribution of forward interest rates

Recently, I was thinking about the distribution of expectations about forward interest rates, and how that would bias the market rates in the interest rate futures market.  I had been modeling the Eurodollars market as if expectations are normally distributed.  But, recently I considered the possibility that there is a negative skew.  The skew would cause the mode expectation to be above the median and the mean expectation (which we would expect to estimate the market rate) to be below the median.  But, the zero lower bound (ZLB) truncates the bottom portion of the expectations distribution.  This truncation would have the effect of pushing the mean expectation back above the median expectation.

I think it is plausible that these effects are at work in forward rate markets.  But, within my model, where I estimate the date of the first hike and the slope of the yield curve, I had only thought of the effect this would have on the yield curve slope and on rates at the very long end of the curve.

Previously, conceiving of the expectations of the rate hike in terms of time as having a normal distribution, we could estimate the mean, median, and mode date, which would all be the same date.  But, a skewed distribution means that the mode date is earlier than the median date and the mean date is later.  And, on the date of the mode expected first rate hike, the distortion of the ZLB would eliminate almost all of the information coming from expectations below the mode.  On the contract date that is the mode expectation, we would just see the distribution of rate expectations above the ZLB, so that it wouldn't look like a skewed distribution.  It would look a lot like a normal distribution with a very low variance.  So, the yield curve would turn up at the mode date.  The mode date would look just like the mean date of normally distributed expectations.

So, if this skewed distribution is accurate, then the mean expected date of the rate rise could be later than the date estimated by my model, and for that matter, than the date that would look obvious just by eyeballing the yield curve.  June could be the mode expected date, and that would cause the yield curve to start moving up in June.  But, as we move past June in forward rate contracts, the mode rate would move up off of the ZLB, and the market rate would move down, away from the mode, because the ZLB would have less of an effect on the mean.  The yield curve would begin rising earlier (June, the mode date, instead of the mean date, which is probably after September) and would rise at a less steep rate than it would if the mean expectations were the same but if there were no skew in expectations.

In fact, there is quite a negative skew in the Fed's dealer survey.  In the December dealer survey, the mode was June, the median was September, and the mean is likely even later.

Rates are now lower than they were when QE3 began.  Recently, both the expected date and the slope of the curve have fallen.  I have been attributing this to uncertainty.  But, if part of the lower measured slope of the yield curve could be from increasing skew, then the mean expected rate hike is actually later than it appears.

Possibly a Eurodollar position that was long (gains from falling rates) at the short end and short (gains from climbing rates) on forward dates could serve two purposes:

1) in the mode scenario, the near term position would not have significant gains or losses, and the forward position would have gains.

2) in the mean scenario, contingent on rates rising, the position might gain at both contracts.

3) in the mean scenario contingent on rates remaining at the ZLB persistently, then the two positions would act as a hedge for one another.  The forward position would lose more than the near term position would gain.  But, if most of the change in market expectations was from a delay in the first rate hike, or if this led to an expectation of a more dovish Fed, then the yield curve would shift to the right instead of shifting down, and the losses for the forward contracts would not necessarily be higher than the gains for the near term contracts.

In effect, this is a kind of arbitrage on the de-skewing of the expectations curve as information reduces uncertainty over time.  I haven't totally thought this through, but maybe my imagined gains here come from the inevitable shift to either a normal distribution if rates rise or a positive skew if rates remain at the ZLB.  A third moment arbitrage.  Is that something traders do?  Can you arbitrage the third moment?

So, instead of a June rate hike followed by a 20 basis point per quarter climb to 1.3% by September 2016, the true mean expectation might be a rate hike in December 2015, with the market divided between thinking rates will climb closer to 50bp per quarter, as is typical, to 2.25% (and rising) by December 2016, and expecting that we will still be sitting at the ZLB by then.

Possibly there are two types of positions here:

1) The Fed will be committed to raising rates:
Rates begin to rise in June (with very little variance in expectations), and rise to 3-4%, typical of past experience.  This creates the kink in the curve in June, and, before the influence of the other position, would have rates at 3% or more by the end of 2016.  (Rates in the graph are equal to 100 minus the y-axis number.)
2) The Fed will be willing to move rate hikes back:
There is a huge amount of variance in expectations, with many participants expecting rates to remain at ZLB for a long time.  This has the effect of pulling the market rate down from the level implied by the first position, across the yield curve.  But, since this is a proportional effect pulling the yield down, it doesn't pull the apparent date of the first hike to the right (forward in time).

I think this, more or less, describes my skewed distribution model, with maybe a sharper hump in the expectations curve than I have drawn.  Note that the change in slope after 2016 is subtle.  The ZLB effect on the mean rate expectation would slowly decrease over time as the ZLB truncates a smaller portion of expected outcomes.  In addition, there might start to be some positive skew in expectations from investors with long term inflation fears as the Fed unwinds their balance sheet.
I was hoping for confirmation of this week's MBA report of increased mortgage activity.  Friday's Federal Reserve H.8 report didn't have any sign of increasing closed end residential mortgage credit.  We have seen 2 weeks of increased home equity credit, though.  If next week shows an increase, that would be the first time since the crisis that it increased 3 weeks in a row.

Friday, January 16, 2015

December 2014 Inflation and an Adjusted Taylor Rule

YOY "Core minus Shelter" inflation is now down to 0.7%, and all of that came before June.  There has been no "Core minus Shelter" inflation since then.

A version of the Taylor Rule that I track (8.5% - 1.4*(Full Time Unemployment Rate - Core Inflation)) is still nearing 3%, because of strong employment trends.

This version of the Taylor Rule, as well as some others, looks like it tended to be too high in the 2000s, by as much as 2%, and now looks like it might be giving a reading that is still too high.  I have considered the possibility that real estate credit is the main governor on monetary growth right now, so that it may be that raising short term interest rates to 1% or 2% wouldn't have much effect on industrial credit.  But, wherever that number is, this version of the Taylor Rule still looks like it is too high by 1% to 3%.

I had been attributing this to broad secular trends, like the large pool of savings from developing markets and from aging developed world savers.  But, I wonder if part of this is due to the supply problem in housing.  The Taylor Rule treats inflation as a monetary phenomenon, so it might break down as a useful indicator if inflation is coming from a supply shock.

And, interestingly, adjusting this version of the Taylor Rule so that it uses "Core minus Shelter" inflation makes it conform much more closely to the interest rates of the last decade.  When the Fed pushed rates too far up in 2006, turning a plateau in homebuilding into a free-fall, the "Core minus Shelter" Taylor Rule avoids the error of treating the ensuing rent inflation as a demand issue.  We can see that here, as the adjusted Taylor Rule recommends a drop in short term interest rates by late 2006.

By January 2008, the adjusted Taylor Rule has reverted back to the regular Taylor Rule, but by then, rates were in free-fall, and the 1 year backward perspective of the rule would have been distorting its usefulness, as both indicators were negative by the beginning of 2009.

Rent inflation subsided briefly in 2004 and 2005, so that the Fed Funds rate trailed both of these indicators coming out of the previous recession.  Could it be that the core time period of the housing boom was the only time where homebuilding was actually able to meet demand?

The adjusted Taylor Rule is up to 1.5%.  Can I take this as evidence that my indifference to 2015 rate hikes is the correct approach?

If mortgage credit is normalized and equity recovery continues among existing homes, I would expect home building to rise steeply.  This will create new strength in employment and the new building will lower rent inflation.  This will cause these two Taylor Rules to converge, at a higher level.  Possibly this is an initial guide to the two scenarios for 2015 and these two Taylor Rules give an idea of where interest rates will peak, depending on what happens to the housing market.

If the Taylor Rule is useful at all, this might lend hope to the idea that, even if QE3 ended too soon and will lead to significant lingering disinflation, the economy will continue to rebound.  And, if some of the recent inflation softness is due to a positive supply shock in commodities, then both Taylor Rules would move higher.

Thursday, January 15, 2015

Regulation will be more important than the Fed Funds rate in 2015.

Bond yields have been taking a dive - dropping even more yesterday.  Most of the drop has been from inflation expectations, although some of the drop over the past week has come from expectations of a later rate hike date.  Yesterday, very long term forward rates were stable, but both the date of the first hike and the slope of the curve coming off the hike continued to fall steeply.

The other day, I was thinking about the shape of the distribution of rate expectations, and how different shapes would have different implications for the yield curve.  The short version - a bifurcated expectation would pull the entire yield curve down below the median expected forward rates.  A normal distribution in which the lower tail hit the zero lower bound would push the yield curve above the median expected forward rates.  The slope of the yield curve in the near term suggests bifurcation, but the long end of the curve suggests normal distribution.

Thinking about it some more, and watching recent movements, I think the most realistic interpretation is that expectations have a negative skew, with the negative end of rate expectations hitting the zero lower bound (ZLB) across the entire curve.  So, above ZLB, the mode expectation is above the market rate, the median expectation is below the market rate, and the mean expectation is (approximately) the market rate.  Normally, a negative skew would push the mean below the median, but the ZLB truncates the distribution.  So, both of my explanations were part of the answer.
Here is a graph of this new interpretation.  I suspect that the mode expectations remain fairly stable, and what we see in yield curve movements is increasing negative skew related to increasing uncertainty, which does basically increase the hump at ZLB and decrease the hump at the mode.  So, for the full distribution, higher uncertainty pulls the market price down from the mode.  If not for the ZLB truncation, the mean would be moving down even further, so that we would also expect the market rate to fall below the median expectation.  But, since the increasing uncertainty makes the ZLB more binding, that effect pushes the market price above the median expectation.  Both effects from my earlier post are at work at the same time.

(Caution.  I know I'm spit-balling here.  I'm not trying to get published in an academic model or to change anyone's mind who isn't interested in my analysis.  I'm just trying to get a grasp on the world in real time.  I'm very interested in comments regarding mistakes I might be making here, or basic improvements I could make to the thought experiment.  But, I know that there is a lot of speculation here, and a lot of arguable priors are informing this work.  If you are more than 2 steps away from me here in agreement, or if my priors bother you, then there may not be much for us to discuss.)

Rate Movements in 2015-2016

If rates do rise, I think we can expect a 50-75 basis point move, per quarter after they begin to rise.  This is an interesting situation.  I suspect that in the industrial economy, natural interest rates are already rising above ZLB.  They aren't binding.  What is acting as a governor on market interest rates and NGDP growth is the limit on mortgage credit.  This is not related to interest rates.  Interest rates and the level of monthly mortgage payments for new home buyers are not the constraint on housing markets right now.  This leaves trillions of dollars of savings looking for an outlet, and the industrial economy simply can't utilize all the capital.

So, interest rate movements simply may not be that important.  Rates could rise to 2% by the end of 2015, and industrial credit will probably just keep churning along.  Expansion of industrial credit looks like it is running at the high end of typical potential growth.  This will probably continue.  Interest rates will have little effect until they reach some higher level where they become the constraint.

In the meantime, regulatory issues in the mortgage market and household real estate leverage levels will determine real growth, real rates, and inflation.  If this can loosen, all of those measures will move up.  If it doesn't loosen, then I suspect that the main issue will be whether the Fed stops hiking rates before rates become the binding constraint, or if they hike them enough to further hobble the economy.

In other words, we have 2 important factors to watch, in series.

1) Does the mortgage credit market begin to expand again.  If it does, it's Party in the USA.

2) If not, then the date of the first rate hike is not as important as the level of the last rate hike.  How will we know if it is too high?  Industrial credit will start to decelerate.  But, that can be a lagging indicator.  Likely, the earlier indicator will be a flattening yield curve.

If mortgages can begin expanding at any point before that happens, that will be very bullish.  It will mean higher interest rates, increased employment, higher real wages, and healthy equity markets.

From a speculative point of view, as rates begin to rise, uncertainty about rate levels will begin to decline.  This will reduce the skew of forward expectations, and the yield curve will steepen.  The trick is always in the timing and the details in these types of situations.  There will be a window where a position that gains from rising near-term rates is lucrative.  But, considering the policy inertia of the Fed and their general stance, I would expect them to overshoot, which would call for reversing the position and then speculating on near-term yield declines.

Really, the same pattern will probably play out even if mortgage markets do recover in the meantime, but we will likely see much more growth and higher interest rates in that scenario before the reversal.

Wednesday, January 14, 2015

The complicated role of homeowners in income and production growth

I have been thinking through the relationships between housing, inflation, wages, and economic growth.  My basic narrative is that severe restrictions in mortgage credit from regulatory and market constraints on banks, coming out of the crisis, have crushed access to home ownership.  This has pushed home prices well below their intrinsic value.  There are many implications from this, including low single family home prices, low levels of single family home building, heavy relative all-cash and investor buying, strong growth in multi-unit housing, high rent inflation (Shelter CPI has been around 3%. Trulia is reporting 6% YOY rent inflation.), shrinking home ownership rates, high levels of net capital income to real estate owners, limited monetary expansion through bank credit (crimping both real production of new real estate assets and nominal inflation of existing assets), regressive transfer of income from renters to owners, and downward pressure on real wages (since nominal wage increases are devalued by rent inflation that is a result of the negative supply shock in housing).
Source: Calculated Risk

This last point is one that I would like to think about here, and as I begin writing this, I am not sure if I grasp all of the implications.

Through November
As a starting point, I don't agree with concerns I am seeing about slow wage growth.  Average hourly wages for production & non-supervisory workers have been growing at over 2%, but in December dropped to 1.6%, YOY.  Since this is an outlier compared to recent trends, it is possible that this will be revised up.  And, inflation is also falling precipitously, so real wages in December may not have dropped that much.  In any case, as my graph here suggests, real wage growth has been unusually high in this recession and recovery (which has been part of the problem).  But, my more specific argument here is that real wages are growing faster than even this data implies, because while YOY core CPI inflation is 1.7% (as of November), core minus shelter inflation is only 0.9%.  Shelter inflation is coming from negative supply forces, not positive demand forces, so real wages are increasing at a very healthy pace, once we account for the housing supply issue.  But, this is complicated, and I want to think through this out loud to consider the implications.

Dwarfing the Cantillon Effect

There is some debate over whether there is any significant transfer of value to financial interests who are the first to receive new dollars when the Fed expands the money supply.  Despite it's questionable significance and likely small size, this idea holds some status in the populist economics mind, (usually only noticed during an expanding money supply, but forgotten during monetary contractions when the effect would reverse) along with the notion that monetary expansion is a payoff to Wall Street, pushing equities to unsustainably high levels.

I am not going to unpack the whole issue here, but if loose money caused unsustainably high equity gains, wouldn't the stock market of the 1970's have been going gangbusters?  The reason equities have been rising during periods of monetary expansion is because monetary expansion has been what the economy needs.  The rise has been mostly a product of healed economic wounds and improved expectations.  The fact that so many observers see rising stock prices as a reason to complain about the effects of Fed policy is just one more reason that we are lucky it hasn't been worse than it has.  There is a mood of self-destruction in the air.

But, here, again, housing is different.  Because banks and regulatory constraints are binding, monetary expansion through the credit channel does have direct effects on home prices - but it's complicated.  Since 2007 when the mortgage market froze up, home prices have been held below intrinsic value.  Because returns move inversely to prices, this has countervailing effects.  Rent is more stable than price, so falling prices meant that the returns to home ownership have increased.  In addition, the shortage in housing that this has created has pushed rent inflation higher, making returns to home ownership even higher.  But, these gains are only captured over time, as rents are collected (or implied).  And, since owner-occupiers aren't exactly marking their implied rent to market, this effect among owner-occupiers will tend to have considerable lag as an influence on behavior.  For renters, on the other hand, the hit to real income is felt immediately.  New homeowners feel the effect immediately, because they are marking to market when they engage in a real estate transaction, but this is countered by the home seller who realizes his relative capital loss.  Homeowners who buy new homes might feel some of this positive effect without as many mitigating effects on the selling party, but of course new home building has been very low.  And institutional renters do feel the positive effects of higher rents, which is leading to strong growth in construction for rentals.

The decline in price that leads to this higher implied return, on the other hand, is felt immediately by many homeowners.  Homeowners with high equity levels may have the same lagged reaction to unrealized capital gains as they do to the high implied return.  But, homeowners that are leveraged or that would be tapping into home equity for cash, would feel this effect immediately.

So, frozen mortgage markets and tight monetary policy have lowered home prices.  This simultaneously creates higher inequality in reported real household incomes, deflationary pressures through the credit channel, and a supply shock felt only by renters.  It's a sort of bizarro Cantillon effect - the economic advantaged gaining income at the expense of the disadvantaged, simply as a result of the change in an asset price because of Fed cash.  Except this is due to a dearth of Fed cash and a decline in asset prices.

If mortgage markets can expand, the direct effect on asset prices (specifically homes) would be much stronger than it is in the regular Cantillon effect.  But, it would lead to a reversal of the problematic effects of the bizarro Cantillon.

Here is a provocative new paper that points to the credit problem as a cause of the employment crisis.

Revisiting my Housing Narrative

Real estate values would be nearly double where they are if credit markets had been functioning.  Some of that value has probably been lost forever because of the depth and length of the downturn.  And, that value would have been roughly divided between the value of new building and the nominal increase in the value of existing real estate.  One reason home prices were so high in the 2000s was because long term interest rates were low.  But, one reason was that we weren't increasing home supply fast enough, so that less of that growth in real estate value was coming from growing supply, and more was coming from price appreciation.  As I mentioned, rent inflation was pretty high throughout the period.  Also, excess returns to real estate, according to the BEA, were high throughout the period - another sign that there were frictions to supply growth.

Monetary policy was so tight in the 2000s that the increase in equity that came from rising prices did not lead to excessive inflation.  The late 1970's is a good example of loose money and high inflation, and during that time real estate market values also increased above trend.  But, as we know, inflation ran between 6% and 12% during that period.  To call the 2000's loose and inflationary, in the absence of consumption inflation is very debatable.  I'm surprised at how confidently this narrative is so widely accepted.  I argue that frictions in home supply were creating a supply shock in the 2000s, which was creating supply-based inflation.  Tight monetary policy was pulling "Core minus Shelter" inflation below 2% for this entire period.  Shelter inflation was above 2% for the entire period.

So, I'm crazy, right?  Well, here is a graph of annual growth in closed end real estate loans.  The trend through 2006 was for more than 11% growth, annually.  What looks more out of line?  The 2000's or the 2008-2014 period?

Here's a graph of the market value of household real estate, with a log scale.  Until 2006, there was a quite stable 8% annual growth rate.  (Again, this is roughly divided between nominal price appreciation and new building.)  Again, I ask.  What looks more out of line?  The 2000's or the 2008-2014 period?  The trend line through 2006, when real estate market values were at about $25 trillion, would now be above $45 trillion.  So, should the burden of proof be on the narrative that says we should be on the 50 year long trend line?  Or, should the burden of proof be on the narrative that says we should have seen a sudden and unprecedented destruction of $20 trillion in household nominal wealth?

Man, it's hard for me to remain civil when I work through all of this stuff and think of SJWs sniveling "The banks did this to us." while they complain about loose money as if it's a "bailout" and constantly remind us how concerned they are about inequality.

A very early sign of recovery in this area appeared today.  This needs confirmation, as it can be a noisy indicator.  A big jump in mortgage applications.  Since interest rates aren't the binding constraint on monetary expansion right now (probably the topic of my next post), mortgage expansion should create a fantastic economic context, regardless of Fed rate decisions this year.  Let's hope.

Tuesday, January 13, 2015

November 2014 JOLTS

JOLTS data continue to come in strong.  Older workers tend to have lower unemployment but longer unemployment durations.  Specialization, better personal financial safety nets, etc. create frictions in re-employment so that I think we are seeing demographic-based movements in openings and quits.  Openings are higher, quits are lower, and hires are lower.  If we split the difference between openings and quits, JOLTS data suggest the labor markets are comparable to something around late 2005.

The unemployment rate was under 5% then.  I had been expecting this to mean that we would continue to see the unemployment rate drop sharply, but I am starting to think that there is some persistence in measured unemployment.  There might have been a shift right in the Beveridge Curve (x=unemployment, y=openings), even after adjusting for demographics, as there had been in the 1970's and 1980's after there had been several recessions in relatively short succession.  Basically a higher NAIRU, I guess, which I guess is basically the consensus view right now.

So, while I don't abide the supposed fear of wage-inflation, I do think that we should see the benefits of an economy running at full employment, even if the unemployment rate is a little high, which should mean higher RGDP growth, decreasing risk premiums, increasing real wage growth, and increasing real interest rates.  According to JOLTS, we are a long way from any concerns.

Monday, January 12, 2015

December 2014 Employment

The unemployment rate came in at 5.56% - almost down to 5.5%.  Hmm.  Going back as far as the beginning of 2014, the Fed was forecasting around 6.2%-6.4%.  Was there anyone with the foresight and courage to forecast something like 5.6% by year end?  'Cause that would be somethin'.  HmmHuh. :-)  (Note for journalists, that's "Erdmann", with two n's.  The second "n" stands for "Nailed it!")

Here is my chart of total unemployment, compared to insured unemployment - plus the zoomed in version.  The black line was my unemployment forecast based on a linear trend in the decline of very long term unemployment.  The red line was my unemployment forecast, assuming that after June 2014, very long term unemployment would stop declining.  December came in between the two.

I will probably revisit these charts less often now, since the trend will be less pronounced as we proceed from here.

Next, a look at unemployment by duration and at my estimate of very long term unemployed workers.  Most of this month's decline came in the shorter durations.  Durations above 26 weeks are still a little persistent, suggesting that further declines in the unemployment rate might be tougher to come by.  My estimate of very long term unemployment is showing a slight downward trend again, but this will probably not be a steep as it has been.

Following the graphs is a table comparing September unemployment with the estimates I had at the time for December and with actual December numbers.  Here again, we can see that the declines even over that time frame have been due to unexpected declines at the short end, and that long term unemployment is near my pessimistic projections.

There could be some bounce back from here.  I wouldn't be surprised if, several months from now, there is at least one month that still prints at 5.5% or 5.6%.

Looking at employment flows paints a more positive picture.  Flows between unemployment and employment were an aberration in November, I think.  This month flows into employment bounced back, and I don't see any reason to expect this net flow (from UtoE) to suddenly tighten up.  We should expect about 0.2% of the labor force to continue to move from UtoE in a typical month.  At this point in the recovery, we might expect there to be flows back into the labor force.  But, this month, there was a tick up in flows directly from Employment to Not in the Labor Force.  And, there were no net flows from Not in Labor Force to Unemployed.  Of those two flows, it is NtoU that would affect the unemployment rate, and as with UtoE, there is no reason to expect a sudden change in behavior.  Outside of the most intense parts of labor corrections, like in 2008-9, the net flow NtoU hasn't typically been more than about 0.1% of the labor force since the 1990's.  Looking at flows suggests that inertia in these flows should continue to pull unemployment down at a nice pace.  Maybe each of these flows will converge a little bit, causing unemployment to level off without any sharp changes in trends.  But, as the last graph suggests, UtoE flows dropping below about 0.14% has been a recessionary signal, and I just don't see that happening soon unless the Fed knocks the wind out of us very quickly.

Looking at flows data makes sub-5% by the end of 2015 seem very possible.  And, looking at flows data, it's hard to imagine that we could still be teasing around 5.5% a few months from now.  We'll have to wait and see how the mystery unravels.

Interestingly, interest rates have dropped after the employment rate was published Friday.  I think the market has moved on from unemployment levels to wage growth as the focus of potential Fed policy triggers.

I think I'll have a post up tomorrow on that topic.