Is the U.S. housing market about to recover? In my opinion, that depends on whether houses have become affordable for Middle Americans (defined below). If houses have become affordable, excess inventory will be absorbed and new home construction will grow. Until houses are affordable for Middle Americans, housing market progress may come in fits and starts, but it is not likely to be sustained.
Housing led America out of the 2001 recession, as well as the recessions of 1973-74, 1981 and 1991. It is difficult to see a true economic recovery from the Great Recession if housing does not play a positive role.
As you will see from the various data discussed below, affordability is hard to measure. It has many components, some of which are readily measurable, but some of which are not. And those that are measurable do not all point in one direction. My conclusion is that a prudent investor should now be in the housing market to some extent, but I am not ready to declare that the bull is at the door.
I am going to try to give you the tools you need to make your own evaluation, not of what other people are thinking right now, but of what is likely to happen to the housing market in the next year or so. For discussions of immediate trends, look here and here.
Can Middle Americans afford to buy the same home as in 1991 at the current inflation-adjusted price?
How will we know when homes are affordable? I think the key determinant is whether the third quintile of Americans by income - the people I call Middle Americans - are earning enough to be able to pay about the same amount that they paid in 1991 (in inflation-adjusted dollars). I pick the year 1991 because it was the beginning of the 15-year bull market in house prices that ended in 2005. 1991 also was the last year before the Great Recession in which personal incomes fell in the U.S.
I select the third quintile because it is, by definition, Middle America. To get a sense of where Middle Americans stand economically, let us begin by reviewing how income distribution has changed since 1991, which is the subject of our first table:
|Quintile||2010 mean family income||change from 1991|
As we can see, over the last 20 years, income distribution has changed significantly. The top 5% of earners should have no problem buying houses, and the next 15% should be just fine, as well. But when we look at the second and third quintiles, where, by definition, a large part of the middle class are found, we can see that they have made relatively little economic progress over the last 20 years. If we divided up the last 20 years, we would find that they did OK for the first 10 years, but stagnated or backslid over the last 10 years, particularly since the advent of the Great Recession. The mean income of the third quintile, the one on which we are focusing, grew only 5% since 1991. But even the second quintile's mean family income grew only 10%.
We can see the income progress or lack thereof from a slightly different perspective in the next graph, which I have made from data available at bea.gov. This graph shows how average disposable personal income continued to grow until 2009. The two top lines are disposable income per capita. The dark blue line is disposable income per capita in constant dollars, which I believe is the most important line to look at. But if you look closely, you will see that wages and fringes (the gray line) did not grow as fast as disposable income as whole (the brown line). This divergence reflects the growing disparity between the top earners, who often make entrepreneurial and investment profits, and the bulk of the population.
House Prices Since 1991
With this background of family and personal income growth, we turn next to the question of whether house prices, expressed in constant dollars, have come down sufficiently that they are within 5% of their 1991 level. (At that level, the growth in house prices since 1991 would match the growth in third quintile mean family income.)
There is no single standard for evaluating national house prices, much less the improvements in construction and amenities, as well as larger house sizes, that have taken place over the last 20 years. Blogger JP has made an attempt to compare house prices over the last 40 years, taking changes in house size into account. His data is reflected in the following graph.
As you can see, the narrow red trend line suggests that house prices are now back to trend. But they are still, even in constant dollar terms, and after having adjusted for house sizes, above 1991 prices by approximately 15%.
If we simply compared the 15% increase in house prices since 1991 with the 5% increase in mean family income of the third quintile of American families, we would conclude that house prices remain, from an affordability point of view, higher than they were in 1991, at the beginning of the last housing boom.
But first I will show you my attempt to see whether JP's data is believable.
JP bases his data on the National Association of Realtors data series. If we look at house prices from a different data series, the Case-Shiller series put out by S&P, and deflate that data using the GDP deflator, we get a graph as follows, in which Series 1 is the basic data and series 2 is the deflated data.
What we can see is that, based on the deflated data, current house prices remain about 14% above the 1991 benchmark.
It is convenient that the two data series give us similar results. That similarity suggests that we may be accurate if we suppose that the 15% ball park is realistic.
It would be facile to conclude that house prices therefore still have 15% to fall. Or it would be facile to conclude that, since our third quintile of families are 5% better off than they were in 1991, house prices still have 10% to go to put us where we were when the last housing boom began. But price alone does not determine affordability. Interest rates, mortgage underwriting criteria, down payment requirements, real estate taxes, and natural gas, electricity and heating oil prices also may play a role in affordability.
Monthly Mortgage Payments Today Versus 1991
In 1991, the interest rate on a conforming 30-year mortgage was about 9%, according to Freddie Mac (FMCC.OB) data. Today, the interest rate on a conforming 30-year mortgage is closer to 4%, and the federal government is committed to keeping the rate low, which it can do, since it controls Fannie (FNMA.OB) and Freddie. The Federal Reserve also has an incentive to keep mortgage interest rates low because it now owns a great deal of long-term mortgage bonds that it has funded with short-term liabilities. For these reasons, although we do not know whether a Republican administration would follow the same policies as the current administration, it seems reasonable to expect that rates on conforming 30-year mortgages will remain in the 4%-to-5% area for several years.
The difference between the payments on a $200,000 mortgage at 5% and at 9% is substantial. The difference is $1,073 a month versus $1,609, or about 33%. If the 1991 loan were $200,000 and the 2012 loan were 15% more--$230,000-the payment differential would be $1,609 versus $1,234, still a substantial difference. Thus, from a payments point of view, the interest rate differential dwarfs the 15% differential in prices, and the house appears far more affordable today than it was in 1991. (At least anecdotally, house buyers are far more concerned with monthly payments than with absolute prices.)
Property Taxes Today Versus 1991
Property taxes also have an impact on monthly payments. Such taxes, are, by their nature, local, and I have found no national repository of data. However, it should be possible to back into an estimate by comparing aggregate home real estate tax deductions by all U.S. taxpayers in 1991 and 2010, dividing by the number of returns on which such deductions were claimed, and deflating the result to constant dollars. I have not yet succeeded in getting these numbers from the IRS. I will supply them in a comment when I have them. My guess is that increased real estate taxes, in constant dollars, will add $50 a month to the carrying costs of the average house. That will not be sufficient to offset the gains from lower interest rates.
Electricity, Natural Gas and Heating Oil Today Versus 1991
The cost of electricity to residences does not appear to have gone up significantly in constant dollars. It is up about the same amount as the CPI in general. Heating oil prices, in constant dollars, have about doubled. Thus, a house that cost $100 a month to heat in 1991 costs $200 to heat today. This would add about $100 per month to the cost of owning an average home in the northern states, but that would not be sufficient to offset the advantage that lower interest rates have conferred on the same home. Natural gas prices, by contrast, are up by less than 50% in constant dollars, so the additional cost would be less than $50 per month. Again, that increase is dwarfed by the benefit of lower interest rates.
My conclusion is that, although energy costs are higher than they were in 1991, they should not have a major impact on house affordability at this time. But if we add higher energy costs and higher taxes, we will see a significant erosion of the benefits of lower interest rates.
Thus I conclude that homes are, based on monthly costs, more affordable today than they were in 1991, due to the interest rate differential. But we may want to note that in the mid-1990s, mortgage interest rates came down substantially. A comparison with, say, 1996 rather than 1991 might well yield much closer monthly costs.
National Association of Realtors (NAR) Affordability Index
The NAR maintains a housing affordability index based on current interest rates, the price of a median home, and the earnings of a median family. The Index over the last 25 years is shown in the following table:
The HA Index (HAI) measures house affordability as a percentage of monthly cost of a mortgage. For example, an HAI of 120.0 means a family earning the median family income has 120% of the income necessary to qualify for a conventional loan covering 80% of a median-priced existing single-family home. Therefore, according to the HAI, whereas a median family had only slightly more income than needed to buy a median home in 1991, in 2010 such a family had 169% of the necessary income.
The HAI estimate of affordability is considerably more optimistic than my current estimates.
The Down Payment Issue/Housing Gridlock
But even if the house is affordable, can our Middle American get a mortgage loan today as readily as in 1991? Formal underwriting standards seem to have gone back to something similar to 1991. In large measure, this will depend on whether our Average American family that has continued to pay their bills on time has the money for a down payment. Although down payment requirements for conforming loans have returned approximately to where they were in 1991, Middle Americans may not be as well positioned to make the downpayment today as they were in 1991. This is because Middle Americans who took out mortgage loans in the period 2003 to 2007, either through cash-out refis or new home purchases, largely had their equity wiped out by the decline in house prices from 2007 to 2011.
Nationally, there are about 50 million home mortgages. Of that 50 million mortgagors, 10.7 million (or about 21%) are estimated to have negative equity. See CoreLogic data here. Since, historically, most families that were not first time home buyers got their down payments directly or indirectly from the sale of their prior residences, those 10 million or so families may be shut out of the market to buy a home for some time. In addition, families that have lost their homes to foreclosure (several million more families) are likely to have a hard time getting another mortgage for a number of years.
There is some literature on the extent to which gridlock results from the large number of underwater mortgages. My thanks to "urbanexus" for pointing me in the direction of this literature in a comment on my article here. My review of that literature leaves me continuing to believe that for households that are making ends meet and that are neither unemployed nor behind in their mortgage payments, the lack of equity in their homes is a significant impediment to moving up, downsizing or moving to a different climate, all of which have typically been important reasons for buying another home. Moving to take a new job may be an exception.
Thus, although on the surface, credit availability looks about the same as in 1991, there are large numbers of Middle Americans who will not be able to qualify because the real estate market decline over the past five years has destroyed their equity. There also may be many families that have experienced financial difficulties and have not been able to pay all their bills on time. These families may not have adequate credit scores. I have not been able to quantify these impediments, but it seems likely that they will reduce the pool of potential buyers by a substantial percentage.
The FHA to the Rescue?
On the other hand, FHA loans are now far more readily available than they were in 1991, and they allow purchases with low down payments. But FHA loans are available only to borrowers whose income is not greater than 80% of the local median income. That would cover our fourth quintile and, probably, the bottom of our third quintile, but not most of it. Since the fourth quintile has a median family income of only $28,000 and has not improved its position at all since 1991, it seems to me that prices would have to fall a good deal further before the bulk of the fourth quintile could afford to buy. Moreover, the fourth quintile are many of the same people who got put into homes by subprime mortgages they could not afford. They are likely to be more cautious this time around.
The Impact of Household Formation during the 2000s
The following graph from the National Association of Realtors shows housing starts since World War II. The site is here.
We might expect that the very low level of house building since 2007 would indicate that the excess building of 2003-2007 would have been absorbed by population growth and the formation of new households over the last decade.
But we also need to evaluate the impact of changes in household formation due to changes in culture. Calculated Risk has a nice table that shows household formation over the last few decades. See here. The table shows that although the U.S. added 11.2 million households in the 2000s, that was fewer than in any of the preceding three decades. Homebuilding in the first six years of the decade of course exceeded homebuilding in any previous period, leaving the potential for oversupply to persist for several years.
A recent New York Times article here discussed significant changes in American demography, largely based on the fact that over 50% of births to women under 30 years old are out of wedlock. This statistic and related reporting show that household formation in recent times is quite different from household formation in earlier eras. Whereas in earlier eras household formation typically involved marriage and its relative permanence, today's household formation very frequently - perhaps even predominantly - involves unmarried couples or single women with children. The capacity of these new households to buy homes may well be significantly less than the capacity of new households in earlier eras, and therefore the normal push from the bottom of the market by first-time home buyers may be diminished.
Thus, although the population growth of the last decade, combined with the low rate at which new houses have been built for the last four years, might have been expected to absorb the excess house building of 2003-2007, the changes in household formation may well have negated that possibility. For this reason, I am unable to tell exactly where we are in the process of absorption of the excess supply.
Unsold Homes and Foreclosures
Obviously, I have not yet dealt with the overhang of unsold houses or foreclosed houses. I have not done so because I believe that, although they certainly constitute a headwind, those problems will work themselves through relatively quickly once houses are affordable for Middle Americans-and the process of working them through will be profitable for much of the real-estate-related industry, even if not for the homebuilders themselves.
Recent evidence suggests that the absorption process already is well along. See the following graph from Calculated Risk:
The Great Recession Will Not Really End until House Building Recovers
As I said at the outset, housing typically leads America out of recessions. I do not expect this time to be any different. Therefore I do not expect a vigorous recovery until housing begins to recover. This is because (1) housing was such an important part of the Great Recession, (2) global competition has made jobs scarcer, and house building is the kind of local job that is not yet outsourceable, and (3) a good housing market spurs so many other kinds of consumer expenditures for durable goods. All homebuilders, suppliers, real estate agents, title companies, appliance manufacturers, etc. etc. will benefit once the les bon temps roulerai. There is tremendous pent up demand, as houses have tended to deteriorate, band aids have been put on appliances, and home improvements have been deferred. But is now the time, as some of the possible early indicators suggest?
Remodeling Activity-Harbinger or Substitute?
Remodeling activity in some parts of the country already has begun to grow. See the following graph, courtesy of calculatedriskblog.com.
Does this mean that people are giving up on moving? Or does it mean that foreclosed properties have to be extensively remodeled in order to be habitable? I do not know. But it does foreshadow, I believe, the huge potential for spending on remodeling in the near future, which should be good at least for Home Depot (HD) and Lowes (LOW), both of which have gone up recently.
Homebuilder Stocks and ETFs
The homebuilder stocks also have gone up recently, from their summer 2011 lows (see here). But if we look at the five-year chart of iShares Dow Jones Home Con (ITB), which is the next graph, we can see that it is still basically mired in the trough created early in 2009. Therefore, if you conclude that houses are nearing affordability for Middle Americans, it is not too late to jump on the bandwagon. ETFs with similar housing exposure include iShares Dow Jones US Real Estate ETF (IYR), iShares Cohen and Steers Realty Majors Index Fund ETF (ICF), SPDR DJ Wilshire REIT ETF (RWR), and SPDR Homebuilders ETF (XHB).
My own favorite housing stock is Simpson Manufacturing Co. (SSD), manufacturers of Simpson Strong Ties, among other products. The stock has held up pretty well through the hard times, due, I believe, to good management as well as sound products. But the ETFs are a good way to play the recovery, if it is coming.
The Data Is Equivocal
I have provided so much data in this article because I believe the data is equivocal. Each investor will have to make a judgment. The market already has made a judgment that homebuilders will make more money in the future than they have in the recent past. But that judgment has yet to be confirmed.
At the present time, I am not confident that Middle Americans have the ability to make the necessary down payment to move from one house to another. The fourth quintile of income can get FHA loans, but I do not see that demographic group leading the economy, especially in light of the other demographic factors identified in the New York Times article I discussed above. For these reasons, I believe that signs of recovery are likely to be harbingers of some progress but not likely to be signs that the housing leadership that we are used to associating with strong economic recoveries is about to begin. Personally, I will be putting my toes a little further into the water so as not to be completely left behind if I am wrong, but I am not ready to make any big bets.