As everyone who has taken Introduction to Voodoo Bullshit, better known Econ 101, can attest the following chart is basically as ugly as it gets: in simple terms when you have a collapsing demand curve coupled with a surge in supply, the bottom line is that no matter how much intervention is involved, nothing can help to restore the pricing equilibrium to its old level (at least not for a long, long time).
And as can be expected from economists, despite having come up with the S-D concept, they consistently focus on the part that's (relatively) easy to control - the supply side, and tend to ignore the "demand" aspect, which is far more difficult to jigger in the desired direction (think the constant blaming of banks for not lending when it is in fact the consumers who do not want loans). As such, using data from Bank of America, we focus on the complete picture, with an emphasis on the much ignored Demand side of the home price equilibrium, to conclude that prices are set to drop much lower from current levels.
The simplest way to analyze the outlook for the housing market is to compare the evolution of housing demand and supply. Housing demand comes from the creation of new households and purchases of vacation homes. Households are defined as the number of separate housing units, either families or individuals. A household can be created when children move out of their parents’ homes, couples separate, or roommates decide to live apart. Housing supply is a function of new construction and, in today’s market, foreclosures. The foreclosure will only create net new supply if the former homeowner moves in with friends or family rather than becoming a renter. By this reasoning, we also do not count “turnover” – transactions from existing owners – when analyzing demand and supply. If homeowners decide to sell their home and move to a new house or a rental unit, it will show up as both supply and demand. As a result, we are only focused on new housing demand and vacant housing supply.
Of the above paragraph, the key notable is the highlighted sentence, or the ever critical "household formation" variable. Unfortunately as the chart below shows, household growth has plummeted over the past 3 years, declining by about 700,000, after clocking in materially positive numbers in the 6 years prior.
So focusing on the far more critical Demand side:
Demand: fewer new households
Household creation depends on the state of the economy. The combination of high unemployment, weak wage and salary growth, and tight credit has led to a decline in household growth over the past few years. The two main surveys of household formation from the Census Bureau – the Housing Vacancy Survey and Current Population Survey – show that about 500,000 households were created annually over the past three years compared to an annual average of about 1.2 million during the first half of the decade (Figure 6). How can we explain such a notable drop in household formation?
Moving in with the folks
The obvious answer is to look at homeownership rates, which have tumbled to 66.9% from a peak of 69.2% in 4Q04. This translates to a loss of nearly 2.5 mn homeowners. Most of these homeowners became renters, which means they remain a household, but not all. As can be seen by the surge in the rental vacancy rate to 10.6%, it seems that there was not a perfect shift from homeowners to renters (Figure 7). This begs the question: what happened to these former households? There was doubling up among economically stressed households; in other words people moved in with friends or family. Many of these former homeowners were probably foreclosure victims (Figure 8).
As Figure 8 shows, household formation can also decline if there are fewer young households created to replace the aging homeowners. Given the nearly 10 point surge in the unemployment rate among 16 to 24 year olds from the trough to peak during this cycle, it seems like this was a considerable factor. A recent paper sponsored by the Research Institute for Housing America estimates that the probability of a young adult forming a household declines by 4% during a recession, and up to 10% if unemployed. In addition to the slowdown in “headship rates” domestically, there was a drop in household formation from immigration. According to the Office of Immigration Statistics at the Department of Homeland Security, the number of unauthorized immigrants decline by 1.0 million from 2007 to 2009 compared to a net gain of 1.3 million from 2005 to 2007.
Household growth to improve, but with a lag
Household formation will naturally pick up as the economy improves, but if our forecast for a sluggish recovery is realized, household growth will also be lackluster. The main factor influencing household growth will be the state of the labor market. The above-referenced paper finds that the unemployment rate must fall by 2pp from current levels to return to normal rates of household formation of about 1.2-1.4 million a year. We do not expect the unemployment rate to reach the mid-7% range until 2013, implying another two and a half years of sluggish household formation of about 800,000 a year. This is also when we expect the pace of foreclosures to slow notably, which means that fewer households will have to double-up.
Looking ahead to 2013 and beyond, we use forecasts from the Joint Center for Housing Studies at Harvard University. They present two possible trajectories for household growth: 1) an average of 1.48 million annually through 2020 assuming net immigration returns to the 2000-05 pace and headship rates at 2008 levels; and 2) an average of 1.25 million annually through 2020 assuming the same 2008 headship rates but slower immigration. We believe the latter is more likely and use this as our baseline forecast (Figure 9).
Renters will take market share
Although we expect household formation to start to improve in 2013, the homeownership rate should still fall further, suggesting that most of the gain in households will be due to an increase in renters. This is because there is still a considerable number of homeowners with mortgages in some stage of delinquency that are likely to end in foreclosure. Based on data from the Mortgage Bankers Association, there are about 5.5 mn seriously delinquent mortgages currently outstanding.
A recent paper by economists at the NY Federal Reserve (Haughwout, Andrew, Richard Peach, Joseph Tracy. “The Homeownership Gap”, Federal Reserve Bank of New York Current Issues in Economics and Finance, Volume 16, Number 5, May 2010) attempts to quantify the effective lower bound for the homeownership rate. They make the assumption that underwater borrowers (negative equity), who currently account for about a quarter of mortgage holders, will transition to renters over time. Subtracting these underwater borrowers yields an “effective homeownership rate” of 61.6% (Figure 10). This would be a record low in the data which goes back to 1965. We do not expect such a precipitous drop because not all underwater homeowners will become renters. Indeed, a recent study by Trulia.com and RealtyTrac found that 59% of respondents would not go into foreclosure simply because of negative equity. We believe it is more likely that the homeownership rate will bottom at 65%, returning to mid-1990s levels.
It is plainly obvious why the demand-side is so often ignored in polite conversation: it is the consumer-driven aspect of the house price variable, over which neither the Fed, nor the Treasury, nor the FHA has any authority, and which is a function purely of expectations of the future. Alas, those right now are lously and getting worse. We expect that Demand-side housing economics will take on progressively more importance in the future, as it becomes obvious that no amount of Supply-side tinkering will prevent another 20% drop in prices.
And speaking of Supply, this is also a critical factor, if much more prevalent in the daily media. Alas, that in itself does not make the problem any easier to resolve.
Supply: out of balance
The drop in housing demand triggered homebuilders to slash new construction. In theory, this should have balanced the market. A shift lower in the demand curve will temporarily depress prices until the supply curve shifts to balance the market at a lower quantity. The problem is that this model does not take into account the additional source of supply: foreclosed properties that have returned to the market for sale. As a result, housing supply has not normalized and greatly exceeds housing demand, creating a large imbalance in the housing market.
We can measure this imbalance two ways. The simplest method is to calculate the number of excess vacant homes for sale. Assuming a normal homeowner vacancy rate of 1.7% and rental vacancy rate of 8%, there is an excess of 1.87 mn vacant homes (again, see Figure 7).
Another, more detailed, way to measure the imbalance is to estimate the gap between housing supply and demand over the past few years. We define housing demand as the sum of household formation, demolitions and purchases of vacation homes, and housing supply as the sum of new construction, mobile homes and foreclosures from households that double up (Demolitions: Assume historical average of 0.24% of housing stock; Vacation homes: annual survey from National Association of Realtors which is about 10% of sales – we assume that 80% plan to use the home for vacation rather than rent it out and of which about three quarters plan to keep their first home; New construction: single and multi-family completions; Foreclosures: measure REOs (real estate owned) – we assume that half of the former homeowners will double-up). We estimate an excess of about 2 mn homes created over the past four years, consistent with our estimate from the vacancy data.
It will take years to clear the excess
Our mortgage strategists expect approximately 6 mn additional foreclosures to enter the market for resale over the next three years. It will then take a few years for the pace of foreclosures to gradually normalize to about 200,000 a year. If we plug in our baseline forecast for household formation (as explained above) and assumptions for the remaining variables (as explained in the footnote), we can run a few scenarios for the amount of time it will take to clear the imbalance under different paths of housing starts. At one extreme, if housing starts fall to zero, translating to zero completions next year, the excess supply will be cleared by early 2013. At the other extreme, if housing starts surge back to the historical norm of 1.5 mn next year and hold indefinitely, there would be another 2 mn homes added to the excess, bringing the total to 4mn, which would take until 2027 to offset. We believe the truth falls somewhere between these two extreme scenarios.
If we pencil in our baseline forecast for housing starts of 590,000 this year and 690,000 next year, another 500,000 excess homes will be created. Looking ahead, we must be more judgmental. A reasonable scenario is that starts slowly edge higher to 1 mn by 2013 and reach the “normal pace” of 1.5 mn by 2015. At this point, most of the excess supply will have nearly cleared, allowing starts to pick up to match the pace of demand (Figure 12).
Why any building given shadow inventory?
We often get asked why builders would start new construction given the considerable number of vacant homes on the market for sale. As the above example showed, if housing starts fell to zero, the market would return to normal much quicker.
There is a good reason for new construction: a foreclosed home is not a perfect substitute for a new home. Foreclosures will sell at a discount to a new home, but will often require a great deal of renovations. This will discourage some buyers. In addition, foreclosures are not equally distributed geographically or across price ranges. The bulk of foreclosures are in the lower end of the housing market. This creates two types of markets: “passive” and “active:” The passive market has the undesirable foreclosures, which are either in very poor condition and/or in a foreclosure-dense, and therefore suffering, neighborhood. It is likely that many of these homes will remain vacant and on the market for sale or rent. In contrast, the active market will contain foreclosures that can compete with regular homes in a market with housing demand.
The regional difference is particularly interesting. According to RealtyTrac, 37% of foreclosures are in either California or Florida and 65% are in the 10 states with the highest foreclosure rates. Even within these states, there are stark differences between zip codes. Homebuilders will avoid building new homes in the areas with a large presence of foreclosures. In the first half of the year, only 27% of housing starts were the top 10 foreclosure states (Figure 13). Comparing two large states is noteworthy: about 22% of foreclosures were in California, but only about 6% of housing starts, compared to Texas, which had only 4% of foreclosures and 15% of housing starts.
So now that both the very dire Demand and Supply sides of the pricing equilibrium have been discusses, what does this imply for the broader economy?
Unlike in prior recoveries, it is clear that housing will lag rather than lead the recovery. The monetary policy transmission typically has a very strong impact in the housing market – low rates encourage home sales and greater residential investment. In turn, job creation picks up in the construction sector, further supporting consumer spending and home sales, and creating a virtuous cycle. This feedback loop is currently broken. Mortgage rates have plunged to record lows, and yet have done little to stimulate home purchases because credit conditions are still incredibly tight and consumer confidence is depressed. That said, we believe the direct drag from housing, through construction, is nearly over. Our forecasts for housing starts imply that residential investment will subtract from growth in 3Q, but then consistently add to output going forward. Still, the contribution to growth will be feeble relative to prior cycles, where housing was decidedly the force of growth (Figure 14).
While housing construction is the most direct link to the economy, it is not the only one. The path of home prices is also very important as consumers respond to changes in housing wealth. Our baseline view is that national home prices will edge lower over the rest of the year and then bounce around the bottom for some time. The downside risk is that foreclosures flood the market at a rapid pace, depressing home prices greatly, which could tip the overall economy back into recession.
Alas, we are already there, and yes, it is BofA's job to put a favorable spin on the data. Look for another 15-25% drop in home prices from here on out, and another wave of hundred billion+ charge offs at undercapitalized banks.