National Housing Market Update

Here's an overview of recent stories about the nation's housing market.

Housing starts fell more than 10% in October, hitting a six-month low. Pundits are attributing the unexpected decline to uncertainty about the fate of the first-time homebuyer credit. The decline is troubling because housing starts historically have been a good leading indicator of overall economic activity.

The delinquency rate on residential mortgages hit a record high of almost 10% in the third quarter. That doesn't include the 4.5% of mortgages that are in foreclosure. The combined total of over 14% is also a record, going back to 1972 when the Mortgage Bankers Association initiated its delinquency survey.

The vacancy rate for apartments hit a 23-year high of 7.8% in the third quarter. The vacancy rate is expected to climb further during the fall and winter quarters, when rental demand is seasonally weak.

The unemployment rate rose to 10.2% in October, reaching double digits for the first time in 26 years. The highest rate recorded during the six decades since the government began compiling data was 10.8%, which occurred in 1982.

Despite all the bad news, existing home sales rose 9.4% in September, to a two-year high of 5.6 million units on an annualized basis. The recent strength of demand is at least partially attributable to the first-time homebuyer credit, which is scheduled to expire in the middle of 2010.

Don't Count on Silicon Valley to Revive Housing Market

Silicon Valley is the economic engine of the Bay Area, but it is not as dominant as many people seem to believe. Take a look at the chart, below, which shows technology sector employment in the Bay Area over the last twenty years.

The numbers in the chart were obtained from California's Employment Development Department. I've defined the technology sector to include computer design and manufacturing, research and development, internet, and software development.

The first thing to notice is that the tech sector is not all that large. There were 250,000 tech jobs in September of 2009. That represents only 15% of the 1.7 million jobs in the San Francisco and San Jose metro areas. And keep in mind that many tech jobs are in supporting roles such as secretarial, accounting, and sales.

The relative size of the tech sector surprised me. What surprised me even more, however, was that the number of tech jobs has never re-approached its dotcom peak. The number of tech jobs in September of 2009 was 33% lower than it was at the peak, and only 10% higher than it was 20 years ago.

The tech sector exerts a significant influence on the Bay Area housing market. Technology employees usually earn high salaries, so they can afford to pay top dollar for homes. But at only 15% of all jobs, the tech sector doesn't dominate the housing market. And judging from the record of the last twenty years, it isn't becoming significantly more important.

Note: Thanks to one of my readers, who pointed out that I had mistakenly compared the number of tech jobs in the San Francisco and San Jose metro areas to the total number of jobs in California. That was an embarrassing mistake. The correct comparison is not as striking as my original, erroneous comparison, but the original conclusion remains valid: The tech sector does not dominate the Bay Area housing market, and does not appear to be headed that way.

Banks Are Still Tightening Mortgage Standards

Mortgage delinquencies have been rising rapidly for the last two years. Not surprisingly, banks have been tightening their standards for new mortgages. Take a look at the chart, below, which shows the net percentage of banks that tightened credit standards for residential mortgages in each quarter.

The chart was derived from the Federal Reserve's quarterly survey of senior loan officers. The survey asks, among other things, if banks tightened or loosened credit standards during the current quarter. The numbers in the chart were calculated by subtracting the percentage of banks that loosened standards from the percentage of banks that tightened standards.

Banks have gone through cycles of loosening and tightening in the past, but the current cycle is unprecedented. Beginning in the fourth quarter of 2006, banks began tightening credit, reaching a peak level of activity in the third quarter of 2008. At that point, roughly 75% of banks were tightening credit standards for mortgages.

In some senses, the panic has receded. In the most recent survey, only 24% of banks said that they were tightening standards. It's important to interpret the chart correctly, however. It does not indicate that credit standards are actually loosening, only that fewer banks are still tightening. (Only 2% of banks actually relaxed their lending standards during the current quarter. That's the first time in a year that any survey respondent has relaxed its standards.)

The continuing clamp-down on mortgage standards probably explains why record-low mortgage interest rates have not had a larger impact on housing prices. It may also provide a clue about when the housing market will recover. The last time that the housing market went from boom to bust, prices didn't bottom out until the mid-1990's, long after banks had begun loosening credit standards.

Mortgage Delinquencies Are Still Rising

The housing market is showing tentative signs of stabilization, but you wouldn't know it from looking at mortgage delinquency data. Nationally, the delinquency rate on residential mortgages is at record levels, and still appears to be rising. Take a look at the chart, below, which shows the percentage of Fannie Mae's single-family home loans that are at least 90 days past due.

The delinquency rate for August of 2009 (i.e., the latest available) was 4.45%. That's almost three times higher than the 1.57% delinquency rate that was recorded in August of 2008, when the financial crisis was already well underway. While it's certainly not scientific, I think it's reasonable to look at the chart and conclude that delinquencies are likely to continue rising for some time.

Note: The Mortgage Bankers Association (MBA) publishes its own National Delinquency Survey. The headline number from the second quarter survey was 9.24%. The MBA figure is heavily skewed by the inclusion of subprime mortgages, which are largely absent from Fannie Mae's portfolio. More importantly, the MBA uses a lower threshold for 'delinquent', requiring only that the borrower has missed a payment.

At first blush, the situation appears to be somewhat better here in the Bay Area. Take a look at the chart, below, which shows the number of default notices recorded in San Francisco as well as the greater Bay Area.

Recorded default notices appear to be stabilizing. There's an important distinction, however, between a default notice and a delinquency. A borrower is said to be delinquent when he falls behind on his loan payments. When he falls far enough behind, the lender may record a Notice of Default, which is the first step in the foreclosure process. Recent changes in California law make the foreclosure process more difficult, so the number of recorded default notices probably understates current delinquency rates.

In other words, it seems likely that delinquency rates here in the Bay Area are still rising, just as they are at the national level.

Impact of Rent Control: $325,000 per Unit

San Francisco has strong rent-control laws. Once you put a tenant in a rent-controlled apartment, you can't raise his rent by more than about 2% per year, and you can't evict him except under special circumstances. Meanwhile, there is never a shortage of people who are looking for a place to live in San Francisco. That combination of laws and circumstances is guaranteed to depress the value of residential property, at least as far as property owners are concerned. But what is the financial impact? Here’s a way to think about it, based on recent sale prices.

The first clue that rent control depresses property values is that vacant apartments are usually worth more than occupied apartments. A vacancy represents an opportunity to replace an outgoing tenant (who typically would have been paying less than market rent) with someone who is willing to pay full price for a home. Take a look at the chart, below, which shows median prices for three-unit buildings sold in San Francisco since 2005.

The median price difference between an empty three-unit building and a full one is $265,000, or approximately $90,000 per unit. Roughly speaking, that’s the value of a one-time opportunity to increase the rent to current market value. (How much of an increase are we talking about? Apartment buildings typically sell for about 14 times gross rent, so $90,000 of extra value equates to about $535 per month of extra rent.)

Unfortunately for property owners, the impact of rent control is much greater than $90,000. Installing a new tenant usually will result in higher rent, but the new tenant will benefit from the same 2% annual limitation on rent increases as the old tenant. For comparison purposes, market rents in San Francisco historically have increased at around 4% per year. At those rates, the new rent will be 10% below market in only five years.

It’s straightforward to calculate the cost (in present value terms) of a 2% annual limitation on rent increases. However, it’s easier (and probably more reliable) to work with current market values. Specifically, since condos are exempt from the most onerous aspects of rent control, we can estimate the impact of rent control by comparing condo prices to apartment prices. Take a look at the chart, below, which shows the median price per unit for occupied apartments, vacant apartments, tenancy-in-common (TIC) units, and condos.

The price difference between a vacant apartment and a condo is $355,000. To be fair, some of the difference in price is attributable to differences in quality. As indicated in the chart, TIC’s typically sell for around $120,000 more than vacant apartments. Since there is very little difference (from a legal point of view) between a TIC and a vacant apartment, we should attribute most of the $120,000 price difference to quality improvements that generally are made when apartments are ‘converted’ to TIC’s.

That still leaves a difference of $235,000 to be accounted for. That’s the typical price difference between a condo and a TIC. The most obvious reason for the price difference is that interest rates on TIC loans are higher (by around two percentage points) than interest rates on condo loans. Ultimately, however, the price difference is attributable to rent control. Why? A TIC owner could reduce his interest rate simply by converting his TIC to a condo. However, the City makes condo conversion difficult, precisely because condos are largely exempt from rent control.

Adding it all up, the impact of rent control comes to roughly $325,000 per unit. That’s for the special case of three-unit buildings, but the impact is likely to be similar for other buildings. I wonder if the creators of rent control had any idea that they would be imposing such a large burden on property owners.

TIC’s Are Still Attractive for Developers

Developers have been buying San Francisco apartment buildings and converting them to tenancy-in-common (TIC) buildings for years. That’s because owner-occupiers are willing to pay more for homes than investors are willing to pay for rent-controlled apartments. Take a look at the chart, below, which shows historical sale prices for TIC’s and multi-unit (apartment) buildings.

TIC prices leveled out at around $500,000 in 2005, and have remained fairly stable since then. Over the same five-year period, multi-unit building prices have hovered in the neighborhood of $250,000 per unit. In other words, TIC’s have been selling for roughly twice as much as apartments. The price premium is easier to see in the chart, below, which shows the per-unit price difference between TIC’s and apartments.

What I find surprising about this chart is that the TIC price premium does not seem to have diminished over time. It dipped modestly in 2008, but has since bounced back to around $250,000 – the same level that it achieved during the peak years of the housing boom.

If I had several hundred thousand dollars in cash and were looking to invest in San Francisco real estate, I’d think about buying an apartment building and converting it to TIC’s. It’s not cheap to remove tenants and renovate a building, but it doesn't cost $250,000 per unit.

Short Term Rates Likely to Stay Low for Another Two Years

Short-term interest rates are basically at 0%. That’s good news for borrowers with floating-rate loans (which are tied to short-term rates), but how long will it last?

In January, Paul Krugman estimated that short-term rates would remain low through the end of 2011. He based his estimate on historical precedent: The Federal Reserve typically doesn’t start raising rates until two and a half years after recessions end. (At the time, most people were expecting the current recession to run at least another six months.) Now Krugman has made a more careful estimate, based on economic principles rather than historical precedent. His new estimate: The Fed can wait at least another two years to start raising rates, and probably longer than that.

Krugman’s argument is based on two common economic rules of thumb, namely, the Taylor rule and Okun’s Law. His argument is simple, but I thought it would be helpful to explain his rules of thumb.

The Taylor Rule

The Federal Reserve has two distinct mandates, namely, to maintain price stability as well as full employment. The primary tool for achieving these mandates is the federal funds rate. When inflation gets too high, the Fed increases the Fed funds rate in order to slow the economy and dampen inflationary pressures. On the other hand, when unemployment gets too high, the Fed cuts the Fed funds rate in order to stimulate the economy and increase demand for labor. Adjustments to the Fed funds rate are decided by the Fed’s Board of Governors, but a Stanford Professor named John Taylor devised a simple rule that historically has done a good job of mimicking their decisions.

One version of the Taylor rule says that:

Fed Funds Rate = 2.1 + 1.3 x Inflation – 2.0 x Excess Unemployment

Excess unemployment is defined as the difference between the actual unemployment rate and the ‘natural’ rate, i.e., the rate that would prevail if the economy were operating at its potential. Take a look at the chart, below, which plots the Taylor rule formula along with the actual Fed funds rate.

The Taylor rule does a pretty good job of explaining the past behavior of the Federal Reserve. The largest deviation between the rule and the Fed funds rate occurred between 2004 and 2005, when the Fed held short-term rates well below the level suggested by the Taylor rule. (The Fed was criticized at the time – partly on the basis of the Taylor rule – for helping to inflate the housing bubble by keeping rates too low for too long.)

The September unemployment rate was 9.8%. The natural unemployment rate is estimated to be about 4.8%, so excess unemployment is around 5%. Meanwhile, inflation has been running at around 1.6% (and appears to be trending down). If you plug those numbers into the Taylor rule, you get a Fed funds rate of -5.8%. Negative interest rates can’t exist in the real world (people would hold hard currency rather than lend it at negative interest), so the Fed has settled for the next best thing: 0% interest.

As long as the Taylor rule yields a negative result, the Fed will presumably want to keep the Fed funds rate at 0%. So when will the Taylor rule yield a positive result?

Krugman argues that inflation is unlikely to increase during the next couple of years. (Given the amount of idle capacity in the economy, it will be some time before competition for resources puts upward pressure on prices.) Assuming that he’s correct, unemployment must fall to around 7% in order for the Taylor rule to yield a positive result. How long will that take? That’s where Okun’s Law comes in.

Okun’s Law

Increases in population and improvements in technology drive long-term increases in economic output. Over shorter time periods, changes in output are also influenced by changes in the intensity with which resources are used. In particular, if there is a surge in the percentage of the population that is working, there will be a surge in output as well. Or to put it in more familiar terms, GDP generally rises faster than trend when the unemployment rate is falling. It turns out that this relationship can be readily quantified, in the form of Okun’s Law.

Take a look at the chart, below, which plots changes in unemployment against changes in GDP over the last 40 years.

Each data point represents a calendar year between 1969 and 2008. If you fit a line to the data, you get the following equation:

Change in Unemployment = - 0.4 x ( Change in GDP – 3% )

As an example, if GDP rises by 4%, unemployment should fall by 0.4%. (Incidentally, the equation also says that if GDP rises by 3%, unemployment shouldn’t change. In other words, the equation implies that 3% is the trend rate of GDP growth.)

With the economy coming out of recession, GDP is likely to grow faster than trend for the next few years. How much faster? Krugman suggests that 5% is a safe upper bound for GDP growth. (I agree. Full-year GDP growth has exceeded 5% only once in the last 30 years. Most economists expect the current recovery to be slower than normal.) With 5% GDP growth, Okun’s Law says that the unemployment rate will fall by 0.8% per year. With unemployment currently at 9.8%, it will be another three and a half years before the recovery pushes unemployment down to the 7% threshold which would justify a positive Fed funds rate.

Note: Krugman’s two-year estimate was based on a more conservative set of assumptions. I assume that his conservatism was intentional. In any event, my goal is to help explain his reasoning, not to quibble with his numbers.

What does this mean for homeowners? Most floating rate loans are tied to LIBOR, which is driven by the Fed funds rate. Krugman’s argument suggests that borrowers with floating rate loans can breathe easy for at least another two years, and probably a good deal longer than that.