No Sign of Economic Recovery

In a recent posting, I said that the financial panic has subsided, but that the housing market is unlikely to recover soon. Paul Krugman made a similar point last week, about the broader economy. I won't reproduce his comments here, but I will borrow his chart (which he borrowed from Goldman Sachs, in any case).

Together, the four series included in the chart represent a sizable portion of overall economic activity. They all fell significantly during 2008, and they've all been flat since the beginning of 2009. Economic activity remains at depressed levels, with few signs of improvement.

San Francisco Housing Market Recovering Strongly - For How Long?

The San Francisco housing market has strengthened significantly since the beginning of the year. Prices are still 10% to 15% lower than they were last summer, but that represents a marked improvement from January, when they were down by roughly 25%. Take a look at the chart, below, which shows the recent history of home prices in the City.

The solid blue line shows median sale prices for 2-4 bedroom single family homes in San Francisco. The dashed purple line shows median sale prices for 1-3 bedroom condos. I smoothed both series slightly, using a trailing two-month average.

In January, San Francisco home prices were roughly 25% lower than they had been just six months earlier. And they were on a trajectory that seemed to imply disaster. Since then, the economy appears to have stabilized. The recession isn't over yet, but the panic has subsided. Not surprisingly, home prices have bounced back somewhat, rising by about 10% since January.

Is the recent price trajectory sustainable? I'll save that for a future blog entry. Meanwhile, the chart below seems worth considering.

The solid blue line shows median sale prices for 2-4 bedroom single family homes in San Francisco. The dashed purple line shows the S&P 500 stock price index.

It seems clear that over the last couple of years, the housing market and the stock market both have been driven by the same underlying factors. That's hardly surprising, given the dramatic events of the last 18 months. The recent close correlation is unusual, however. Over longer time horizons, houses and stocks generally don't track each other very closely. Stock prices respond rapidly to economic news, limited only by investors' imaginations. In contrast, home prices are sharply constrained by the ability of consumers to pay for them. So while expectations of improvement in corporate profits may drive stock prices higher, continuing deterioration in the job market is likely to put a ceiling on home prices.

Note: You can find additional price charts for San Francisco and the Bay Area in the Market Stats section of my website.

The Panic Is Over but Don't Expect Quick Recovery in Housing

The World panicked when Lehman Brothers collapsed. Credit markets froze and stock markets nosedived as investors fretted about the risk of a second Great Depression. Inevitably, the housing market got caught up in the crisis of confidence. Starting from already low levels, most indicators of housing activity fell to record lows in the opening months of 2009.

Several months later, it appears that disaster has been averted. The direst indicators of distress have returned to levels that are merely concerning, and not terrifying. It shouldn’t come as a surprise, therefore, that housing market indicators have bounced back from recent lows. But stepping back from the brink of disaster isn’t the same thing as recovery. The housing market downturn was well underway before the panic started, and may still have some ways to go now that the panic is over.

Consider the following series of charts, which illustrate the spike in panic levels. The first chart, below, shows the so-called TED spread.

The TED spread is the difference between the interest rates on 3-month interbank loans and 3-month Treasury bills. The former is the interest rate that (non-US) banks pay when they borrow from each other; it’s commonly known as LIBOR. The difference between LIBOR and the Treasury rate is an indicator of the perceived riskiness of lending money to commercial banks.

By September of 2008, the TED spread had already widened significantly from pre-crisis levels. When Lehman Brothers collapsed, however, the spread spiked to more than ten times its pre-crisis levels. Evidently, major banks suddenly became concerned that if Lehman could fail, so could any other major bank. Indeed, almost every bank in the country has since received federal bailout money to keep it afloat.

The TED spread has fallen steadily over the last several months. It’s still at elevated levels – suggesting that banks aren’t out of the woods yet – but the worst of the crisis has subsided. Some of the larger banks have actually begun returning their bailout money.

The stock market also went into a panic in September. Take a look at the chart, below, which shows the VIX volatility index.

Volatility is a measure of stock price fluctuations. High volatility indicates that stock prices are changing rapidly. When Lehman Brothers collapsed, the VIX index went from around 20 to a record of over 60. The index has declined significantly since then, to its current level of around 25. That’s still higher than the levels that prevailed immediately before the crisis, but it’s not much higher than the long-run average of around 19.

It’s instructive to make a quick detour into option theory. Options can be thought of as insurance against big stock price movements. When volatility rises, the insurance gets more expensive. What the VIX index actually measures is not current volatility, but the volatility that’s implied by option prices. In other words, the VIX index is essentially measuring the cost of insuring against big stock price movements. Immediately before the crisis, insurance was much cheaper than it had been historically, suggesting that investors may have been overly complacent about risk. When the crisis was in full swing, the cost of insurance hit record levels.

The bond market provided what was perhaps the most alarming indicator of panic. Take a look at the chart, below, which shows the difference between the yields on ordinary 10-year Treasury notes and 10-year inflation-indexed Treasurys.

When investors lend money to the Government, they usually demand a certain base rate of return (the so-called ‘real return’) and then add something extra to compensate for expected inflation. For instance, if investors normally require a 3% real return, they’ll demand a yield of 5% on Treasury notes if they’re expecting an inflation rate of 2%. That’s because inflation will erode the purchasing power of their loan principal by 2% per year. The extra 2% interest is compensation for the 2% annual erosion of the purchasing power of their principal.

Inflation-indexed Treasurys are designed to compensate investors directly for the eroding effect of inflation. If inflation runs at 2%, the Government automatically increases the loan principal at the same rate of 2%. Since the purchasing power of the loan principal is protected from inflation, the regular interest payments don’t need to include any extra compensation. The yield on inflation-indexed Treasurys can therefore be taken as a direct measure of the real return that investors require on loans to the Government.

The difference between the yields on ordinary 10-year Treasury notes and 10-year inflation-indexed Treasurys provides an indication of the inflation rate that’s expected to prevail over the next ten years. Before the crisis began, expected inflation was running at around 2.5%. That was in line with the historical average for (expected) inflation. It was also consistent with the 2%-3% inflation range that the Federal Reserve seems to have targeted in the past.

When the crisis began, expected inflation fell essentially to zero, reflecting investor concerns that the US might experience a ‘lost decade’ similar to what Japan went through in the 1990’s. As with the TED spread and the VIX index, however, inflation expectations have reverted toward normal levels lately, and are now only somewhat lower than they were before the crisis. Presumably, investors are no longer concerned about the possibility of a protracted recession (although many economists believe that there is still a substantial risk that the US will experience a lost decade).

What does all of this have to do with housing? Much has been made of the recent upturn in housing market activity. Housing starts have increased for several months in a row, as have sales and even, in some markets, prices. Take a look at the chart, below, which shows the Housing Market Index, published by the National Association of Home Builders.

The Housing Market Index (HMI) is a measure of builder confidence. It’s a good proxy for residential construction activity.

The HMI index fell significantly when Lehman Brothers collapsed. From a level of 17 in September, it fell to a record low of 8 in January. Unlike the other distress indicators above, however, the HMI index had begun deteriorating long before the panic began. It bounced back to 15 in the latest survey, but the bounce is mainly attributable to the end of widespread panic, not a general recovery in the housing market. The index is still far below the neutral level of 50, which would indicate an even balance of optimism and pessimism among builders.

Bay Area home prices have risen from their recent lows. Is that an indication that demand is finally turning around, or is it a one-time bounce, resulting from the end of the panic? My money is on the latter hypothesis. Considering the severity of the panic, it would be surprising if home prices didn’t bounce back when the panic subsided. The recession is still in full swing, however, and jobs are still being lost at a rapid pace. (Unemployment rose by 0.5% in May alone.) People who are banking on a continuing recovery in home prices are likely to be disappointed.

KPIX Story about Multiple Offers - Tying Up Loose Ends

KPIX ran a story in March about a house that received 42 offers. I said at the time that the house was severely underpriced. Out of curiosity, I decided to see what happened to it.

The house in question was 555 Edinburg. It was listed at $459,000, which is equivalent to $367 per square foot. By way of comparison, the median price-per-square-foot multiple for the 16 comparable houses that I identified was $451. In other words, the house on Edinburg was priced at a large discount relative to the comps, so it should have received multiple offers.

The sale closed on April 22nd, at a price of $570,000. That's equivalent to $456 per square foot -- right in line with the comps. Take a look at the chart, below, which shows the distribution of price-per-square-foot multiples for the comps.

555 Edinburg landed almost precisely in the middle of the distribution. Keep in mind that the comps were sold over a five-month period between October, 2008 and February, 2009. Those were the darkest months of the financial panic that ensued following the collapse of Lehman Brothers. In other words, far from signaling a recovery in the San Francisco housing market, the sale of 555 Edinburg signaled continuing softness.

The good news, if there is any, is that the market clearly did its job. Despite what was no doubt a bidding frenzy, 555 Edinburg sold right where it should have.

Falling Prices and Rising Sales

In an earlier posting , I said that sharp price reductions have had much to do with the rising volume of home sales in the Bay Area. I should have included this chart with the earlier posting.

Each dot represents one of the nine Bay Area counties. The horizontal axis shows the year-on-year change in the median home price. The vertical axis shows the year-on-year change in the volume of sales. The dashed line is a least-squares fit of the data points.

The theoretical value of the line is dubious, but it does help to illustrate my point. The line intersects the horizontal axis at a value of -30%. That's how far prices had to fall in a typical Bay Area county in order to keep sales at the same level as a year ago. In counties where prices fell further, sales volumes have increased significantly compared to last year's levels. But that clearly shouldn't be taken as an indication of health in the housing market.

San Francisco Apartment Rents Falling Rapidly

Near-term fluctuations in apartment rents have little impact on long-term housing values. But they do offer a useful window into the health of the local economy. The view from that window doesn't look good right now. Take a look at the chart, below, which comes from SFRentStats.

The data comes from Craigslist, and covers the City of San Francisco.

Rents have fallen by about 20% since the spring of 2008. That doesn't mean that the sky is falling. Even after the decline, rents are no lower than they were two years ago, when the economy was generally thought to be on solid footing. As long as rents are falling, however, it will be difficult to argue that the San Francisco economy has turned the corner.

Price Declines Likely to Push Foreclosure Rates Higher

Why do homeowners default on their mortgages? Researchers generally focus on two main reasons: job losses and price declines. Job losses directly impact default activity, because they reduce borrowers' ability to service their mortgages. In contrast, the main impact of price declines is to reduce borrowers' incentives to service their mortgages. Which of these two reasons is more important?

Let's consider two scenarios. In scenario 1, the borrower loses his job but still has positive equity in his home. In this case, he may be unable to continue paying his mortgage, but he can preserve his equity (and his credit rating) simply by selling his home. Default is therefore unlikely.

In scenario 2, the borrower keeps his job but the market value of his home falls below the outstanding balance on his mortgage. In common parlance, the borrower is 'underwater' on his mortgage. If he falls far enough underwater, he may make a purely financial decision to walk away from his home, rather than continue paying his mortgage. Default therefore becomes increasingly likely as the value of the home falls, even though the borrower’s ability to service his mortgage remains unchanged.

The Great Recession has brought job losses as well as home price declines. To some extent, the two effects have fed on each other: Job losses have resulted in lower demand for housing, which has led to price declines; and price declines have resulted in lower housing construction, which has led to job losses. Nonetheless, the scenarios considered above suggest that an underwater borrower is a bigger default risk than a borrower who has simply lost his job. This is borne out by recent default activity in the Bay Area.

Consider the chart, below, which compares default rates for the first quarter of 2009 (vertical axis) to the percentage of mortgages that were underwater in November of 2008 (horizontal axis). Each dot represents one of the nine Bay Area counties.

The relationship between the two variables is clearly strong, and seems to indicate that being underwater is an excellent predictor of the likelihood that a borrower will default in the near term. (Note: Information on default activity was obtained from Dataquick. Information on underwater mortgages was obtained from Zillow, by way of the San Francisco Chronicle.)

Since I don’t have ongoing access to data about underwater mortgages, I thought it would be useful to create an alternative version of the chart above, using recent price declines as a proxy for the percentage of mortgages that are underwater.

The vertical axis shows the Q1 2009 default rate for each county, as before. The horizontal axis shows a measure of the recent price change for each county. (Using the average price for the three-year period from July, 2005 to June, 2008 as a base, I calculated the percentage change in price through the end of 2008. I reversed the axis to facilitate comparison with the first chart, above.)

The relationship isn’t as strong as before, but still seems compelling. In a future blog posting, I’ll develop a better proxy for the percentage of mortgages that are underwater. For now, I’ll point out that prices have continued to fall in every county except Marin (where they’ve risen by a meager 3% since the end of the year) and San Francisco (where they’ve remained flat). The wave of foreclosures in the Bay Area therefore seems likely to continue building for some time to come.