A healthy housing market is important to the well-being of the US economy. Unfortunately, the market doesn't appear to be turning around just yet, with housing starts dropping to an average annual rate of 1.02 million over the past three months from 2.07 million in 2005. The S&P/Case-Shiller index of 20 major metropolitan areas shows that home prices have fallen 15.8% from a year ago and are 18.4% below their July 2006 peak. By historical standards, the drop in housing starts and sales is still about average; in the average post-war recession, housing starts have plunged 49% from peak to trough. However, the decline in the median home price is the largest since the 1930s and far greater than in any other post-war downturn.
So, is the market bottoming out? It seems to be nearing a bottom, at least in terms of sales and starts, but we thought we were seeing that last year as well. Seasonal factors are volatile, making month-to-month movements difficult to interpret. Current data suggest that starts are levelling out near 1 million, with home sales (new and existing) near 5 million. The inventory of unsold new homes has come down to 426,000 from 543,000 last June. We don't think we have hit bottom yet, but we expect to do so in the second half of the year. Even so, this year will be the worst for housing starts since World War II.
So far, housing sales and starts have done a bit better than we feared, while prices have dropped more than we expected. In a perverse way, this could be good news, suggesting that homeowners and builders are accepting the lower prices more quickly than we had feared. However, it could also just be a reflection of the very low down-payments with which buyers acquired many of these houses, now pushing more homeowners into foreclosure or forced sales.
No sale
Home sales have been dropping from their 2005 peak. In July, existing home sales were at 5.00 million, down 13% from July 2007 and 29% below the record 7.08 million sales in 2005. New home sales have been hit even harder, plunging to a 17-year low of 513,000 (annual rate) in March, before edging up to 515,000 in July.
The reason for the decline is that homes became too expensive. To the average buyer, the price of a home is effectively the size of the monthly mortgage payment. As mortgage rates rose with the Federal Reserve rate hikes from 2004 to 2006, the effective cost of a house increased. The change pushed the National Association of Realtors' affordability index, which is based on the monthly income required to qualify to buy the median existing home with a conventional mortgage, down to 102.7 in the second quarter of 2006 from 136.5 in the first quarter of 2003. (The higher the number, the more affordable the average home is.)
With home prices now falling and interest rates lower, affordability improved to 132.4 in the first quarter of 2008, which brings it nearly back to its 2003 peak. However, the inventory of unsold homes on the market (an 11.1-month supply of existing homes) and loss of confidence among potential buyers are keeping sales down. Potential purchasers are afraid they are buying a depreciating asset, so to get ready to buy they need assurance that prices have finished falling.
We expect the Federal Reserve to begin raising rates next spring. However, just as there has been little impact on long-term bond yields (and thus fixed-rate mortgages) since the Fed began to lower rates in mid-2007, the rate hikes will also have little impact. Adjustable-rate mortgages will become more expensive, but borrowing has shifted to fixed-rate mortgages in any event.
House prices continue to drop
The current drop in the median home price is unique. The price dropped 2.9% in 2007, marking the first annual decline in the history of the series (since 1966). We expect a total decline of 19% by early 2010.
The decline is bringing home prices back in line with incomes. The ratio of the average home price to average after-tax household income hit a record 3.4 in the third quarter of 2005, but the lower house prices and higher incomes brought it back to its historical average of 2.7 (1960-2007) as of the first quarter of 2008. We had originally expected the ratio to remain near its historical average because mortgage rates are below their historical average, but we now think that prices will overreact on the downside, resulting in the ratio sliding to 2.2 by late 2010 - the lowest since 1975.
But traditional financing (20% down and a 28% ratio of debt service to income) is at least available and helps support moderate-priced housing. Harder-to-secure jumbo loan financing has hurt the sale of high-end homes more, but jumbo loans are still available; they're just more expensive.
A better measure of home prices is the S&P/Case-Shiller home price index, which looks at resales of the same houses and thus is less subject to changes in the mix of sales. While the median home price is down 7.1% from a year ago (July), the S&P/Case-Shiller index is down 15.9% (June data). A weakness of the S&P/Case-Shiller index is that it covers only the 20 largest metropolitan areas. However, it is an index that is tradable in the Chicago market.
The other often-quoted measure of home prices is the House Price Index from the Office of Federal Housing Enterprise Oversight (OFHEO). This index is similar to the S&P/Case-Shiller index, but covers all of the US. However, it only tracks homes purchased with mortgages obtained through Freddie Mac or Fannie Mae, which excludes both subprime and jumbo loans. This index has dropped only 4.8% from a year ago, based on second-quarter data.
OFHEO has studied the difference between the two series at length (see "Revisiting the Differences Between the OFHEO and S&P/Case-Shiller House Price Indices," published in January 2008 at www.ofheo.gov/media/research/OFHEOSPCS12008.pdf). On average, the OFHEO index moves about half as much as the S&P/Case-Shiller ù both up and down. The differences by metropolitan area are largely the result of the inclusion of high- and low-end properties in the S&P/Case-Shiller index, with the low end accounting for most of the difference during the study period and the lower weights the S&P/Case-Shiller index assigned to sales with long periods in between. (The largest difference originally was the inclusion of refinancing appraisals in the OFHEO index, but OFHEO has since emphasised its purchase-only index.) The coverage is also a factor because prices in the major metro areas move more than prices in smaller cities, and the S&P/Case-Shiller cities are somewhat concentrated in the South and West, where prices were most bubbly.
Although the biggest declines in home prices are in the regions that had the biggest run-ups and highest ratios of home prices to income, the relationship is not absolute. The biggest drop in home prices in 2007 was in Michigan, where home-price appreciation had been weak and the ratio of home price to average income was among the lowest in the country. But jobs in this region have been disappearing because of the problems in the auto industry, and home prices have dropped accordingly. On the other hand, Seattle and Portland have high ratios of home price to income, but home prices have been stable there because of strong job growth and in-migration.
Impact on mortgage holders
On average, households appear to be in good shape. Although household wealth remains well below its 1999 record ratio of 615% of after-tax income, at 533% in the first quarter 2008 it is still well above its 1959-2005 average of 480%. Debt-service payments are 14.1% of household income, down from 14.5% in mid-2006, though still very high by historical standards. Total financial obligations, including debt service plus rents, leases, property taxes, and insurance payments, are at 19.2% of income. This is near the record high of 19.5% set in late 2006, although only slightly above the 18.3% average of the last decade. Moreover, homeowners have a financial obligations ratio of only 17.8% compared with 26.0% for renters. Owning a home still beats paying rent in the long run, though not by as much as it did five years ago.
However, some households are going to be in trouble. The 2004 Survey of Consumer Finances showed that 12.2% of US households (up from 11.8% in 2001) have debt-service burdens that exceed 40% of their income. The share is even higher in major cities. We suspect that this category also has a higher share of adjustable-rate mortgages.
Impact on the economy
Housing is the major factor slowing economic growth in the US. We expect housing to subtract another percentage point from growth in 2008, about the same as it did in 2007.
The indirect impact of housing on the economy so far appears small. Consumers have not backed away from spending and the personal savings rate remains near zero (0.6% in 2007). The stimulus package has distorted recent data. One of the biggest questions, however, is whether the higher cost of borrowing and the slower rise in the value of home equity will curtail borrowing.
Americans have been using their homes as ATM machines through home equity loans and cash-out refinancings. Last year, homeowners took $364 billion (3.8% of disposable income) out of their homes, down from $640 billion in 2006. Low interest rates made these loans cheap, especially because they tended to be tax deductible.
So far, this activity doesn't seem to be slowing down very much. Refinancing activity remains high, though some of this is probably refinancing adjustable-rate loans into fixed-rate ones as mortgage holders got scared about rising monthly payments. Americans still have a lot of untapped home equity, but the average loan-to-value ratio in the US housing market has risen to 54% in the first quarter of 2008 compared with 43% at the end of 2001. This reflects both higher borrowing and the drop in home prices.
The higher interest rates and reduced willingness of banks to provide second mortgages and home equity loans will cut down on real estate-backed borrowing and, eventually, spending. Lower car sales have reduced auto loans as well. Credit card activity has accelerated, but only modestly, suggesting that Americans are becoming a bit more cautious about using their credit (or perhaps banks are becoming more cautious about granting it). We expect this to slow consumer spending once Americans have spent their tax rebates.
It could be worse
Our baseline forecast includes a further drop of 10% in the S&P/Case-Shiller home price index over the next year, bringing the total decline to 29%. Other prices will decline proportionately, with the average home price dropping 16% from the second quarter of 2006 to the first quarter of 2010. Along with income growth, this decline brings the ratio of home prices to income down to 2.2 by 2010, well below its long-term average of 2.7 and the lowest level in 35 years. The correction could be greater, especially if mortgages become even more difficult to get and the economy falls into a deeper recession. In our pessimistic scenario (See "US Risks To The Forecast: How Far Down?" published on July 25, 2008, on Ratings Direct, S&P's real-time, web-based source for credit ratings, research, and risk analysis), the average home price drops a total of 29% (more than 40% on an S&P/Case-Shiller basis).
In our alternative projection, we also assume bond yields will rise sharply, increasing the mortgage rate to 9% by mid 2010. Home sales and prices will drop and housing starts will fall to a post-war low of 497,000 in the first quarter of 2009.
We intend this projection to be a worst case scenario and we believe it has about a 20% probability of occurring. It if does, the correction will be severe, and in fact, it would be unprecedented for prices and the 60-year low for housing sales and starts.
David Wyss is the chief economist at Standard & Poor's in New York.
¬ Haymarket Media Limited. All rights reserved.
