by Calculated Risk on 7/20/2007 04:20:00 PM

This is a preliminary update to my 2007 housing forecast (see bottom of post for graphs on inventory).

Let's review the basics of supply and demand.

This diagram shows a normal Supply and Demand relationship. When the supply curve shifts (dark blue to light blue) then the price falls from P0 to P1.

And when the demand curve shifts, perhaps due to the changes in lending standards (from dark red to light red), the price falls again, this time from P1 to P2.

So, given a normal market, with the given shifts in supply and demand, we would expect prices to fall, keeping the quantity demanded in balance with the quantity supplied.

The above diagram works well for commodities, like corn, and these facts - shifts in supply and demand - would lead to falling prices, bringing the quantity demanded and the quantity supplied back into equilibrium. But, for housing, prices are sticky because sellers tend to want a price close to recent sales in their neighborhood, and buyers, sensing prices are declining, will wait for even lower prices.

With perfect information, sellers would reduce their prices to P2, and the markets would clear. However, with imperfect information and sticky prices, we usually see a precipitous decline in the quantity demanded instead.

Look at the diagram and imagine that the price stays at P0. What happens? The quantity demanded is where the light red line and P0 cross. Meanwhile the quantity supplied is determined by where the light blue line would cross P0 (imagine the P0 line extends to the light blue line). This implies transaction volumes would decline and inventory levels would rise - exactly what is happening right now.

The Supply and Demand Curves shifts

This post would get very long if I tried to explain why and when the demand and supply curves shifted. Obviously the demand curve has shifted to the left recently due to the tightening of lending standards. But the first shift in the demand curve actually occurred towards the end of 2005 - because so many buyers were "speculating" with leverage and pulling demand forward during the boom. (Not to be confused with speculators buying investment homes).

As a reminder of the buying frenzy back in April, 2005, in an LA Times article "They're In — but Not Home Free", the writer described a woman that was "able to afford, barely, her first home". She had taken out "an adjustable-rate mortgage that won't require her to pay any principal for three years". She was already strapped, working overtime to pay her bills, and didn't know what she would do in early 2008. She was gambling that either her income would increase or that the value of her home would rise enough to sell at a profit. From the article:

Californians are adopting a "buy now, pay later" strategy on a massive scale. The boom in interest-only loans — nearly half the state's home buyers used them last year, up from virtually none in 2001— is the engine behind California's surging home prices.

This type of leveraged activity pulled demand from future periods (i.e. now) shifting the demand curve to the right during the boom, and has shifted the demand curve to the left during the bust.

The supply curve was shifted to the right during the bust for at least two reasons: 1) investors were buying homes in record numbers to flip during the boom. This can be described as "storage", reducing supply during the boom, and increasing supply during the bust as investors try to unload their properties, and 2) rising foreclosures also increases the supply during the bust.

I'll update the 2007 housing forecast next week, after the June New and Existing home sales reports are released. Meanwhile, here is a look at inventories through May:

Inventories are at record levels and rising.

This graph shows the year end total houses for sale since 1982 (2007 inventory is for May). Existing home inventories are currently at a record 4.431 million units, and new home inventories of 0.536 million units are near the record.

And the situation appears to be getting worse. Based on ZipRealty data, existing home inventories were probably over 4.5 million units in June. And since cancellations for new home builders are increasing - for examples, see here and here - the number of new homes for sale is probably understated (note: see Caroline Baum on how cancellations impact new home inventory levels).

In the second graph, inventory levels are normalized by the number of owner occupied units (OOU). Normalizing by OOU isn't perfect, since it doesn't account for changes in 2nd home ownership, but it does show that inventories are at record levels even compared to the growth in homeownership.

The third graph uses the "months of supply" metric for both new and existing homes. The "months of supply" metric is now above the level of the previous housing slump in the early '90s, but still below the peak levels of the housing bust in the early '80s.

The "months of supply" is calculated by dividing the total inventory by the seasonally adjusted annual rate (SAAR) of sales, and multiplying by 12. Currently inventory is 4.431 million, SAAR sales are 5.99 million giving 8.9 months of supply.

Both the numerator and the denominator are moving in the wrong direction. Not only is inventory at record levels and increasing, but sales are falling ...

Demand is falling.

It is difficult to graph "demand", but we should be aware of the incredible increase in turnover of existing homes during the recent housing boom (due to excessive leverage as described above). This graph shows existing and new home sales normalized by the number of owner occupied units. Existing home sales reached 9.5% of owner occupied units in 2005, far above the median level of 6.0% for the last 35 years.

Some of the sales were for investment and second homes, but normalizing by owner occupied units probably provides a good estimate of normal turnover. If existing home sales fell back to 6% that would be about 4.6 million units. If sales fell back to the level of 1998 to 2001 (7.3% of total owner occupied units sold) that would be about 5.6 million units in 2007.

This graph also shows that new home sales, as a percent of owner occupied units, has already fallen back to the median level of the last 35 years. However existing homes are a competing product for new homes, and the record inventory of existing homes for sale will probably continue to put downward pressure on new home sales.

And finally a word on prices: Sticky doesn't mean stuck. Prices are now falling by most measures, and will probably continue to fall slowly over the next several years. While we wait for prices to fall to P2, it is easy to predict that transaction volumes will decrease, but it is difficult to predict by how much. The housing market moves in slow motion, so I think my sales prediction of between 5.6 and 5.8 million existing homes in 2007 still looks pretty good. Sorry for the long post!