Speculative Logic Improved

By the time any real estate market reaches its seventh consecutive good year, it's wise to check that itchy spot at the nape of the neck, the one that warns of overconfidence (and toothy critters gaining ground, unseen).

There are many measures of real estate market health, but most are defective. Rising home prices are an effect of a healthy market, not a cause. If prices rise faster than local incomes, it's a bad sign, not a good one. Home sales by themselves don't tell you much, prone as they are to sudden leaps up and off sequential cliffs. Job creation statistics are nice, but most jobs are created by people who want to work, not as gratuities to supplicants. A wave of new construction can presage an overbuilt, pre-crash market as easily as a healthy one.

One good measure of real estate market health -- especially in Colorado -- is the relationship between building permits and net population migration. A history of the permit/migration relationship since the early '80's shows the consequences of the old-style boom/bust cycle, but also shows a change to reassuringly prudent practices by today's developers.

1983 1984 1988 1989 1993 1997

Permits 32,207 27,262 6,853 5,762 15,759 23,320

Net Migration 37,111 3,263 (24,046) (18,698) 72,667 76,465

Sales 26,000 22,000 23,460 25,289 38,598 35,963

The permit/migration relationship in 1983 and 1984 reveals the old-style danger in a hot market. At the peak, anybody who could drive a nail had a "spec" (short for speculative, meaning no buyer) home underway. Migration was strong in 1983, and supported the huge number of permits; but in 1984, builders built in anticipation of continuing demand, and nobody came to the party. From 1985 through 1990, migration to Colorado was net-negative by 65,000 people.

Similar permit/migration boom/bust relationships can be found in each post-war cycle, with permit peaks in 1955, 1961, 1972, and 1977 followed by migration dips in 1957-58, 1964-66, 1974-76, and 1980.

Sales figures for the seven Metro counties are misleading, as they seem to maintain a fairly steady pace. However, maybe a third of 1988 and 1989 sales were distressed sales or remarketed foreclosures. Conversely, sales in boom-year 1993 would have been higher had there been more homes to sell.

Wisdom acquired through suffering in the 1980's together with the demise of foolhardy S&Ls, and tightened rules at banks have transformed the boom cycle. Today, in a hot market, there is negligible "standing inventory" of completed but unsold homes in Front Range markets; and for that matter, damn few spec homes started at all.

Builders have learned. And how: there were only two-thirds as many permits issued in 1997 as in 1983, though the State population is almost 25% bigger, sales are 50% higher, and migration is double the 1983 pace. It's unusual to find the words "developer" and "prudent" in the same sentence, but there they are.

There is a tremendous inventory of land, subdivided and ready to go (maybe 10,000 lots in Boulder County alone), but land is much cheaper to carry through a down cycle than finished homes. A decline in migration could soften our market at any time, but with no huge inventory of overbuilt homes to work off, any sales trough is likely to be much shallower and shorter than the 1980's.

If you think this 1998 market is too hot, check out the all-time record year for building permits. In 1972, 45,063 permits were pulled (double the 1997 total), and Colorado's population grew by 111,000 people -- almost five percent in a single year. By comparison, the 1990's are on a gentle, sustainable cruise.

(Notes to data: The State Office of Demographics produces reliable migration and population numbers back to 1950, and the Home Builders Association of Denver keeps a good permit series for seven Metro counties going all the way back to 1955. However, reliable Metro sales data are unobtainable before the 1985 start of Perry & Butler's ongoing study, and the Genesis Group's comprehensive series beginning in 1986. 1983 and 1984 sales figures here are educated guesses.)