Reading the Tides

Whoever said a rising tide lifts all boats didn't know boats. All boats, like all building companies, fare very differently in the same conditions. No two voyages are exactly the same. Even with the best captain and an experienced crew, navigating rough seas and storms takes a blend of technical know-how and sheer brass.
By Bob Sperber, Senior Editor | December 31, 2005


Whoever said a rising tide lifts all boats didn't know boats.

All boats, like all building companies, fare very differently in the same conditions. No two voyages are exactly the same. Even with the best captain and an experienced crew, navigating rough seas and storms takes a blend of technical know-how and sheer brass. But with the right team and some discipline, a well-maintained cruising dinghy can overtake the sleekest high-tech racer with a mediocre crew.

No one can control nature or the economy. At times, environmental elements are too great for any captain, or captain of industry. But a good sailor can read the tide charts before setting out, adjust for changing winds and pick the wisest course. For starters, a builder can run a tight ship with a good internal budget and efficient processes. But the minute he thinks about moving dirt, the business becomes part of a larger ecosystem.

Therefore, builders need to understand the economic indicators that buffet the industry. Knowing which indicators matter and why can help a builder sail ahead on sunny days, and know what course to chart when the boom comes down.

Below are the most important indicators for a builder to watch:

Interest Rates: 30-yr fixed; 1-yr ARM, 10-year T-bill set by the Federal Reserve Bank and reported monthly.

Employment: Total change in U.S. employment reported monthly by the Bureau of Labor Statistics (BLS).

Permits: Total U.S. permits and regional permits for new home construction in both single family and multifamily. Reported monthly.

Starts: Total U.S. housing starts reported monthly by the U.S. Census Bureau.

New-Home Sales: Sales of new homes reported by U.S. Census Bureau monthly.

Existing Home Sales: National Association of Realtors reports existing home sales monthly.

New Single-Family Home Prices: Prices for new single-family homes reported by Census Bureau on a monthly basis.

Existing Single-Family Home Prices: National Association of Realtors reports prices on single-family homes.

Inventory: Census Bureau reports available inventory of new homes and the National Association of Realtors reports existing homes. Reports are available monthly.

Materials Prices: Producer Price Index (PPI) for Residential and/or single-family construction reported monthly by the Bureau of Labor Statistics.

Consumer Confidence: Michigan Consumer Sentiment Index from the University of Michigan, reported monthly. Also, the Conference Board's Consumer Confidence Index, reported monthly.

NAHB/Wells Fargo Housing Market Index: The Housing Market Index (HMI) is a composite index that measure builder confidence and reported monthly.

Interest Rates

When it comes to housing market forecasts, mortgage rates are "clearly the biggest single factor," Michael Carliner, economist with NAHB, says. Add interest rates, widely recognized as the most important variable to Wall Street housing industry watchers for a complete and very leading indicator for builders.

Builders can easily track 30-year fixed and adjustable-rate mortgages with the Weekly Primary Mortgage Market Survey from Freddie Mac. These can be viewed against the open-market long-term rates, otherwise known as 10-year Treasury bills, which lead those mortgage rates. Viewing mortgage rates against the backdrop of long-term interest rates — and keeping an ear on news announcements from the Federal Reserve Bank — can help builders predict demand patterns months, a year or more in advance.

Some economists follow the yield curve, which is the spread between long and short rates. But rates that are easily tracked individually, so this statistical abstraction is a bit removed for builder's needs.

Pros: The "most important" and leading indicator of housing demand.

Cons: Short-term trends can mislead.

Employment Growth

With some caveats, an increase in jobs produces an increase in households and therefore housing demand. The BLS Current Employment Statistics (CES) program is the benchmark source. The most comprehensive number for viewing job market health is the total change in U.S. employment, that is, the number of jobs lost or gained each month.

The CES surveys 160,000 businesses drawn from a sample of 400,000 in 850 non-farm industries. Monthly data include total employment, number of women employed, number of production or non-supervisory workers, average hourly earnings, average weekly hours, average weekly earnings, and average weekly overtime hours in manufacturing industries. The survey covers production workers in natural resources, mining, manufacturing, and construction, as well as non-supervisory workers in the service industries. It does not cover smaller groups including the self-employed and one big group: military personnel, who builders can, however, factor-in based on military base locations.

BLS also reports initial unemployment insurance claims and the unemployment rate. The first is an accurate tally from BLS. The second is less trusted by the experts; it's an extrapolation based on 2000 Census data supplemented by a relatively small, monthly Current Population Survey of households by the Bureau. Generally, housing economists prefer to track employment — potential buyers — than to follow non-prospects.

But in the current economy, experts are uncertain about recent events. Since 1970, employment and housing starts dipped whenever employment fell. Then, conventional wisdom was shattered as employment track flat-to-negative while housing boomed from 2001 to 2003.

"If builders based all their plans on employment during those boom years," Carliner says, "they wouldn't have built anything... Population would be a better indicator of housing demand." The Census Bureau's Current Population Survey provides monthly updates of household and employment data. But this survey is relatively small and highly extrapolated from a larger, annual survey. "And none of it is timely or directly observable," Carliner says. "So we use employment — because it's there. It isn't perfect but it's the best combination of a statistic being important and available."

The BLS data are trusted at the national and major-regional level but are "spotty" data below the major metropolitan level, Carliner says. For example, it's hard to trust part-time vs. full-time job counts. Also, when viewing monthly levels, builders should avoid short-term conclusions. Inclement weather, for example, can keep outdoor workers off the job, but their hours will bounce back the next week or month. Jobs also differ in their quality and effect on local economies.

"Some jobs create other jobs, and some don't," Jim Haughey, director of Economics for Reed Business Information, says. (Disclosure: RBI publishes Professional Builder.) A military base, factory or large financial employer is more likely to generate more jobs to serve its employees. "These types of jobs are a lot more useful to builders than a situation where the only jobs being created are at the local pizza parlor, the grocery store or a strip mall."

There's one more related threat: Since 2000, 38 percent are in real estate. This could represent an artificially strong job economy. It looks good, but it "may not be all that all that rosy when you look at what's driving it," Justin Speer, researcher with Credit Suisse First Boston, says.

Pros: Without timely and accurate population demographics, it's the best and most leading indicator to forecast demand.

Cons: Trust issues. Counter to convention, employment fell 2001 to 2003 but housing boomed, confounding all predictions.

Permit Approvals

Housing permits are a good short-term leading indicator, weeks if not months ahead of starts and sales. It's useful to track the total number of permits, seasonally adjusted, for single-family units. Data are also available for two-to-four unit buildings and five-plus units. Single-family trending provides a useful trend, but builders can adjust this to their needs.

For a major portion of houses, construction begins the month of permit issuance. The remainder of homes breaks ground within three months, Census reports.

Permits are tracked in Census' monthly New Residential Construction release. This includes permits by building type as well as starts and completions. Data are arranged nationally; by the four major Census regions (Northeast, Midwest, South and West); and by state, metro and county levels. (Or county-like groupings in New England.)

The number of permits is based on the number of housing units jurisdictions authorize. Permit data are collected from individual municipalities where available, or else counties, townships, or New England and Mid-Atlantic towns. As builders know, permits are on the public record, which makes them a highly accurate national as well as local indicator.

This statistic is based on a monthly survey of approximately 9,000 of the 20,000 permit issuing places in the United States. (The rest are surveyed annually.) When one of the jurisdictions doesn't reply to the survey, the missing information is filled-in using a separate survey of 900 places Census field representatives visit, or is statistically estimated using year-to-year trends.

A longer-term view may also be helpful for understanding industry dynamics. For example, new home sales hit a record in 2004 with 1.2 million single-family sales. But demand hadn't peaked; 1.6 million permits were approved for single-family units the same year. "These data underscore our view that builders are being more restrained in their business practices, focusing on 'build to order' rather than 'build to inventory,'" according to a UBS Investment research report that noted a trend over the last decade. The report added that the ratio of permits issued to new homes sold declined from 37 percent to 25 percent. This, UBS took to mean that there are fewer speculative permits being sought today. Whatever else this means, it solidifies permits as a leading indicator of sales.

Permit statistics only represent housing in places where permits are required, which excludes a Census-estimated three percent of jurisdictions. Up to two percent of permits never result in construction. At the local level, permits can come in bunches, so builders should be cautious of short-term trend assumptions.

Permits are "probably are not quite as reliable as starts, nationally, but are the best way to get more regional detail," says RBI's Haughey. After several months lag, permits are available across the nation by zip code.

Pros: Permits provide an accurate, leading look at pending starts and sales, and help gauge near-future housing supply.

Cons: Permits are still permits; total industry figures for starts are best for national trends.

Housing Starts

CSFB calls starts "critical to the housing market, since they are a leading indicator of sales and earnings." For the public home builders it tracks, improving starts are likely to be reflected in home-builders' performance three to six months later. While most builders are more concerned with Main Street than with Wall Street, industry performance measured in starts gives an accurate read on industry momentum.

Seasonally adjusted housing starts for single- and multi-family units, again, are derived from Census' monthly New Residential Construction report. Starts are measured in all jurisdictions, including those that don't issue permits. And while permits generally predate starts, starts are valued indicator that helps predict housing coming online in the coming months.

Census field personnel check actual sites to fill-out surveys but data rely on private developer/builder response. While national estimates are trusted, Census admits it "does not have a large enough sample size to make state or local area estimates." Therefore, aside from permits, all other statistics in the New Residential Construction release are tabulated no further down than the four major Census regions.

Pros: Starts are a forward-looking measure of construction actually underway, nationally and in four major regions.

Cons: Starts data are not as reliable an indicator as permits below the four major regions. Data do not reflect the value of the houses.

New-Home Sales

The Census covers sales for new single-family units. This is based on the Bureau's monthly New Residential Sales report, which is "probably one of the better and more timely and accurate statistics that the government collects," Carliner says.

Here, a sale is based on the "the signing of a sales contract or the acceptance of a deposit," whether the house has yet to start construction, is under construction or is complete. About one-quarter of houses are already completed. The rest, Census reports, "are evenly split between those not yet started and those under construction." The report is national and by the four major Census regions. Closings typically lag by one to two months.

For permits, sales and other reports, Census defines a single-family home as a detached house as well as attached units with floor-to-ceiling construction and other criteria generally describing a townhouse — as opposed to stacked multi-family units. The single-family definition excludes homes built for rent, homes built by the owner or on the owner's land... and manufactured homes.

Pros: An accurate estimation of near-current housing industry health.

Cons: Sales are recorded at varying, unknown stages of construction. Local data are unavailable.

Existing Home Sales

Along with new home sales, sales of existing homes are an important, if trailing, gauge of housing-market health. The National Association of Realtors releases its Existing Home Sales (EHS) report monthly. Unlike the Census' definition of a sale as a signed contract, NAR tallies sales based on closings. But the survey sample is large. NAR surveys 160 of 700 local partners to get data on 30 to 40 percent of existing home sales.

The final report estimates the total number of closings nationwide and for the four main census regions. Homes are separated by single-family, condos and co-op.

There are a couple of options for attempting to match Census and NAR sales volume for a given month in the real world. One is to compare the most recent NAR sales numbers to Census new-home sales data from a month or two ago. This discrepancy in definitions led NAR to come with another alternative that serves as a partial solution. In January 2005, NAR began its Pending Home Sales report based on the date of contract signing, which makes for a roughly apples-to-apples comparison with the Census' report. Also like Census, this Index reports trends nationally and in the four major Census regions. The downside: Monthly estimates are expressed as an index (with 2001 set to equal 100); NAR does not release actual volume of units.

While this falls short of a direct comparison, the key issue remains trending, not number tracking. "The number itself does not really indicate too much," NAR economist Lawrence Yun says. Because what matters are the trends, or "whether conditions are improving or declining, and how strongly," Yun says.

Pros: A definitive if lagging view of existing home sales at closing and a new at-contract index to match Census' new-home data.

Cons: NAR doesn't release sales volume numbers for its Index, so there's no instant match-up with new home sales data.

New Single-Family Home Prices

It's useful for builders to track home prices, not just for the number but the trend. For example, a surge in prices beyond the population's norm could signal an undersupply or other issues related to affordability.

Therefore, builders should view price trends alongside sales. Census' monthly New Residential Sales release gives median and average sales prices of new single-family homes, again at the time of deposit or sales contract signing. The median price is more accurate for trending because it represents the midpoint; half the houses sold have a greater price and half have a lesser price. As with other data in this Census report, tabulations are national and for the four Census regions; the sample size isn't sufficient for state and local estimates.

Pros: Prices are a good general indicator relating to affordability and market opportunities.

Cons: State and local prices aren't reported.

Existing Single-Family Home Prices

Like Census, NAR reports median or midpoint prices each month as part of its Existing Home Sales release. Between mean and average, builders should track median prices. Data are reported nationally and the four census regions each month.

This indicator isn't leading but still is a necessary companion to new-home prices because trends can develop. For example, Carliner has noticed that existing home prices have been rising faster than new-home prices. "Is it because we sold more large vs. small houses, or more houses in expensive areas vs. less expensive areas?" He's not sure because "it's not easy to adjust for a perfect apples-to-apples comparison."

In response to this difficulty, NAHB and others routinely check the House Price Index from OFHEO, the Office of Federal Housing Enterprise Oversight.

"We prefer OFHEO because it's apples to apples, as opposed to looking at different sets of inputs," says Ivy Zelman, analyst in charge of CSFB's home building group. But this is a quarterly index. She finds it more reliable for tracking changes in home value, by nine Census regional divisions, and "many" metropolitan areas. RBI's Haughey believes the quarterly lag make it useful to still view Census and NAR data by the month.

Pros: NAR provides timely data on home prices for a majority of homes sold.

Cons: Monthly comparisons with new-home prices are difficult. OFHEO provides a quarterly analysis.

Housing Inventory

The total inventory of completed unsold homes and a monthly supply of unsold homes data provide a good gauge of housing supply. It's not a leading indicator but it's important. Too many completed, unsold homes in the ground can reflect poor sales, poor construction planning or too much of a debt burden to builders. Or, a bump in employment can shrink inventories and indicate a rise in demand.

Census' New Residential Sales provides both an absolute total as well as an estimate of how many months' supply of homes this represents. This months-of-supply number is based on an annually adjusted sales rate. It uses the ratio of houses for sale to houses sold. NAR's EHS provides a similar measure for existing homes.

"Both inventories of new and existing homes are very important — one by itself is not enough," says CSFB's Zelman. "But we think the absolute number is more useful because the month's supply is dependent on sales. So if sales are slow one month, obviously, what you thought was a low level of supply can turn out to be a big month's supply."

"Inventory changes as a consequence of everyone's marketing mistakes," says Haughey. "If builders built too much and over-estimated what they could sell, it indicates what they did wrong the prior month. So while inventory is a lagging indicator, it can be used to help builders know what to do the next month."

Locally, builders can use the Realtors' Multiple Listing Service to calculate inventory, price and other factors.

Pros: A strong concurrent-to-lagging indicator of housing supply and future needs with analytical value.

Cons: More Census apples and Realtors' oranges in trying to mesh new and existing inventories.

Materials Prices

To the extent that the price of materials affects building costs — about 30 percent — this is a leading indicator. Top management would be wise to track material costs, specifically one of the measures specifically suited to their type of residential building. The most comprehensive and efficient way to do this is through an appropriate, industry-specific Producers Price Index (PPI).

Unlike the Consumer Price Index associated with predicting inflation, PPIs reflect wholesale prices before added subsidies, taxes and distribution costs. The Bureau of Labor Statistics tracks 8,000 PPIs based on more than 100,000 individual prices that span the economy. For these indexes, 100 is the baseline, tied to prices in 1986.

Within these massive digital catacombs, are industry-specific indexes of material and supply inputs to construction industries including. There are specific indexes for all construction, new construction, residential construction, single-unit construction and multi-unit construction. The "inputs" are similar for all residential indexes, but differ slightly due to weighting. For example, the residential construction index reflects less steel and concrete than the broader new-construction index, which includes non-residential construction. For most builders, the single-unit residential index most reflects their business because it includes single-family houses and floor-to-roof units (townhouses), but excludes apartment buildings with stacked units.

"There's a little bit of many other indexes that go into making up each one of these," says BLS economist Antonio Lombardozzi, "and there's nothing publicly available to know exactly what goes into it." For instance, the residential PPI has "at least 150 PPIs of material and supply inputs to the construction industries in it."

PPI data lag actual prices, because the data, issued roughly mid-month, reflect the prior month's prices. "But it's all relative. You have to define your references," RBI's Haughey says. "For a home builder, if material prices go up it can have a negative effect on pricing and of course starts and sales." Additionally, rising wholesale prices in the PPI tend to predict the CPI and can affect consumer sentiment "because people don't react for a month or two after they see prices move up or down," Haughey says.

This BLS information is practical and comprehensive for top-level viewing and detection of trends. Further investigation, builders can look to individual PPIs for items such as gypsum, cement, plumbing materials, and many kinds lumber. Private sources such as Crow's and Random Lengths provide additional depth in lumber analysis.

Pros: PPI material data are solid economic indicators of future home pricing and/or demand.

Cons: For day-to-day purchasing, seek out more timely commodity-specific services and markets.

Consumer Confidence/Sentiment

Builders value consumer perceptions. Two competing and roughly interchangeable indexes from two competing sources — the University of Michigan and the Conference Board — survey consumer attitudes that correlate with economic ups and downs.

"At the end of the day, for our business, the most important things to know are where interest rates are going and the general psyche of the market at any one time," says Joe Riggs says, group president over roughly 20 percent of K. Hovnanian's $4-plus billion home building business. "The better people feel about their situation, the more likely they are to make a large-scale purchase, and the better it is for our business."

The Michigan Consumer Sentiment Index and Conference Board's Consumer Confidence Index both measure consumers' economic optimism or pessimism with slight differences. For example, Michigan asks consumers how they feel about their own situation, while the Conference Board's questions are more aimed at how people feel about the overall economy. Both ask for perceptions of current and future expectations.

"Watch especially the expectation component," Ken Goldstein, Conference Board economist, says. "Because if people think things are going to sour, they won't take on a mortgage." He advises those viewing the survey to pay more attention to sustained up-or-down trends than arcane survey intricacies.

Both survey sources offer, and sell, deeper industry-specific data. For example, in October, Michigan reported, that "favorable home buying attitudes fell to their lowest level since 1990, with more consumers complaining about high home prices than any time since 1982."

Choosing between the two sources, CSFB's Zelman slightly prefers Michigan but doesn't "necessarily think one is better."

"We have never found them to be very good predictors of what's going to happen," says Carliner. "If you take a look at the things that might influence and help to form consumer confidence, like unemployment rates, income and interest rates, you can look at directly and not wait for a survey. Often they're just measures of what's already happened."

Pros: Consumer attitudes toward the economy correlate with interest rates and employment.

Cons: It's not leading; even future expectations measures rely on past conditions. Other indicators are more immediate.

NAHB/Wells Fargo Housing Market Index

The NAHB/Wells Fargo Housing Market Index of builders' expectations has a "fair amount of predictive power that reaches about a quarter into the future," Carliner says. A recent Federal Reserve study found it more useful an indicator of housing strength, and economists across the housing economy refer to it. Suppliers to the industry who hope to gauge builder demand also use it. But it seems to have at least as much value as consumer readings. For this index, 50 is the break-even point; a value above 50 shows more confidence, below 50 shows less.

Pros: Builders are better economy-watchers than those questioned in general consumer surveys.

Cons: It's a lagging indicator and offers no direct measurement.



"Confidence" isn't alone in hard-to-plumb waters. For measuring income, Census's household income surveys aren't highly reliable and don't include investment income. The Personal Income and Outlays report from the Bureau of Economic Analysis carries more weight. But after the last hurricane season, Mark Vitner, senior economist with Wachovia Securities, switched to wage and salary data, which "are somewhat less distorted than the income statistics." On affordability, NAR calculates an oft-used index that blends calculations of income, price and mortgage rates based on a 20 percent down payment some see as obsolete due to new mortgage products. NAR's index has veered completely 180 degrees off course since 1999 with a negative 0.86 correlation. Rather than ignore imperfect indicators — none are perfect — builders should instead understand as many as possible in triangulating a course by which to steer.