Housing and home building ooze data, but that doesn't seem to take the risk out of the steep, front-end investment of money, time, and talent in developing and building homes and communities.

Fact is, though, data may not be what it claims to be, which can add business and economic risk. Successful home builders, residential developers, and their partners often work in two "now" time continuums. One is about operating today at a relatively high level of effectiveness, efficiency, and customer satisfaction. The other is almost entirely speculative, based on carrying forward assumptions that have proven to be reliable in the past and present, following a path that has worked, and placing bets that, given a number of visible variables, demand will be constant.

Both "nows" are critical. It is essential to the DNA of a builder or developer to ensure ways and means to continue doing that building and developing tomorrow, and tomorrow. This makes data--reliable data--a must.

The past two to three years--especially as mortgage interest rates have run at historic lows and job formations have picked up and companies have even started increasing wages as the United States economy found its post-Great Recession footing--most of the focus on risk to home builders' and developers' business models has been on supply-side constraints. The now-familiar unholy trinity of pain points--lending, lots, and labor--have been hogging the spotlight, at least among housing community insiders.

What's gotten less careful attention, however, is what's really going on on the demand side of housing's equation. One of the more important benchmarks planners, developers, land acquisition experts, and home building division, region, and national executives count on in their budgeting, business modeling, and investment strategies is household formations.

Thing is, the No. 1-cited data source for household formations information, the U.S. Census Bureau, is often dead-wrong in its releases. The reasons are simple, but complicated to describe. The Census gets much better data--because it spends far more money and surveys a huge sample of households--every 10 years, in what's called its decennial Census. But every quarter, it releases its Residential Vacancies and Homeownership report from American Community Survey samples that are far smaller. Those quarterly data reports are suspect. Calculated Risk blog chief Bill McBride notes:

This report is frequently mentioned by analysts and the media to track household formation, the homeownership rate, and the homeowner and rental vacancy rates. However, there are serious questions about the accuracy of this survey.

This survey might show the trend, but I wouldn't rely on the absolute numbers. The Census Bureau is investigating the differences between the HVS, ACS and decennial Census, and analysts probably shouldn't use the HVS to estimate the excess vacant supply or household formation, or rely on the homeownership rate, except as a guide to the trend.

Going even further than that, Zelman & Associates cautions that using this Census yardstick for household formation visibility and projection could be downright dangerous.

Every quarter, the Census Bureau releases the results of its Housing Vacancy and Homeownership Survey (HVS), which is primarily meant to measure the vacancy rate of the housing stock but it is also heavily relied on by the media and macro economists to gauge household formation. The problem is that the data have been incredibly volatile and the conclusion is often at odds with other measures of housing demand as we see it.

As part of the Zelman & Associates Z Report series of thought leadership and practical analysis, this piece looks at an alternative, more sensitive, and perhaps more reliable proxy for estimating household formations.

Based on public electric utility provider results, which aggregate to a 30 million unit sample, residential electric customer counts were up 1.3% year over year in 2Q16, four basis points stronger than in 1Q16 and setting another recovery high. In fact, the year-over-year rate of growth for this subset has improved each quarter since 1Q13 and for 17 out of the last 18. Speaking to the more stable data set, the median change in national customer counts over the last 15 years was a smaller 22%. Contrary to the HVS conclusion, we expect household formation to average 1.34 million in 2016, relatively steady with 2015 as the highest years of the recovery thus far.

Now, this is a big difference from what the Census would lead you to think, as you can see from this chart.

The issue now is, what does that difference translate to as you calculate and model the demand pipeline in the "now" that is your next 24 to 36 month horizon? Register for the Zelman & Associates Z Report for bi-weekly, exclusive, high-level, high-takeaway value analysis from one of housing's most astute observers of who and what makes home building and residential development tick. Simply click here. And remember, Zelman & Associates ceo Ivy Zelman will be with us as one of our HIVE deans, on Sept. 28 & 29, at the JW Marriott/LA Live in Los Angeles.