In February, Housing Starts fell to a seasonally adjusted annual rate of 575,000, a decline of 5.9% from the revised January level and just 0.2% above a year ago. While on first blush the data looks quite negative, there are many caveats.

First, while the data is seasonally adjusted, the weather was much worse this year than last, and that might have had an effect, especially when looking at the regional breakdown. Starts were down by 9.6% in the Northeast on the month, but up 6.5% year over year in the region, while they were down 15.5% in the South for the month and off 12.4% year over year there. The big blizzard mostly affected the East Coast, both above and below the Mason-Dixon line.

Starts in the Midwest rose by 10.6% on the month and were up 11.8% year over year. Yes, snow did fall in the Midwest, but not as much as in the Northeast, and people in Chicago are much more used to dealing with snow than are people in Atlanta. The West was not affected by the blizzards, and there starts rose by 7.9% for the month and are up 21.3% year over year.

Then again, given the state of the economy — and especially the housing market — a year ago, being up from those levels is nothing to brag about…

The next caveat to taking a “too bearish” reading from this report is that the January numbers were revised sharply higher. Starts in January were originally reported as a 591,000 annual rate, so from where we thought we were, the decline was just 3.3%.

Also, single-family starts are holding up much better than the overall starts; most of the weakness was in the very volatile apartment and condo segment. Starts of buildings with five or more units plunged by 43.1% to a seasonally adjusted annual rate of just 58,000. That rate is 71.6% below the year-ago level. Single family starts were down just 0.6% on the month to a rate of 499,000, and are up 39.8% from the extremely depressed level of a year ago.

Building Permits Point Forward

Looking forward, the best indicator of future housing starts is usually Building Permits, which declined to a seasonally adjusted annual rate of 612,000, a decline of 1.6% on the month and up 11.3% from a year ago. The January level of permits was revised slightly higher from 621,000 to 622,000.

There, too, the weakness was mostly in the apartment and condo area, where permits were down 10.1% on the month and are down 41.4% year over year. Single-family permits were down just 0.2% on the month to a seasonally adjusted average of 503.000, and are up 32.0% from a year ago.

Regionally, the Midwest was the strongest, with an 11.7% monthly rise and up 23.5% year over year. The Northeast was unchanged on the month and up a slight 2.8% from a year ago. Permits in the West were down 2,1% on the month but are 38.6% above last year’s levels.

The all-important South region — which typically represents more than half of all housing activity in the country — suffered a 5.8% decline on the month and is just 0.3% above the depressed levels of a year ago. Getting a building permit is a nice indoor activity, and is not very weather-sensitive, unlike actually putting a shovel in the ground.

Do We Really Want Housing Starts to Increase?

Finally, there is the more existential question of if we really want to see housing starts increase. We still have a massive inventory of houses, both new and used. This is especially true if one considers the shadow inventory of people who are far behind on their mortgages, or houses that are already in the foreclosure pipeline but not yet listed for sale. Until we get the inventories under control, each new house built simply adds to the problem.

The U.S. is decidedly not suffering from a shortage of homes, at least not with the current rate of household formation. Household formation is greatly influenced by the number of jobs out there. If a 24-year-old college grad does not have a job, he will move back home with mom and dad instead of getting his own place. People without jobs will tend to double up with friends and family.

However, historically residential investment — the biggest part of which is new home construction — has been the locomotive that helps pull the economy out of recessions. This can be seen in the graph below (from Note how both total housing starts (blue line) and single family starts (red line) rise sharply after each recession (with the bottom often coming during the recession, making starts an important leading indicator of the economy).

Building a house can employ a lot of people. After the 2001 recession, job creation really did not get underway until August 2003, and the bulk of job creation in that expansion happened over the next three years. By August of 2006, the economy had added 6.430 million jobs, of which 965,000 — or 15.0% — were in Construction.

At the start of the period, construction jobs were just 5.2% of all jobs in the country, and by August 2007, they had climbed to 5.7%. While the overall economy continued to gain jobs until November 2007, August 2006 was the peak for construction jobs. Since then, the construction industry has lost 2.170 million jobs, while the economy as a whole as 6.826 million fewer jobs than in August of 2006.

In other words, construction alone is responsible for 31.8% of all jobs lost since that time. In the process, Construction has fallen back to being just 4.3% of all jobs in the country. Those numbers do not count the auxiliary jobs that are created by housing construction. The carpenter working for D.R. Horton (DHI) is counted, but not the guy in the factory making faucets at Masco (MAS) or the gal making kitchen cabinets at Fortune Brands (FO).

A Classic “Chicken & Egg” Situation

Thus we are caught in a classic “chicken and the egg” dilemma. Demand for houses will not pick up and cure the inventory overhang until household formation picks up. Household formation will not pick up until the job market comes back.

Historically, construction has provided a disproportionate number of jobs coming out of recessions (and losses a disproportionate share going into them). Construction jobs are also either semi-skilled, or have very specific skills. Those skills do not necessarily transfer over well to other types of employment.

Non-residential construction is not going to be the answer, at least not in the private sector. The nation is just as overloaded with excess strip malls, office buildings and hotels as it is with houses, at least for the current level of demand. If we get more jobs that might change, as more workers will mean more cubicles to fill. Right now though, there is a lot of empty office space around and rents have generally been falling.

Bad commercial real estate is probably a bigger thereat right now to the banks than are bad residential mortgages. This is particularly true for the small and mid-sized banks that were largely locked out of the housing boom and instead focused on commercial real estate (not all small banks did, but lots are extremely exposed relative to their capital levels).

The Bigger Picture

In the long term, a lower level of housing starts is probably a good thing. Building things that you don’t need, and which do not produce a return on investment, is not the path to a stronger economy over the long term.

However, in terms of the current rate of GDP growth, weak residential investment is a big drag. After 14 straight quarters of subtracting from overall economic growth, in the last half of 2009 residential investment finally started to pull its own weight and helped the economy. That help may be in the process of petering out.

At this point, housing start numbers are extremely low. If we could just recover to the level of starts that was associated with the worst levels of prior bad recessions, we would see very big percentage increases. If we got back to the average of the worst single months for single family starts of the recessions of 1970, 73-74, 1980, 1982 and 1991 (i.e. every recession since the 1960’s except the 2001 recession, where housing did not decline significantly) then we would see a 21.1% increase from current levels.

The size of the population has expanded significantly over the last 45 years, and with more people one would think that we need more places for them to live. We do, but we built too many from 2002 though 2007. In effect, the demand for houses was pulled forward by the housing bubble.

It will be possible to generate pretty healthy percentage gains in housing starts without going to very heroic levels of housing starts, unless you call the worst month of the 1982-83 recession a good level of housing activity. However we are not going to go back to the sorts of levels of housing activity and residential investment that we saw just a few years ago. The low level of residential investment is going to be a principal reason that this economic recovery is going to be very anemic.

Dirk van Dijk, CFA is the Chief Equity Strategist for With more than 25 years investment experience he has become a popular commentator appearing in the Wall Street Journal and on CNBC. Dirk is also the Editor in charge of the market beating Zacks Strategic Investor service.

More about Zacks Strategic Investor >>

Read the full analyst report on “DHI”
Read the full analyst report on “MAS”
Read the full analyst report on “FO”
Zacks Investment Research