Permits & Construction Costs
The response to Noah's piece, The Manhattan Inventory Argument, suggests that folks might like to see some data on Manhattan housing permits (these are planned construction projects; not every one gets built) along with the construction cost data that is included. First, a couple of points: the data below is for New York County which is constituted by Manhattan and I believe a tiny slice of the Bronx (Guru help on this?) and the data I present below only include buildings for five or more families. The buildings could be apartments or condos; the government does not split it out by end-use. I would note that in recent years, the cost of land and construction, made financing the construction of apartment buildings prohibitive, although building apartments is a better alternative in terms of after-tax returns to builders. The apartments built were generally on land people acquired cheaply years before, or for projects where lots of developer equity was available. In recent years, most developers wanted to use as much leverage as possible and get paid back as quickly as possible so they could pay back their construction loan. With stricter lending rules, apartment projects are becoming a bigger piece of the pie these days. Okay, so here is the data on building permits starting in 1990.
You can see the impact of the credit crunch in the early 1990s - it's too bad no data from the late 80s are available to get a sense for how much construction dropped off. You can see the impact of the 2000 to 2003 economic slowdown as well and the big leg up in the 2005 to 2007 time frame. In contrast you can see the H1 2008 fall off versus H1 2007 below. Year-to-date permits indicate 2,037 units to be built, down 59.7% from 5,055 over the January to April period last year.
The 1990 and 1991 construction cost data have got to be an error in the government statistics. But as you can see, there has been a steady rise in construction costs per residential building unit. I am not sure that these numbers are super meaningful as the mix of buildings with large numbers of units versus buildings with small numbers of units will impact this figure a lot, which is why the annual percentage changes are very volatile. So the median annual change of construction cost per unit (starting in 1993 so as to factor out the weird 1990 and 1991 data) at 15% is probably more meaningful than the 47% average. Interestingly, the construction cost per unit rose 15% in 2007 - about the median amount.
Data: U.S. Census Bureau
Charts Courtesy of Guild Partners
