Grandfathering Structure

by Don Friedman on May 23, 2016

The Times published an interesting article last week on the topic of existing buildings that do not meet current zoning. The title tells the story: “40 Percent of the Buildings in Manhattan Could Not Be Built Today.” The term for legal protection for existing conditions that cannot be legally created now is “grandfathering”

Rather than discuss zoning, I want to talk about a topic closer to home: grandfathering of obsolete structure. For example, the structural cast iron of the Haughwout building at 490 Broadway could no longer be legally constructed, regardless of how beautiful it may be (click for a bigger version of the HABS photo):


I’ve gradually come to realize that most people don’t understand how common grandfathering is. Every time the building code for a given location is changed, some existing buildings there change from code-compliant to non-compliant. Since my career started in 1987, the New York City Building Code and administrative rules have changed pretty much every year, with major changes in 1996, 2008, and 2014. The material codes that are the basis of the NYC provisions change regularly, as does ASCE 7, which is the basis for the structural loading provisions. Every one of those code changes made some existing buildings non-compliant, as did every code change going back to the publication of the second NYC code in 1887.

[For the rest of this discussion, I’m going to make the assumption that the existing buildings met the code that was in force at the time they were built. Obviously that assumption is not always true, but the failure of that assumption is a different topic entirely.]

Why is grandfathering allowed? First, most of the changes to codes are not drastic. The inclusion of seismic design in NYC in 1996 was a drastic change in concept, but because we have relatively low levels of possible seismic load, the practical effect was that a lot of existing buildings that were designed for our hurricane-force wind loads will work for earthquakes as well. The change from a relatively simple linear increase in wind loads under the 1968 code to a complex formula in 2008 has little effect on the design wind loads for most buildings.

Second, it is simply not practical to require every building to be upgraded to meet every new code provision. It would mean ripping apart people’s homes and workplaces on a regular basis. For example, every old tenement would lose usable floor area as the current requirement for universal access means that every multiple dwelling would have to get an elevator installed. The majority of buildings constructed before 1950 would also have to get new stairs to meet current geometric requirements.

Third, and most importantly, code provisions are meant to provide a given level of safety based on new construction, and encompass three basic inputs: dead load, live loads, and capacity. The dead loads (effectively the weight of the building) are known reasonably well for any given structure. The live loads (occupancy, wind, snow, and earthquakes) are based on statistical probabilities and are more or less the same for new and old buildings. Capacity is also based on probabilities (for example, the probability of a concrete beam being cast slightly smaller than the design size), and varies greatly with the structural systems used. The aspect of these calculations that is easy to overlook is that the probabilities change with time. The probabilities are ultimately based on observation (i.e. empirical evidence), and therefore the real (possibly uncalculatable) probabilities change as a building ages and its actual conditions can be examined. For example, if the code-specified snow load is based on a 50-year mean recurrence time, that means that, on average, a buildings roof will be subjected to that snow load every fifty years. The words “on average” are doing a lot of work there: a building could see a heavier snow the first winter after it was built, and an old building may never have seen the code load. But, in general, the longer a roof has been in existence, the more likely it is that the roof has supported a large percentage of the code snow load. A small change in that snow load from a new code is therefore less worrying for the old building that has a proven record of supporting snow than for a new one that may never have been heavily loaded.

In short, the mere fact of survival of an old building (and I’m assuming that the building has survived in a usable condition) is proof that it can handle real-life loads with its real-life capacity; the longer it has survived, the more proof it provides. This proof does not supplant the code, but it certainly supplements it. In circumstances when current codes provide little guidance – and cast-iron columns, massive unreinforced masonry walls, and tile-arch floors all spring to mind – the empirical performance may be more useful than the code. Looking at those cases in that light, grandfathering looks more rational than analysis.

Previous post:

Next post: