A few months ago Washington DC released its 2012 energy benchmarking data for private commercial buildings 150,000 sf and larger. Credible energy and water consumption data for some 400 buildings were released, of which 246 were office buildings. A recent article — stemming from the web site LEED Exposed — has claimed that these data show LEED buildings use more energy than non-LEED buildings. Specifically it is claimed that LEED buildings have an average weather normalized source EUI of 205 kBtu/sf whereas non-LEED buildings have an average EUI of 199 kBtu/sf. No details are provided to support this claim.
My students and I have cross-listed the DC benchmarking data with the USGBC LEED Project directory and identified 94 LEED-certified buildings in the 2012 DC benchmarking dataset — all but one being classified as office buildings. The unweighted mean weather normalized source EUI for these 94 LEED certified buildings is 202 kBtu/sf. The unweighted mean weather normalized source EUI for remaining 305 buildings is 198 kBtu/sf. No doubt this is the basis for the claim that LEED buildings use more energy than non-LEED. However, the difference is not statistically significant.
Moreover, the non-LEED dataset, in addition to 154 office buildings, contains 64 (unrefrigerated) warehouses and 90 multifamily housing buildings — all of which use significantly less energy than do office buildings. The comparison of these two average EUI is not useful — just a meaningless sound bite.
It should also be noted that the unweighted mean EUI for a collection of buildings is an inappropriate measure of their total energy consumption. The appropriate measure of energy consumption is their gross energy intensity — their total source energy divided by the total gross square footage. This issue has been discussed in several papers [2008 IEPEC; 2009 Energy & Buildings].
Note that an apples-to-apples comparison of energy consumed by one set of buildings to that consumed by another requires that the two sets contain the same kinds of buildings in similar proportions. When possible this is best accomplished by sticking to one specific building type. Since office buildings are far and away the most common in both datasets it makes sense to make an office-to-office comparison — pun intended.
93 of the LEED-certified buildings are offices. But many of these buildings were not certified during the period for which data were collected. Some were certified during 2012 and others were not certified until 2013 or 2014. Only 46 of the office buildings were certified before Jan. 1, 2012 and are then expected to demonstrate energy and GHG emissions savings for 2012.
The 2012 gross weather-normalized source energy intensity for the 46 LEED certified office buildings is 191 kBtu/sf. This is 16% lower than the gross weather-normalized source energy intensity for the 154 non-certified office buildings in the dataset, 229 kBtu/sf. These modest savings are real and statistically significant, though much lower than the 30-40% savings routinely claimed by the USGBC.
Note that similar savings were not found in 2011 or 2012 NYC energy benchmarking data. Analysis of these data showed that LEED-certified office buildings in NYC used the same amount of primary energy and emitted no less green house gases than did other large NYC office buildings. So the 2012 results from Washington DC are significantly different. It should be noted that NYC office buildings certified at the gold level were found to exhibit similar modest energy savings. Perhaps this is a clue as to why Washington DC LEED buildings show energy savings. More analysis is required.
For the last few years the USGBC has pointed to ENERGY STAR scores for LEED certified buildings as evidence of their energy efficiency. While ENERGY STAR scores have two important characteristics — they use source rather than site energy and they are based on actual energy measurements — they simply do not have sound scientific basis. The science has never been vetted, and my own analysis shows these scores are little more than placebos to encourage energy efficiency. They certainly do not have any quantitative value.
So to summarize, in 2012 LEED offices in Washington used 16% less source energy than did other office buildings in DC. What this means and whether such savings justify the added costs of LEED are open questions.
The Pragmatic Steward 
John, great article. It’s great to see more people actually digging into the “data” that can get haphazardly thrown around without enough actual analysis. I found that same was true for the article in the New Republic that was challenging the effectiveness of the LEED Platinum rating of One Bryant Park–maybe you saw it. Again, it did not really provide an apples-to-apples look against the building’s very unique programmatic components, mainly the energy intensity of its banking trading floors.
Having said that, I still think it’s dangerous for people to be weighing the appropriateness of LEED on just a single metric: energy efficiency on a kwh/SF basis. Energy efficiency is definitely important and reducing power consumption improves our ecological stewardship in a variety of ways, but LEED is not just an efficiency system and sustainability is not just about energy use. The very reason that LEED is made up of multiple components is because energy efficiency isn’t a silver bullet for environmental progress.
Even if the study collectively rendered only minimal energy efficiency benefits, that doesn’t necessarily mean that LEED isn’t valuable, or isn’t working. Issues of water conversation, indoor environmental quality, occupant health and productivity are all important factors that LEED helps us address. I feel like energy gets picked on most often because it’s the easiest for us to measure, but that doesn’t mean it’s all that counts.