In the last blog post I reviewed the process EPA used to update its building data set which had become out of date. The 2003 Commercial Building Energy Consumption (CBECS) building data was updated with survey data collected in 2012. In this post I take a look at the impact this change had on a commercial real estate facility.
I’d liken what happened up until this summer to the grade inflation we’ve seen at some of our schools. Even when you grade on the curve, if everybody is acing all the tests, how do you get an even distribution? Make the tests more difficult. Not everyone can get a trophy. And that’s basically what happened in August of 2018. For some, the test got a lot tougher.
Understanding the impact of these changes can take some patience and a willingness to dig into the vagaries of statistical analysis, not something your average salesperson searches for in their spare time. If you have an Energy Engineer nearby, I suggest you grab a few of their favorite IPAs and settle in. For those less fortunate, let’s take a shot at it and remember we’re here to help if you need us.
What I’m going to try to explain is documented in the graphic below. The red dotted vertical line separates data generated from the 2003 data set from the 2012 data set. The left, blue axis tracks the Energy Use Intensity (EUI) for the building from April 2017 to the end of 2018. Note that there is no significant variation in consumption over the term.
The right, rust colored axis tracks the actual Energy Star Score. So, while consumption remains relatively constant, looking at the Energy Star score using the 2003 data set (rust line left of the red vertical line) and comparing it to the 2012 based score we drop almost 30 points. Why?
To answer the question, we need to look at the relationship between the facility in question and the survey data (see note 1 at bottom of post). Ok, per my warning, it may be time to crack that beer…
Looking at this facility, we can see using the 2003 data the building’s Actual Source EUI was 160.32 (Solid Blue Line). At that time, the Energy Star’s mean Predicted Source EUI fell from 179.31 to 126.36 (Dashed Blue Line). Meaning the School Board got wise and toughened up the tests. Or, mixing metaphors, someone moved the goal posts about 30%. Hmmm…..
Ok, beer #2. Using the following equation (see note 2 at bottom of post), we can see that our Efficiency Ratio rose from roughly 0.89 to 1.3 which corresponds to the Cumulative Percent of 51% and 80%. Since the cumulative percent and Energy Star score both range from 0 to 100 and are inversely correlated, we can subtract the Cumulative Percent from 100 to find the Energy Star Scores of 49 for 2003 and 20 for 2012.
The value at the time of the change in 8/18 are as follows:
While Energy Star predicted that the score for an office would fall on average only 12 points, this case shows how an office could fall by 29 points. It’s a function of its relative performance versus peers. For an office, we can see that change in the Efficiency Ratio has a greater impact on buildings in the center than the outliers so while the average drop was 12 points, buildings in the center would likely see a larger drop, while the outliers would see less of a drop.
Note 1: To form a just peer group, Portfolio Manager first normalizes the CBECS data set for based off space type, weather and facility space type characteristics. The next step is to use the coefficients that result from the normalization process and your unique building characteristics to predict the mean source EUI for a facility like yours. The ratio of your actual EUI over the predicted EUI, or your Efficiency Ratio, is then used to find the percent of facilities that perform better/worse than you.
Note 2: Energy Efficiency Ratio = (Actual Source EUI / Predicted Source EUI)