Monday, November 12, 2007

Is Government Efficient? Evidence from Commercial Building Energy Consumption

I am starting a new empirical project on energy consumption by various cities. Buildings are major energy consumer. I realize that most bloggers do not post econometrics output in their blog outputs but so be it. The unit of analysis is a commercial building so a Starbucks or a post office. There are 4716 in this year 2003 sample. The dependent variable is the natural log of the building's annual energy consumption.

Some Definitions;

1. builtyr = year the building was built
2. size = square feet of the building
3. hdd65 = heating degree days to a base of 65 degree
4. cdd65 = cooling degree days to a base of 65
5. Gov = dummy variable that equals 1 if the building is a Government building
and it equals zero if the building is a private sector building.

So, in English; controlling for the building's size and birth year and local climate,
does the private sector consume less energy than the public sector?

Note that the OLS coefficient on Gov = ".474" so exp(.474)-1 = 60% higher
energy consumption by the public sector!

What is going on here? Is the Postman leaving the lights on or does the
public sector build really inefficient buildings because there is no incentive
to build an energy efficient building?

areg ldep builtyr size Gov hdd658 cdd658 [w=adjwt8], absorb(wkhrs8)
(analytic weights assumed)

X Beta T-Stat

builtyr -0.00127 -1.92000
size 0.00001 30.12000
Gov 0.47391 8.02000
hdd658 0.00005 3.31000
cdd658 -0.00001 -0.18000
_cons 14.58793 11.14000