Calculating Carbon Tax Costs Is Challenging!

Back in May 2017 I wrote a blog post providing estimates of potential carbon tax costs to households for different levels of the carbon tax (this was based on a briefing note I produced at the request of the Senate). These numbers have been used and quoted by others, most recently in a National Post piece by Claudia Cattaneo. As a result, I’ve received quite a few questions about interpretation of the numbers and the assumptions behind them.

Often when the numbers are quoted they are missing much of the nuance in my blog post. I don’t mean to denigrate anyone who uses the numbers (I am very happy they are being used in policy discussions); the reality is that there are a lot of assumptions built into these calculations. Unfortunately, there have to be. Part of the reason is limits imposed by data availability. In addition, there is significant variability across individuals, businesses and other organizations in their responses to a carbon tax. In the absence of some simplifying assumptions, it is difficult to convey the main points in an easily understood manner. The purpose of this post is to provide more detail on the assumptions I made, and explain why I made those assumptions.

First off, how should we interpret the costs I calculated? One of the drawbacks of my methodology is that I assumed no behavioral changes on the part of households. This means that I am assuming households use the same amount of energy (and produce the same amount of emissions) regardless of whether or not there is a carbon tax. This is clearly an over-simplification. We know people will respond to price changes (if not exactly by how much). We also know governments use some of the carbon tax revenue to fund energy efficiency enhancements by households and businesses; this should have some impact on energy consumption, particularly over the longer term. Because of these factors, my estimates can be thought of as an upper bound on potential costs to households.

I also didn’t incorporate policy actions taken by governments to reduce carbon costs to households. For example, the Government of Alberta has recently implemented a policy called the Carbon Competitiveness Incentive (CCI), which reduces the carbon costs to large emitters, including electricity generators. This means the carbon tax costs borne by households may be lower. I tried to compensate for this by excluding electricity cost increases associated with the carbon tax in an alternative set of estimates. You could think of this as a rough approximation of the effect of the CCI on power prices, since the goal of that policy was to buffer power prices from the effect of carbon taxes.

I also had to make assumptions about gasoline use and energy used for home heating. This is again a source of potential error, making my calculations less precise. It means I overestimate the costs for some households and underestimate the costs for others. My calculations were for an average household, based on total energy use (for each energy source) in each province divided by the number of households using that energy source, and then multiplied by the carbon cost for each fuel type. It’s a rough way to estimate average household energy use, but without more detailed data it is very difficult to have more precise cost estimates by energy source. It’s made trickier by the fact that in some provinces, electricity is used for home heating, but without more detailed data and analysis I do not know how much electricity is for heating versus lights and appliances. This likely means that my calculations of home heating costs will be overestimates for those using electricity for home heating, and my calculations of electricity costs will be overestimates for those using natural gas or heating oil for home heating.

To conclude, I want to emphasize that the numbers I presented in that May 2017 blog have limitations, but I think my calculations do provide useful information. At the very least, they provide the approximate upper bound for carbon tax costs to households. As more (and better) data becomes available I hope to keep updating these numbers to keep everyone informed.