My Twitter feed has been alive with news of a “new UBC Report” that according to its author, Dr. Karen Bakker from the UBC  Program on Water Governance, “concludes stopping Site C will create a larger number of sustainable jobs in the province“. This report has been cited in multiple locations in the last week so I have been hoping to get a chance to dig into it and having done so I am, once again, surprised at how easily questionable research can dominate the conversation on the Site C file. It was even cited in the Legislature. The selling point of the report (at least from what I have seen on Twitter) was that it showed that the BCUC Alternative portfolio produced 5 times more jobs in renewable energy than are expected to be produced by Site C. What I found through my research was that the majority of the jobs “created” by the BCUC Alternative Portfolio would likely never actually appear and that the costs for the renewables portfolio promoted by the anti-Site C activists would likely be much higher than was previously suggested. The rest of this blog post will examine why, in my opinion, the results from Dr. Bakker’s study are unreliable.

The report itself is actually an unsigned Briefing note and an accompanying spreadsheet analysis (link opens an Excel file). In the briefing note we are presented with an employment table that presents three columns of employment numbers. One for the BCUC Alternative portfolio, one for the BC Hydro Alternative portfolio and one for Site C Continued. The take-home message is that the cumulative person-years or “jobs” modeled for the BCUC Alternative portfolio is 208,498 person-years by 2094 while Site C only generates 40,578 person-years by 2094.  This is where the headline 5 times the job numbers headline comes from (actually 5.13 but who is counting).

Looking at the spreadsheet we find the basis for this table in  the tab “Comparison-LLF”. The BCUC numbers are derived from the tab “BCUC-LFF”. On that tab we find the “Employment Summary-All Resources (by year)” table starting at column T. Looking at this spreadsheet we can see where all the jobs are coming from. Adding up the columns we discover that the 208,498 person-years for the BCUC Alternative portfolio is made up of the following subtotals:

• 10,296 person-years from decommissioning Site C  (5% of the total)
• 183,600 from demand-side management (DSM) programs (88% of the total)
• 14,602 person-years from the various wind projects  (7% of the total)

Looking at this you immediately recognize that contrary to what the people on Twitter have been claiming, renewables, on their own, do not generate more jobs than Site C. Rather according to the briefing note Site C generates 40,578 jobs, almost 3 times as many jobs as the actual wind projects.

Looking at this table one recognizes an obvious initial flaw of the model with respect to the wind jobs. According to the model the wind facilities will generate a lot of jobs during their initial construction phase (for the Wind -PC18 column (Column X) that would be 310 person-years per year from 2034 through 2038). The authors then confusingly have those facilities providing a constant number of maintenance jobs (52 person-years per year) between 2039 and 2094 (over 55 years).

Anyone who has studied wind energy projects understands that wind facilities are not designed to operate for 55+ years. The typical wind project has a 20-25 year lifespan at which point the turbines have to be decommissioned and new turbines installed if the power is still needed. From a jobs perspective this would significantly increase the number of jobs generated by the wind projects. Tearing down old turbines and building new ones would generate a lot of person-years of work and yet these jobs are completely missing from the analysis. While this necessary component of a wind project life cycle analysis would represent a good thing for the people touting renewables as a job builder; it does pose a bit of a quandary for the the anti Site c activists.

The problem is that many of their cost-calculations omit the need to decommission and re-build facilities 3+ times over the time period Site C will be in service. In order to effectively replace Site C you have to triple the number of turbines AND include the cost to decommission each generation of obsolete turbines during each life cycle. By tripling the number of turbines and incorporating decommissioning costs, suddenly the costs of those replacement renewables roughly triples. This is a fatal flaw in the cost calculations produced by the anti-Site C activists. I have been banging on this drum for a bit but, confusingly, have heard little from the supporters of Site C on this obvious error in price calculations. A single generation of wind turbines cannot replace a dam intended to operate for 70-120 years yet this is what the modelers assume in their analysis.

The BCUC’s actual Alternative Portfolio spreadsheet does partially account for  this cost but makes a number of odd assumptions including reducing the costs for refurbishing facilities (by 30% of original costs); omitting any decommissioning costs; and most strangely assuming that operations and maintenance costs go down over the lifespan of the facilities. I’m not sure about you, but I’ve found that as systems age they need more maintenance not less.

Going back to the Briefing Note, I have no issue with the person-years associated with decommissioning Site C in the BCUC Alternative portfolio but I have a very serious issue with those 183,000+ jobs attributed to demand-side management (DSM).

You might ask where Dr. Bakker and her colleagues got that huge number. Well according to the Briefing Note:

According to a study carried out for BC Hydro, spending on conservation or demand-side management (DSM) programs creates 30 jobs per $1M spent. 1

That 1 identifies Footnote 1 the Power Smart Employment Impacts DSM Programs, Rates and Codes and Standards, F2008 to F2037 (citing p. iv.). Now any time a footnote references a Roman Numeral that means the authors are sending you to the Executive Summary of the cited report. A good scientist never relies on the Executive Summary of a report because the Executive Summary typical does not contain any of the provisos from the body of the work itself. Looking at the Power Smart Report we do see in the Executive Summary that “employment intensity” for Power Smart DSM was an estimated 34.4 person-years per million dollars spent. The question lies: where does that number comes from?

Reading the report the 34 person-years of employment per million spent on DSM includes two components: Investment Effects and Re-Spending Effects. Investment Effects are described as:

These expenditures are required to implement the energy saving measures and are comparable to the construction expenditures and employment from supply-side projects. As depicted in Figure 4-1, investment employment ramps up over the initial years of the DSM Plan and achieves a plateau until F2028. Thereafter, it falls fairly rapidly with expenditures, but the decline is mitigated because some projects will be completed and paid out after F2028. Overall, the pattern of investment employment directly follows the expenditure pattern.

That is to not to say these are all direct jobs associated with spending all those millions. As described in the glossary, Investment Effects include:

Direct, indirect and induced employment estimated from the initial DSM investment expenditures in programs, rates, codes and standard measures.

Thus those original employment boost involves direct, indirect and induced employment. The other half is the “Re-spending Effects” which are described as:

The employment impacts from re-spending activity are estimate at 50,900 PYs [between 2008 and 2037] and are created as a byproduct of the economic benefits associated with the DSM expenditures (see Figure 4-1). Since these employment benefits continue, driven by the ongoing energy bill savings, they can be likened to the operation and maintenance employment from supply-side projects.

Did you get that second type? If not then let’s look at the Glossary [Definitions] that describes “Re-Spending Effects” as:

Direct, indirect and induced employment estimated from consumers’ re-spent electricity bill savings in the economy.

These two definitions should be raising red flags all over the place. First and foremost the Site C numbers are direct numbers not “induced or indirect employment”. Thus they are comparing apples to oranges. Moreover, Re-Spending Effects can only happen if you have substantial reductions in the consumer’s hydro bill associated with the DSM program. Now this is a huge assumption on the Site C file. As has been repeatedly told, if the project is scrapped then the sunk costs have to be recouped, reportedly by a 10% surtax placed on everybody’s hydro bills. This will also help cover the decommissioning costs. That surtax will eat up any initial savings the consumer might see from shutting down the project. It is unclear how people will re-spend savings that they never receive in the first place.

Moreover as I describe in my previous blog post, the only way for DSM to get us where we need to be [dramatically reduce our electricity demand] is by substantially increasing hydro bills. That is how DSM works, you make electricity more expensive to encourage consumers to use less. This brings up a rule-of-thumb I was taught as a student about DSM programs.  Consumers are generally used to how much they are willing to spend on household bills like hydro. If you increase the price of a household budgeting line item (electricity in this case) consumers will work to drop demand until they are spending about the same amount as they were spending previously. If the price rises substantially, they are often willing to spend a bit more to maintain a quality of life but will not make massive changes unless there is a big upside in savings. What this means is that the increase in price will likely drop demand but will not likely have a major effect on consumer hydro bills. Consequently the DSM measures will not result in reduced hydro bills that can be used to generate re-spending effects.

The absence of re-spending of non-existent savings becomes a serious consideration in the BCUC Alternative model because according  to the Power Smart document, the direct jobs (the Investment Effects) associated with the continued investment in DSM quickly disappear. In the Power Smart reference Figure 2-1 shows that under a steady investment state employment increases associated with Investment Effects essentially disappear about ten years into the spending cycle. By their measure only 50,900 person years are generated regardless of how long the money is spent because after that time the monies are spent on established programs and not on lots of people to set up those programs. For the UBC modelers this is a problem since they assume that 183,600 person-years will be generated by DSM between now and 2094 but if the Re-Spending Effects don’t appear then about 132,700 of those jobs simply vanish from the equation. Take those 132,700 person-years out of the equation then all those talking points go away. Instead of the BCUC Alternative portfolio generating 208,498 jobs it only produces 75,798 jobs which is barely double what Site C generates. Given the mushiness of the assumptions used in this report those two number may as well be the same.

Looking at the Briefing note I remain amazed at how easily the public can be swayed by someone with a fancy title and a complicated spreadsheet. Like the Swain model once you dig into the numbers it becomes increasingly clear that the output of the model is entirely dependent on the input assumptions and that, in this case, the input assumptions are demonstrably faulty. Wind farms don’t last for over 55 years they last closer to 25 and if wind is going to replace Site C then you will need to account for the equivalent time frame for a fair comparison.  On the DSM front, a modeling exercise that assumes that 64% of your total jobs will be derived from people spending savings they never obtained is not a good thing. Finally for the anti-Site C folks who keep proclaiming that renewables will make up for the jobs from Site C the Briefing Note makes it clear that the actual renewable job numbers don’t come close to comparing to the number of jobs generated by Site C. On the positive front, as I have pointed out numerous times at this blog, Site C will not come close to supplying the energy we need to electrify our transportation sector. As such we will need DSM (and all the ensuing jobs) in addition to Site C if we are to meet our Paris Agreement energy commitments.