The Initiative

Initiative 732 would have established a tax on carbon emissions at $15 per metric ton of emissions in July 2017, $25 in July 2018, and then 3.5 percent plus inflation each year until the tax reaches $100 per metric ton.

The Business Opportunity

A tax on carbon emissions would increase the cost to generate power inside the state of Washington. Purchasing long-term fixed price contracts that were based on pre-tax fundamentals would generate gains if the legislation was passed. Two questions arise when valuing this opportunity:

1. What is the price impact on WECC, specifically Mid-C, power markets?
2. What is the probability the legislation would pass?

The Methodology

To address the price impact question Ansergy ran two scenarios through its term forecast model:

1. Base Case – current assumptions, no carbon tax
2. Carbon Case – current assumptions plus an adder for the carbon tax

The base case is easy, Ansergy updates its internal forecast 6-8 times a day for changes in all price drivers including the following:

• Loads – the front incorporates the latest weather, the back is adjusted for load growth and changes in distributed generation
• Hydro – the Ansergy hydro energy forecast is updated each day to incorporate changes in snow, reservoir levels, regulation, and precipitation. These changes (snow, reservoirs, and precipitation) affect the current water year (Today through October 2017).
• Fuels – most recent natural gas, coal, and fuel oil prices
• Outages
• TTC
• Generation Capacity

Computing the carbon case is a bit more problematic given the Initiative 732 legislation covered only the state of Washington, though the tax would be assessed on any energy imported that was generated from a carbon-based fuel. The question then becomes which Mid-C plants would be affected by the tax. Clearly the in-state plants all would, but what about the coal in Montana or Wyoming? What about the natural gas plants in Oregon?

Typically energy flows east to west, north to south. Rarely would Oregon gas plants operate to serve Washington loads, but most of the Colstrip energy flows into the state. The Wyoming plants (Bridger) are more problematic as some of that energy flows from Idaho to Utah, some of it is off-loaded into Idaho Power’s system, and some of it flows across Oregon to serve PacifiCorp’s loads.

Rather than attempt to model each plant separately we elected to treat all Mid-C generation the same; every plant would be subject to the tax. You can call this the “worst-case” scenario.

Study Results

Ansergy compared the output from the base case to the carbon tax case and generated deltas (Carbon Tax Case – Base Case). The following chart plots the changes by month for on peak prices:

A couple of observations we noted:

• Q2 and Q3 were impacted the most
• The outside hubs realized a much smaller impact; in other words, the price impact of the  WA Carbon tax did not materially affect the rest of the hubs, though the cost to serve load did (see below).

Off peak had similar ranges though the Q2 had a lower impact than it did in the on.

 

Quarterly Prices

As mentioned above, Q4 was least impacted by the changes and the impact on the other hubs was negligible.

CY 2018 Prices

A couple of observations on the calendar year results:

• The on versus off-peak spread was relatively unchanged
• Locational spreads essentially collapsed
• All adjoining hubs (NP, SP, and GB) were impacted about the same

Change in Exports

The above table returns the net change in exports out of Mid-C and into the adjoining hubs. The MidC exported 600 aMW less on peak and 450 aMW less off peak with the balance shared across all three adjoining hubs. On the surface does not seem to be too big of a deal, 100-200 aMW lower imports, but the cost of those imports to the receiving hubs rose substantially.

The northwest would have received a net positive cash flow from exporting more expensive energy to the outside hubs, though the quantity was lower,  the price was substantially higher, resulting in a net inflow of cash to Mid-C. The amount of cash inflows was reduced by a net outflow of cash to British Columbia as it was assumed the Canadians exported the same amount of energy but received the higher price.

Observations on the Market

Though it is to late to act upon this study we still found value in both exmaining the fundamental impact of the failed legislation and how the market traded the potential opportunity.  To put it another way, we  now have the luxury of knowing the legislation failed which provides an excellent opportunity to armchair quarterback the market’s reaction to this proposed legislation. A convenient way to do that is to plot the Cal 2018 price changes before and after the election.

Cal 2018 On Peak

In this chart we plotted the closing prices for CY18 on peak for the four traded WECC hubs. The change is all relative to the closing prices as of October 7, 2016 through the day after the election, November 9, 2016. Interesting chart, right?  

Observations

1. We think the market got the Mid-C premium right, it capped at out at $2.50, whereas our model had the change at $4.00, but that assumed that all northwest units would be subject to the cap. That said, one can impute the implied probability of the election passing at somewhere north of 50%, probably closer to 80%. Therein lies the mistakes made, since the initiative failed substantially, a more reasonable probability would have been in the 10-20% range, implying an expected value closer to $0.50/mwh, not the $2.50 that traded. Here is a perfect example of the long-held trading axiom of “Buy the rumor, sell the fact.” The savvy trader would have bought in at a $0.50 to $1.00 premium, realizing that he has already over-paid, and sold on either Friday or Monday. No way should anyone have carried this trade into the election, not given the expected values nor the huge risks associated with energy traders playing political pundits.
2. The market also rallied the NP15 which was a huge blunder since the computed premium (worst case, that is) was only $0.60. Apply a generous 20% expected value on the legislation passing and the fair value premium should have been a dime. And note the market ignored SP15 which totally teed up selling the spread for size and huge profits.
3. Poor Palo, it received no Carbon love, nor should it have, instead it was punished and pounded down, probably as a parking garage for an offsetting short. Now it is the buy on the curve and that hedge is going to be unwound at a loss.

Cal 2018 Off Peak

The off peak realized substantially less volatility than did on, though fundamentally they would be equally impacted. Plus the market actually sold down all of the other hubs even though they were equally impacted in the same ratio as the on peak.

Observations

1. The market traded the off-peak clearly not understanding how the tax would impact that market
2. Palo was over-sold and is now a buy, relative to the on peaks at the other hubs
3. All off peaks are a buy, at least relative to both the on and where off traded on Oct 7.

We think the real lesson to be learned from this exercise is the importance of probability; in the case of Initiative 732 the market clearly assigned too high of probability and, consequently, overbid the on peak. Another lesson is the value of a fundamental model, like Ansergy’s, that can quickly, and accurately, compute deltas. Don’t guess, know.