Last week I met with another of Australia’s major energy users, and was asked questions by some of their international representatives about the approach we have taken in assisting C&I energy users in a particular form of Demand Response:
The particular form = spot exposure, and curtailment at times of high prices.
Please call (07) 3368 4064 if you would like to know more
One of the foreign representatives asked what the potential benefits would have been, historically, had they taken such an approach.
Of course, the real answer is “it depends …”, with the suffix added “… on a whole range of different assumptions…”
Our overseas visitors suggested comparing average contract prices with average spot prices to provide an initial sense-check of the possible benefits of the method that I was speaking about.
In over a decade of working with large energy users in this way, we have evolved a 4-step process (outlined here) that we loosely follow to help customers in their journey. The questions we were asked last week fit within Step 2 (i.e. “define business case”).
Rather than just complete some analysis and provide it directly to our clients in return for some modest consulting fee, we opted to make the following analysis freely-accessible on the new DemandResponse site, in order that it might be used by these particular energy users, and other interested parties as well over time.
1) Comparing Spot and Contract Prices
The analysis requested is possible to complete for the 4 mainland regions of the NEM, but not really for Tasmania, give the lack of transparency (and competition) in the contracts market.
1a) Contract and Spot in NSW for CAL 2015
In the following chart, we look at the trend of reference prices for:
(i) the BASE 2015 contract for NSW (from our NEMFuturesUpdater product via ASX Energy); and
(ii) spot prices for NSW across the 2015 calendar year (from our NEM-Review product using AEMO Data).
Hence the chart spans more than 4 calendar years, from the start of when the contract started to trade out till the end of the 2015 calendar year.
Our overseas visitors asked us to compare the average contract price for the period (i.e. the solid blue line) with the average spot price for the period (i.e. the solid maroon line).
Using this metric, one might conclude that spot would have been better than contract – however readers will be aware of the many simplifying assumptions that would be implicit in drawing such a high-level conclusion (some of which I outline below).
With respect to the same region, and using an approach that is similar in some ways, Mike Williams recently posted his own analysis here on the site.
1b) Contract and Spot in QLD for CAL 2015
Performing the same sort of analysis for Queensland, we see a different chart appear – by virtue of different dynamics in the region over the period chosen:
As we can see in this case, the average of the contract prices over the 3 preceding years slightly undercuts the simple time-weighted average of spot prices over the calendar year.
1a) Contract and Spot in VIC for CAL 2015
Turning to Victoria, we see a picture that’s similar to that for NSW – albeit with a lower average cost:
In this case, there’s a gap of approximately $10/MWh between average contract and average spot – though we see that purchase of a hedge close to the start of the calendar year would have secured an average price across the year down near the average spot.
Those interested in reading more should read Mike William’s take for Victoria here.
1a) Contract and Spot in SA for CAL 2015
Finally we turn to South Australia:
In this case we see that South Australia saw higher volatility than seen in VIC, which is one of the reasons why the average spot price is seen up at $50/MWh. We also see the volatility spread more uniformly through the year compared to in QLD.
Mike Williams also posted some analysis for South Australia here.
2) (Some of) the Simplifying Assumptions in the Above
With only limited time today, I have included below a quick dump of a number of the simplifying assumptions made in the crude analysis above.
2a) Only looked at CAL 2015
For a start, we’ve only looked at calendar 2015 – not individual quarters (or months), and not other years.
Those seeking to take this type of analysis further should study a longer-term time period (and remember that “past behaviour does not necessarily predict future behaviour”)
2b) Assumed a Flat Load
In this case, we have assumed that the load is fully flat, by basing the analysis on both:
(i) A BASE hedge; and
(ii) The simple time-weighted average of spot prices.
In practical terms, all loads deviate from this “perfect world” scenario, and so would the results.
2c) Overlooked cost of purchase (in either option)
Each procurement model will have different transaction costs. None of these have been considered in this high-level comparison.
2d) Did not assume when contracts were purchased
As can be seen in the charts above, contract prices for CAL 2015 varied significantly across the 3 years prior to the start of that year. In the case of this year (noting point 2a above) the overall trend was down.
Hence an approach of “timing the market” to delay the purchase of the hedge (in this case) would have secured a lower price.
2e) Large volumes might have moved the market
The follow-on point to the above, of course, is that large volumes purchased on the market might have moved the market to a degree that is not possible to ascertain in the above.
2f) Did not factor specific curtailability
On the spot side, the time-weighted average price for the year did not factor in any Demand Response at times of high prices, at all.
Hence the “average spot” line should be viewed as a starting point, with the average cost achieved with curtailment being below this (with the difference depending on how much curtailment is chosen).
Understanding this potential is where we help some of our clients, as noted in step 2 here (by working through several sequential levels of sophistication in analysis).