Supply_Costs

This entry tries to balance the costs of supply serving Ontario's electricity sector, and the costs to Ontario's consumers when they purchase their electricity.


Figures shown are estimates using publicly available information.  The final data for a monthly report is not available until the Global Adjustment figures for a month are finalized midway through the following month, however many of the estimates are valid, and will change little,  following the end of a month.  This page may be updated multiple times each month, as additional data is made available.  Rates were estimated for 2012 - 2013's estimates are likely to be low by whatever inflation agreements are in place for contracts (CPI inflation was below 2% in 2012).  Rates have not been adjusted for 2014 (as of the first post in February), and I have not yet reviewed the capacity/contingency fees paid for now closed OPG coal facilities (these were contracted through 2014 and closed a year early- may be some time before I figure out how that impacts the fees).

Confidence in the estimates is derived from checking both sides of an equation separately (accounting for all variables).  In this case the equation, oversimplified, is::
Cost of Supply = Total Revenues

Total Revenue

The Total Revenues from the system is the simpler side of the equation as it can be calculated with only a view publicly available data sources.

Sum([Total Market Demand] * [HOEP]) + [TTL GA] = Total Revenue

The Total Market Demand includes all Ontario Demand and all export demand (IESO's hourly .csv file starting May 2002 is here).
The TTL GA is the total amount of the global adjustment, meaning the Total GA (M$) value provided by the IESO, here.  Since this figure is a monthly value, the check in my estimating costs of each generator will be of monthly totals -  the hourly calculation of demand multiplies by the HOEP is summed to the monthly total, and added to this GA.

The graph shows the resulting total value of the total market, plus what that total value, divided by the total MWh, totals on both the monthly level, and as a 12-month moving average.  The chart is dynamic, starting one year after the market opened in May 2002.  Demand peaked in 2005.  The price dropped along with demand, with the 12-month moving average hitting $52.49 in January of 2007.  It has since been moving upwards to the $66 level in April 2012.  If you are a resident, or small business, in Ontario, these figures don't look familiar, as the average regulated price plan price moved moved from $75.65 up to $80.69 as of May 1, 2012.  I'll review the difference later.

Cost of Supply

The process to determine the cost of supply is far more complex. The basis for power production data is the IESO's Hourly Output and Capability Report, with daily data captured since September 1, 2010.  Added to that is: 
  • estimation of solar output (based on hourly figures for a London Ontario location I took from the net, and total solar capacity estimated from quarterly OPA reports)
  • estimation of CDM costs from the OPA (treated as only 1MWh a month)
  • hourly imports, at the hourly price (HOEP)
  • balancing entry, to force total generation to equate to total market (this is done at the monthly summary level, not the base hourly level)
Estimating the CDM costs is the recognition that the Ontario Power Authority (OPA) also has fundingm through the Global Adjustment mechanism; for spending primarily on 'conservation programs', which carry a cost of approximately $320 million a year (page 15 here).  The OPA does attempt to quantify demand reductions resulting from this spending (here), and it is commonplace to see a value for "negawatts" alongside values for other generation options; Ontario's Long Term Energy Plan goes so far as to forecast 7100 MW of conservation supply, of a total 48000, by 2030.   However, I have not attempted to quantify how much avoided demand to match up to the OPA's ~$320 million a year conservation budget.  When quantifying any price per MWh of power, there will be a discrepancy as some of the spending is on avoidance of generating power..

The balancing is necessary because the output and capability report does not include many small generators, including the solar I appended estimates for (overestimates of solar may be reason the adjusting entry is occasionally negative).  The challenges in estimating the total power produced are minor compared to estimating the payments for that power.  While Ontario as a market for power, and a market price, very little supply is purchased without an existing power purchasing agreement (PPA).

Public power, from Ontario Power Generation (OPG) is fairly straightforward as it has regulated rates for much of it's generation, and the unregulated supply can be assigned HOEP rates according the hourly demand.  But OPG also receives payments related to the availability of some coal units,  from the Ontario Electricity Finance Corporation (OEFC), as well payments for the availability of the huge Lennox Generating Station; these based on an agreement with the Ontario Power Authority (OPA).  
The contracts that private producers hold with the OPA, or with the OEFC (Ontario Electricity Finance Corporation) generally aren't visible.  Trailing through a variety of documents, including annual reports from various companies, and Navigant Consulting's documents setting the regulated price plans' rate, I've estimated costs for all.   As with the public coal generators, and Lennox, these contracts also are either structured to pay for power output, or the availability of capacity to produce power, or both.

The main data table I have created, based on my research (largely undocumented) is for individual generators.  I've shared the table as a google doc, and below indicate the rows relevant to costing supply:
PPA_Out holds the estimated value of power purchase agreements that are based on output (production) - a null value indicates the HOEP will be considered the price paid
PPA_Cap holds an hourly value per MW of capacity that is the estimated value of a contract that pays a price for capacity
GA_Group is the Global Adjustment group, or pot, that will be needed to pay out the difference beween the HOEP (market price), and the power purchase agreement price (PPA_Out from above)


Generator_List


Adding this table to a data structure comprised of the hourly generator output, and other hourly data tables from the IESO, a lot of analysis becomes possible.  In costing, and verifying against revenues, I have grouped the hourly data at the monthly level, as it is at this level the Global Adjustment is calculated, and therefore, it's at this level the comparison between this estimation of supplier costs, and total system's revenue, can be made.

Variance between System Revenue and Supplier Payments

Global Adjustment Database Estimates with Reported Actuals

The database view that I find most useful in assessing how valid the estimates are is comparing the components of the global adjustment as reported on the IESO's website, against my calculations.  
My calculations of the Global Adjustment is the valuation of power produced the PPA rate less the valuation at the HOEP, or market, rate, plus all capacity payments.  There is a lesson here about the Global Adjustment reporting categories; Ontario Power Generation (OPG) is a category in the reporting of the GA, but there are payments to OPG that fall within both the OEFC-NUG (non-utility generator) pot, and other payments within the OPA totals.

Global Adjustment Comparison

The comparison indicates I am a little off in my estimates - with the bulk of the difference being in the OPA (Ontario Power Authority) classification.  This isn't surprising as the OPA contracts are hidden, and it is where the bulk of the charges exist.
It is encouraging that the OPG calculations are very close.  OPG rates, largely regulated, are the most visible, and OPG also does not get paid, generally, when power is curtailed due to surplus (SBG).

I have ignored power curtailments in my costing.  I do estimate these in my weekly reporting, but I do not estimate them by supplier.  One non-utility generator dropping to no output is normal, but the entire group dropping to 500MW occurs only on weekends of very low demand.  Similarly, a nuclear unit at Bruce derating overnight is clearly a curtailment maneuver, but a shut-down for 3 days may be curtailment, or a service outage.

It is notable that the variance between my estimates and the reported figures had been increasing.

NOTE:  Since I wrote that I encountered a ministry communication on the Net Revenue Requirement (NRR's) of natural gas generators which resolved much of the discrepancy.  Post explaining changes is here.

Capacity Payments/LUEC

There are many obstacles in assigning pricing to individual generators, but the largest one of these obstacles is the costs of sufficient capacity to meet peak demand.

2011 Average Costs

The more recent contracts for the larger natural gas generators, OPG's coal units, and Lennox (essentially a strategic reserve option), all have capacity payments.   In the case of the larger CCGT (combined cycle gas turbine plants), the assumption is made that the entire net revenue requirement, estimated at $15000/MW month, or $20.50/MW per hour (by capacity), is a capacity payment.(updated values on 19/7/2012)  These plants are capable of earning revenues, but operating well below 40% capacity factors, at very depressed HOEP pricing, it's likely most of the net revenue requirement is meant as a capacity payment.  The Combined Heat and Power generator contracts are more challenging, and likely far more varied amongst suppliers.  I have started with the assumption their net revenue requirement is $15,970/MWmonth (from here), but recognizing some value for the steam cogeneration sales, I am treating the NRR as a capacity payment of $16.30/MWh (of capacity).  These would be extremely difficult to estimate any other way, as the actual 'capacity' payment depends on the steam sales, or the profit that should be made when a gas generator sells production above the marginal cost of production.  Conversely, I suspect my OPA calculations are low in part due to natural gas generators operating at a loss  - meaning they could sell output below the marginal cost of production and still be made whole to their net revenue requirements.

Regardless, the point to be made with capacity payments is they are offered not due to the generator, but due to Ontario's generation mix, and it's policies.  Coal plants are required currently, but we don't desire to use them.   This is also true of Lennox GS- which is cheap as insurance against a 20000MW shortfall (ie. Pickering B!), but not desirable to actually run. Natural gas plants were needed to replace coal plants, add intermediate supply capability, and balance intermittent, 'green' generation from renewables.  
Similarly, the cost of generation in Ontario has been a small component of rate hikes since the market opened in May 2002.  That's explained by a large component of supply continuing to come from assets (nuclear and hydro) that have not seen significant rate increases.  The majority of increases on residential bills involves transmission and delivery, which would involve the new, largely intermittent, renewable generators, as well as 'smart' grid initiatives bundled into the same 'green' package.

Having noted the limitations on the levelized unit electricity cost (LUEC) figures I have provided the figures.  It's difficult to show everything indicated with these figures in one simple graphic, and I'd suggest opening the spreadsheet and setting some filters to demonstrate the pricing for coal, and natural gas, is distorted in favour of consuming more (the higher the consumption, the lower the cost/MWh of the capacity charge. This has caused lower pricing in Ontario during higher demand periods.  For instance:
  • natural gas production is lowest in April 2011, where the rate is highest, and;
  • natural gas production is  highest in January 2012, where the rate is lowest
Accounting for all capacity charges is interpretive.  The capacity payments for Lennox are (if I recall correctly) incorporated into 'on peak' pricing for Ontario's time-of-use customers; and it would seem likely that much of the payments for gas-fired generation could have a similar treatment (I don't believe that to be true).  All could be allocated not to the cost of producing with the fuel, but the cost of a system with low capacity factors desired from higher emissions generators.

Monthly_LUEC


Subsidies

The accusation of subsidy is frequently made, and often the claim is based on the fallacy that all payments above the Hourly Ontario Energy Price are a 'subsidy' Most suppliers are contracted, receiving the market payments (sold at the HOEP), and then payments to settle the difference out of the global adjustment pot - which is fund, in different degrees, by ratepayers.  But the Global Adjustment mechanism cannot be used, in part because it includes charges that don't relate to generation (he opposite is true as the mechanism is used to fund the OPA's conservation program).
Almost all supply is subsidized if we treat any supply with payments above the HOEP a subsidy - and yet it's all paid for within the system. Ontario's market is not settingt a market price (HOEP) related to the true value.  The HOEP functions only to call for more supply (price is increasing) or less supply (price is declining).   The HOEP price is set by the highest accepted bid in the market. Because many suppliers are guaranteed pricing - thus bidding at very low, or negative, pricing -  the price is set by relatively few participants, including OPG's non-regulated hydroelectric assets, and coal and natural gas-fired generators.  The former has no/low fuel cost, and the latter generally have the plant paid for through the capacity contracts.  The HOEP therefore moves simply with the commodity pricing of natural gas (as do most North American market), and due to capacity payments for these generators, Ontario's HOEP is essentially only the marginal cost component of producing with natural gas.

Contribution To GA

I would prefer to see the term subsidy deal with the variance to the average price.  Because all supply is paid for, if not through the HOEP then by the global adjustment mechanism, the measure of which supply is being subsidized is the 'net' contribution to the global adjustment.  This approach I'll demonstrate with the Beck 2 hydro generating station:  
  • Production in May had an average HOEP value of approximately $19.99/MWh ($15.9Million)
  • Regulated pricing for the generator is $34.13 ($27 Million)
  • The regulated price is $14.14 above HOEP ($15.8 Million)
  • On production of 793,621MWh, treating payments above the HOEP as a subsidy means Beck 2's May production was subsidized $15.8 million
But
  • the production was sold, to the end market, for an average of $67.42
  • the system collected about $53.5 million for that supply, but the market HOEP price only paid OPG $15.0 Million
  • the production from Beck 2 therefore contributed $38.5 million to the global adjustment pot, from which it was reimbursed $11.2 million
  • $27.3 million was transferred from OPG to somewhere else, based on production from the Bruce 2 units.
If the term 'subsidy' is used, it should reference either payments outside of the system altogether (such as the Ontario Clean Energy Benefit), or the distribution of funds from one supplier to another.   Assuming all product is sold at the average price (Total Revenue from the top of this entry, divided by total market demand), I have programmed the calculation of what might be called a subsidy (when negative), but is more accurately described as the net contribution to the global adjustment pot.

The figures are not surprising.  Hydro, and nuclear, generators are both older generation, and the handling of debt acquired to construct the facilities may not reflect the full costs (to put it mildly).
Regardless, the low-cost of that generation provides a stabilizing influence on the average cost of supply.
A combination of intermittent renewable supply, no supply (CDM programs), and capacity payments to allow controllable supply (coal and natural gas) is putting upwards pressure on the average price.

Subsidies - Customer Segments

There is a second type of redistribution of costs that can be called a subsidy.  It is the movement of costs from one customer class to another.

When Ontario's move to a competitive market resulted in a panicky price freeze, in 2002, it was primarily aimed at residential and smaller business customers.  As that was ended in 2005, the Global Adjustment (GA) mechanism was introduced to allow Ontario's consumers to be billed/credited for the difference between charges at the marker, HOEP, rate, and the rates contracted/regulated from suppliers.  At that point most customers within Ontario were relatively equal, and because the value recovered by the HOEP ended the year above the payments to suppliers, there was a slight credit in the GA - overall export customers were also paying a similar rate (discussing only the commodity charge, not delivery or transmission).
But demand started dropping after 2005, and, predictably, so did pricing.  In 2009, the HOEP dropped almost to the same amount as the GA had grown to - and for 2011 the HOEP averaged only $31.46 while the GA was above $40/MWh.  As the GA is only chargeable in Ontario, this indicated a growing discrepancy with domestic, and export, consumers.
The recession alerted many to the fact that Ontario's largest industrial consumers had been shedding consumption since 2005, with the most noticeable example being the exit of Ontario's largest individual user, from Kidd Creek to electricity at half the price in Quebec.  In 2011 a second class of customer was created for the allocation of the global adjustment.  Class A customers are charges based on the 5 highest daily consumption levels over the year.  The reality is that this has resulted in lower payments for those customers.
So there are now 3, or 4,  consumer segment prices in the Ontario market: export,  Class A (large industry), residential (Regulated Price Plan - RPP), and Class B (the last 2 should end up the same, except residential rates are set in advance whereas class B rates are monthly (variance accounts should balance them out over time).

The shifting of costs from export, and Class A, segments is what drives the rates up from the HOEP average rate, in April about $66.10/MWh, to the $77.93 Class B customers averaged.

Rates


A couple of comments on the chart: The "unknown field" is largely because the global adjustment isn't allocated across all "Ontario Demand" indicated in various IESO reports, but by a slightly smaller figure they do not show in reports (the allocated quantity of energy withdrawn).
Separate figures for Class A and Class B are not shown until the middle of the following month.  Until that time, the figures shown on this page will be preliminary/incomplete estimates.


Comments