A step into the AGL FiT fight

Glaring errors are littered throughout a recently released paper being used to lobby the Queensland government to eliminate its feed-in tariff.

AGL's chief economist, AGL's head of economic policy and sustainability and an AGL energy market analyst recently released a paper “Queensland solar feed-in tariffs and the merit-order effect: economic benefit, or regressive taxation and wealth transfers?” Many of its calculations are laughable however, as are their assumptions and arguments. This article presents a small sample of such failings, which are shown to overstate costs, ignore the benefits to welfare of PV purchasers who are predominantly lower-income, and lead the reader astray.

Nelson and Simshauser's calculations are appallingly inaccurate, as detailed below. Together these add ~24-40 per cent to the costs claimed.

1. They ‘assume' that all systems are 1.5kW in size, when the Queensland average is 2.16kW.

2. The authors assume that every system performs optimally, producing 4.1 kWh/kWp/day, 17-32 per cent higher than measured recordings of 8300 systems in NSW, which showed median system performance to be 3.1-3.5 kWh/kWp/day. (The last consultant to make over-optimistic assumptions about solar performance severely embarrassed the NSW government).

3. An adjustment does need to be made for average system performance in QLD, which is 7.7 per cent higher than in NSW.

4. They then add 1.5 years worth of costs due to a simple calculation error, assuming a system installed in 2011 will produce 18 years worth of power by July 2028, when (last time I checked) there is 16.5 years between the 31/12/2011 and 1/7/2028 when the feed-in tariff will conclude. Net effect: 9.1 per cent.

5. They assume PV systems will export 30 per cent of their power generation. IPART “found that net exports and net export ratios for individual customers varied significantly across all PV unit sizes. This was largely due to the individual customers' different electricity consumption behaviour”. A more appropriate calculation methodology would be to determine the average export volume for each system size, and perform a cross multiplication.

6. The writers assume the power price and line losses in NSW are the same as in Queensland.

7. With reference to this (understated) figure, they conclude that (1-77.5/440) = 82 per cent of the value of the feed in tariff is a wealth transfer, but that somehow $978 million of $1062 million is the total transfer, which is equal to 92 per cent not 82 per cent.

8. If a discount rate was used, it is not mentioned anywhere.

9. No mention is made of increasing power prices over the period to 2028, which would substantially reduce the gap between the feed-in tariff and the export ‘value' of the power. It is as if they have assumed that power prices do not increase.

10. No allowance is made for panel degradation over the 16.5 year period, whereas PV panels reduce output by 10-20 per cent over the course of 20-25 years. This reduces the amount of export by a disproportionately larger amount, though to be conservative we could say 5 per cent.

This has the effect of vastly overstating the ‘wealth transfer' that occurs from all electricity consumers to solar owners as a result of a feed-in tariff policy. Serious ambiguity arises over the accuracy of Nelson and Simshauser work when so many mistakes have been made in the basic calculations. In addition the question of accuracy is raised again when the document goes on to further over-exaggerate the impact of PV:

1. Having claimed the policy is costing $1.56/MWh and that the average Queensland house currently consumes 7MWh/year, then the average annual cost for a household should (on the basis of their figures, however questionable) be $10.92 per year. Yes, this is how big an impact they are kicking up a fuss over.

2. They next present a table claiming the weighted average annual cost per household is between “83 and 121” without listing the units, which further misleads readers, rather than listing a value nearer to $10 per year.

3. These numbers are used to conclude there is a wealth transfer from the poor to the rich (having made the assumption that the rich are PV owners). The critical flaw in their methodology however, is failure to account for the fact that solar ownership is concentrated amongst low-to-medium-income households. The benefits that flow to lower income solar owners mean that the lower income brackets benefit from the feed-in tariff, thus improving the average welfare of their demographic by $40-$80/year per household, and those that don't install PV are shielded from $10 average cost of the policy (their calculations) by Queensland's $230/year pensioner electricity discount.

Furthermore, the paper repeatedly misleads its readers:

“It can be demonstrated that solar PV units avoid the wholesale cost of power generation (albeit on a discounted basis due to inherent variability and therefore uncertainty of production output)”- No regulator has discounted the value of PV because of its inherent variability.

“There is little doubt that this extraordinary growth was due to the premium feed-in tariff incentives available from 2009.” – PV industry growth was mostly driven by federal policy settings that made PV affordable, whereas system size was driven by feed-in tariff settings.

“In addition, the South Australian Government has legislated that beyond 30 September 2013, no mandated feed-in tariff will be payable for new installations. “ – The SA government has mandated a minimum retailer contribution, which is in effect a modest feed-in tariff.

“In theory, the value of ‘renewable energy' should not be ascribed to this $77.50 ‘fair value' because PV units have already received a balloon payment of Renewable Energy Certificates for its useful life upon installation via the deeming process under the Renewable Energy Target.” The $77.50 mentioned has no theoretical or inherent ‘renewability' about it, it is simply the wholesale cost of electricity at the time of generation, adjusted for losses, and some avoided market fees.

“The direct cost of renewable resources is ‘unambiguously more expensive than thermal plant'” – Not true – at the point of consumption, distributed PV is now less expensive than power transported from a distant generator. Many view solar farms as being competitive with coal power within a decade.

“If renewable energy capacity dominates too much… and spot prices are sustained below total plant costs even in the long run, the entry of requisite balancing plant becomes intractable. In the short and intermediate run, such a scenario poses no insurmountable threat to the physical operation of the grid. Over the long run given gradual growth in demand…” – Consumption is presently reducing, thus lessening the need for new entry of power stations – the existing ones can balance what's left. Furthermore, other network balancing activities are increasingly becoming commonplace.

“The position that the energy industry is left in today as a result of such policies was best summarised by the Australian Capital Territory Chief Minister who remarked on their own feed-in tariff arrangements – ‘the whole episode has demonstrated the folly of making policy on the run'” – The ACT Chief Minister was not referring to the energy industry's position as a result of feed-in tariffs, nor was she advocating against feed in tariffs. Instead, she was arguing against the unconsidered stop-start-stop end to the ACT Feed-in Tariff.  Here's the full quote from the Media Release: “Government warned quite clearly that re-opening the scheme did not offer transition, rather simply offered another gallop towards a new cap. ‘It has proven to be a very fast gallop that has not provided the ‘softer landing' for local solar industry promised by Mr Rattenbury. The whole episode has demonstrated the folly of making policy on the run,' said Ms Gallagher.”

In a lengthy investigation into the merit order effect, Simshauser and Nelson conclude that it does exist, but they use wind farms to show it to only be a temporary effect. Simshauser and Nelson reference as fact another paper (that they wrote) that states renewable resources are “unambiguously more expensive than thermal plant”. Because it is the average cost of generation that must be recovered by the market in the long run, they argue that more expensive plant will ultimately drive average prices up.

However, rooftop PV is clearly not a wind farm. Distributed PV having now reached grid parity, solar power has both lower average cost and instantaneous cost than conventional supply, at least proportional to the extent that generated power is consumed on site. This means that its ongoing deployment is certain in many situations, which will thus lead to a lasting reduction in power prices. Even if it was only a hypothetical example in their minds, at least Nelson and Simshauser concur, “If a merit order effect occurred due to natural market disruption events such as the entry of a new renewable or low emission technology with lower marginal running costs and lower total costs than incumbent generators, it would produce an unambiguous improvement in welfare.”

This is not to say that Nelson and Simshauser do not make some good, if somewhat moot, points. It is just a shame their arguments are so tarnished by misleading language, assumptions, calculations and appearance. Though posing as an academic paper, their calculations require more rigour. Their language betrays a lack of impartiality: “excessively generous”, “too successful”, “surely make for sobering reading”, “it has been surprisingly well documented in academic literature”, “an oversupply of higher cost renewable capacity was purposefully engineered”, “would represent courageous assumptions at best”. 

It is quite concerning that employees of a large utility have created an academic paper with so many shortcomings. Clearly this document cannot be relied upon to factually report the situation and should be read with as much scepticism as any other document that will be employed to lobby government.

Warwick Johnston is the Managing Director of SunWiz Consulting.