Efficiency gains and wealth transfers: they're not the same thing!

When decisions are being made about whether to intervene in a market – say, by introducing or modifying pricing regulations – there’s a key distinction to be drawn between potentially achievable efficiency gains -(which are relevant to such decisions) and bare transfers of wealth from one group to another (which very often aren’t). I’m constantly surprised – and alarmed - by the frequency with which this distinction is overlooked or misunderstood.

For example, the New Zealand Electricity Authority (EA) claimed recently that certain revisions to the transmission pricing methodology (TPM) would deliver billions of dollars in efficiency gains.[i] However, that contention had no foundation,[ii] because (amongst other things) those ‘gains’ were largely bare transfers. The EA is by no means the first to have made that mistake. In this Explainer I seek to clear up this ongoing source of confusion. 

Genuine efficiency gains

Economists recognise three different – but related – types of efficiency that can be enhanced or improved either through improved competitive rivalry between sellers or, when competition is lacking (e.g., when substantial market power exists), through market interventions such as regulated pricing. These three forms of efficiency are:

• Productive (or technical) efficiency, which refers to a market outcome whereby products and services are provided at the lowest possible cost, using facilities of optimal scale, over the long run, with existing technology;
• Allocative efficiency, which refers to market outcomes whereby prices and profit levels are consistent with the real resource cost of supplying each product, including a normal profit reward to suppliers – where this is case, society’s resources will be allocated between end uses in an optimal way, reflecting their preferences; and  
• Dynamic efficiency, which refers to the ability of markets to adapt over time in response to changes in consumer preferences and/or technology through the development of new products and services and/or production processes.

Put more simply, efficiency requires firms to produce the goods and services that consumers want (allocative efficiency) at the lowest cost (productive efficiency) and to continue to do so over the long run (dynamic efficiency). The classic example of the absence of these types of efficiency is where a product is provided by an unregulated, unrivalled monopoly. When a monopolist faces no competition – or any prospect of it emerging – this can:   

• compromise productive and dynamic efficiency, since the monopolist faces limited pressure to reduce its cost of supply or to invest and innovate in response to changes in market conditions;[iii] and
• reduce allocative efficiency, since there’ll be unmet demand from some consumers who aren’t willing to pay the ‘monopoly price’ but would be prepared to pay a lower price that still delivered the seller a reasonable profit.

The second effect described above – the allocative inefficiency – is particularly important to grasp and is best explained using a simple diagram. Figure 1 illustrates that a monopolist maximises its profits by restricting its output – in this case to 40 units[iv] - giving rise to a price of $6 (PM) that’s well above the cost of producing each of those units (a constant $2/unit in this simple example). At this $6 price, there are customers who don’t buy the product who would have done so if it’d been between $2 and $6, i.e., high enough to allow the firm to cover the extra $2 it would have incurred producing each extra unit.

Figure 1: Allocative inefficiency

By increasing its price above its cost of supply ($2), the firm causes $80 in economic welfare (in this case, what economists call ‘consumer surplus’[v]) to be lost altogether from ‘inefficiently unserved demand’, i.e., demand that could have happened at a lower price, but doesn’t due to the monopoly pricing. This is a ‘deadweight loss’ that’s not recovered by anyone else (represented by the yellow triangle). The corollary of this is that, when prices are at this level, it's possible to make an allocative efficiency improvement, by making someone better off without making anyone else equally worse off (also known as a ‘Pareto improvement’).

For example, if the monopolist was to reduce its price to, say, $2/unit, then it would sell (and consumers would buy) 40 extra units that wouldn’t otherwise have been exchanged. Figure 2 highlights that this would eliminate the previous deadweight loss (the yellow triangle in Figure 1) and generate $80 in economic welfare (in the form of extra consumer surplus) that didn’t previously exist. In other words, that $80 welfare gain for consumers has not come at the expense of anyone else – say, producers. It therefore represents a genuine allocative efficiency gain. It’s new wealth.   

Figure 2: Allocative efficiency gain

It’s the potential to improve allocative efficiency by reducing the deadweight loss from unserved demand that’s one of the chief motivations for intervening in markets by introducing regulation, e.g., controlling prices. Specifically, it may be possible to generate additional consumer welfare that is not predicated on an equivalent reduction in welfare for someone else. As Figures 1 and 2 illustrated, this is achieved primarily by reducing the size of the yellow triangle, i.e., reducing deadweight loss.

Bare wealth transfers

A reduction in deadweight loss (of $80) would not be the only outcome from moving away from monopoly pricing to more ‘cost-reflective’ tariffs in Figures 1 and 2. There’d also be a transfer of wealth from producers to consumers. This would be equal to the $160 reduction in ‘producer surplus’[vi] – represented by the red rectangle in Figure 1 – that subsequently becomes ‘consumer surplus’ in Figure 2. However, this $160 wouldn’t represent an allocative efficiency gain. It would instead be a bare transfer of wealth. More specifically:

• it represents additional welfare that consumers would obtain from the reduced price that they would pay for all the units they would have bought anyway, i.e., the 40 units that would also have been bought at the monopoly price; and
• it would come entirely at the expense of the monopoly producer, i.e., the firm’s profit (its producer surplus) would be $160 lower than it’d otherwise have been if it’d sold those 40 units at the monopoly price. 

In other words, the $160 isn’t additional welfare that didn’t exist previously. Rather, it’s a bare transfer of current wealth, and welfare neutral. All that matters to the assessment of allocative efficiency is the units that the customer would not otherwise have consumed, but now does, i.e., the 40 units of previously unserved demand in Figure 1. Only through reducing this previously unmet demand is it possible to ‘make someone better off, without making someone else equally worse off’.

It would therefore be a mistake to characterise the potential economic benefit from setting a regulated price of $2 as being equal to the reduction in deadweight loss ($80) plus the transfer from producers to consumers ($160), i.e., as the total increase in consumer welfare. This would imply inaccurately that the potential benefits of regulation were $240 when, in fact, that is three times higher than the true achievable economic efficiency benefit, i.e., the potential $80 reduction in deadweight loss (the yellow triangle in Figure 1).

Wealth transfers aren’t benefits

Although a consequence of intervening in a market may be a transfer of wealth between producers and consumers (and/or between different types of consumers), it doesn’t follow that these transfers represent benefits of said intervention. Counting wealth transfers as benefits when deciding whether to intervene in markets risks regulation being introduced that makes New Zealanders worse off, i.e., where the true efficiency benefits are outweighed by the costs.[vii] The New Zealand Treasury has highlighted this risk in the past:[viii]  

“…including distribution effects in a cost benefit analysis could justify regulation where there is an inefficient outcome, but offsetting wealth transfers.”

[…]

“Treasury considers that the assessment framework should only include efficiency gains, as these gains are the actual benefit to New Zealand ...”

Even if one believed that, in theory, an extra $1 in the hands of consumers – or some subset of them (e.g., low-income or vulnerable consumers) – was somehow ‘more valuable’ than $1 in the hands of producers, or some other group of consumers, this isn’t an easy distinction to draw in practice. It’s not possible to crisply differentiate between these groups of people when predicting the overall effects of introducing regulation. For example, attempting to isolate and provide greater weight to consumer surplus is complicated by the fact that every ‘consumer’ may also be a ‘producer’ in some capacity. For example:

• some of the businesses being regulated might be government-owned, or owned by local councils – in which case the ‘end-customers’ will also be the ‘ultimate’ shareholders, i.e., any reduction in prices would also lead to a reduction in returns/dividends;
• those businesses that are publicly listed will have consumers who hold shares – either directly or through financial institutions tasked with investing financial assets (e.g., via superannuation funds) – these people may also be effected in their capacities as ‘producers’ by any decision to regulate; and
• the businesses in question might employ a non-trivial proportion of New Zealand’s workforce, and so any effect that the decision to regulate has on worker compensation and conditions, etc., would naturally affect those employees in their capacities as ‘producers’ as well.

For every $1 that is gained by a person in her capacity as a ‘consumer’ of a product following regulation, that person may lose more (or less) than $1 in her capacity as a shareholder, rate/taxpayer, trust recipient or employee. As we noted above, this means that an approach that counted all potential transfers to ‘consumers’ (or even a sub-set of consumers) as potential benefits of regulation might conclude that they’d be better off, even though many of them may be indifferent or worse off. In any event, there are arguably more efficient vehicles to redistribute income than through regulatory policy, as the New Zealand Treasury has observed:[ix]

 “We consider regulation is best used to improve the efficiency of markets. The Government has other policy instruments to address concerns about distribution of income.”

I therefore think the principal policy objective of regulatory interventions should be to maximise total economic welfare by promoting efficient resource allocation, i.e., by focusing on reducing or eliminating deadweight loss (including over time, thereby promoting both allocative and dynamic efficiency). That surplus can then be redistributed through governmental mechanisms such as taxation policies and targeted subsidy schemes to give effect to any desired distributional outcomes. In this way, wealth redistribution would be taking place using the most effective policy tools available.   

When it goes wrong…

Things can go badly awry when these basic principles are misunderstood or ignored. As I foreshadowed earlier, a prominent recent example was the cost-benefit analysis (CBA) that accompanied the EA’s proposed reforms to the TPM late last year. This analysis suggested that net benefits of $2.7b (in present value terms over a 30-year period) could be achieved by reforming the TPM. Regrettably, this analysis was profoundly flawed. Most of the alleged benefits (~95%) were said to come from ‘improved grid use’. In short, the theory was:

• reforming the TPM would lead to an increase in demand and a very large upsurge in generation investment ($1.9b, in NPV terms);
• that influx of new generation (i.e., new sources of supply) would drive down prices by a substantial amount; and 
• those price reductions would increase consumer surplus to the tune of ~$2.6b, which would constitute a benefit from ‘improved grid use’.

There were three key problems with this chain of logic:

1. the notion that generators would be stampeding to invest despite falling prices and plummeting margins was economically nonsensical;
2. the cost of that new (largely unprofitable!) generation - $1.9b – was left out of the CBA, despite it being pivotal to the achievement of the supposed benefits; and
3. most relevantly for present purposes, nearly all of the estimated increase in consumer surplus was a bare wealth transfer, i.e., not an efficiency gain. 

The CBA conflated efficiency gains and transfers because it was mistakenly assumed that every $1 of additional consumer surplus arising from forecast price reductions constituted an economic benefit from ‘more efficient grid use’. This overlooked the fact that, if final consumers were paying less then, tautologically, generators would be receiving less. Unsurprisingly then, almost all of the estimated ‘benefit’ arose from final customers paying less for electricity they would have bought anyway at the higher price, i.e., from transfers.

Put another way, the CBA inadvertently viewed any increase in consumer surplus as virtuous, regardless of whether it was new wealth (i.e., arising from new demand), or a bare transfer of existing wealth – in this case, from generators. Materials released subsequently pursuant to the Official Information Act revealed that this flaw was swiftly identified by an external consultant who peer reviewed the modelling. He posed the following hypothetical:[x]

“Suppose all prices in the model decreased by $10/MWh, for all consumers, at all times, in all years. Suppose further that demand was perfectly inelastic and so there was no corresponding increase in quantity.

Is there an allocative efficiency gain? Clearly not, with no elasticity and no changes in Q. The net economic benefit is nil.

Is there an increase in the … consumer welfare measure? There sure is – billions of dollars PV!”

The point being made here – very effectively – was the model was generating billions of dollars in ‘benefits’, even when no new transactions were taking place. This was a logical impossibility, since paying different amounts for the same quantity of goods and services can only shift existing wealth around – it can’t produce any new wealth. Efficiency gains require an increase in quantity, not just a reduction in price. Unfortunately, this feedback wasn’t reflected in the CBA which continued to overlook the fact that, if wholesale prices fell, generators would then have less money to:

• invest and innovate in ways that could lead to dynamic efficiency improvements (and, potentially, lower prices) over time;
• pay their employees – all of whom would inevitably be final consumers; and
• distribute as dividends – remembering that three of the larger generators (Meridian, Genesis and Mercury) are still majority-owned by taxpayers, i.e., final consumers. 

Failing to properly differentiate between efficiency gains and transfers served to inflate the estimated net benefit by more than $2b. This sum, whilst enormous, wasn’t unexpected. Because triangles tend to be smaller than rectangles (at least in this context), the transfer component of a change in consumer surplus is usually much bigger than the reduction in deadweight loss. This means that, when these errors are made, they can have an enormous adverse effect upon the analysis – as was the case here. It’s therefore essential to understand and appreciate this crucial distinction.     

Hayden Green | Director, Axiom Economics | 26 July 2020

[i]   Electricity Authority, 2019 issues paper Transmission pricing review, Consultation paper, 23 July 2019, pp.20-55.

[ii]  See for example: Axiom Economics, Economic review of transmission pricing review consultation paper, A report for Transpower, September 2019, pp.89-93.

[iii] While the producer may have an incentive to reduce the cost of supply and increase profits, it doesn’t face a penalty in terms of loss of custom to competitors if it supplies a given level of output at a higher unit cost than that which is attainable. The difference between the actual and minimum attainable supply cost is referred to as ‘X-inefficiency’. In the Collins dictionary of Economics, the authors note: “X-inefficiency is likely to be present in large organisations which lack effective competition ‘to keep them on their toes.’” See: Pass, C and Lowes, B, 1993, Collins Dictionary of Economics: Second Edition, Harper Collins, Great Britain, p.568.

[iv]  This is where its “marginal revenue” (MR) equals its “marginal cost” (MC), i.e., where the additional revenue it makes from the sale of one more unit is equal to the additional cost it incurs in producing that unit. If the revenue that the monopolist would earn from selling one more unit exceeds the cost of producing it, then it’s better off expanding its output and producing that additional unit. Similarly, if the revenue that the monopolist earned from selling its last unit was less than the cost of producing it, then it’d increase its profits by cutting back its production and selling fewer units.

[v]   At a price of $6 there are still ‘infra-marginal’ consumers who would have been prepared to pay more for the product, i.e., who derive more than $6 in private benefits. This margin is known as “consumer surplus” and is represented by the blue triangle.

[vi]  At a price of $6 there are infra-marginal units that would have cost less than $6 each to supply, enabling the monopolist to make an “economic profit” on each unit sold. This margin is known as “producer surplus” and is represented by the red rectangle. 

[vii] It’s also worth noting that imposing regulation is not a costless exercise, which will compound existing inefficiencies. 

[viii] The Treasury, Treasury Report No T2004/774: Briefing for EDC Local Loop Unbundling and Fixed PDN in New Zealand, 10 May 2004, pp.18 and 7.

[ix]  Op. cit., p.8.

[x]   Email from [Peer Reviewer] to [Electricity Authority], Re: Wealth transfers in the TPM CBA, 20 March 2019 (names redacted).