Application of the public benefit test to the Air New Zealand/Qantas case.
The proposed cartel between Air New Zealand and Qantas presented an interesting and important application of the public benefit test in action in two jurisdictions. The case began on December 9, 2002, when the two airlines jointly filed applications with the NZ Commerce Commission and the Australian Competition and Consumer Commission seeking authorisation for Qantas to buy a 22.5% shareholding in Air New Zealand for $550 million (1), and that the pair be allowed to form a 'strategic alliance' controlling just about all aspects of the airlines' activities on flights into, out of and within New Zealand, including scheduling, capacity and pricing.
The 'alliance' or cartel, as it will be called here, was firmly rejected by both Commissions in 2003, whereupon the matter went to appeal. The outcome of this process was in effect sealed by the judgement of the NZ High Court on September 17, 2004, rejecting the airlines' appeal against the NZCC's decision. However, subsequent to this, the Australian Competition Tribunal (ACT), which under the Australian procedures had reheard the case, announced that it had found for the airlines, though without, to date, offering any reasons.
The case has been extremely time-consuming and expensive (2), and, given the High Court's judgement and the airlines' acceptance of this, it would probably have been much better if they had never put the proposal forward in the first place. (3) Nevertheless, there may be useful insights and lessons to be gleaned from an analysis of the submissions, determinations and judgements that the case has generated.
There are two major issues, of which this paper will focus on the second. The first issue is the impact on competition of the cartel. Given that the two airlines singly or jointly have market shares ranging from above 80% to 100% on the major markets affected by the cartel (domestic NZ, trans-Tasman, NZ-USA) it might naturally be expected that eliminating actual or potential competition between them would almost inevitably result in a substantial lessening of competition (SLC). However, with respect in particular to the trans-Tasman market, which is key to the Australian hearing of the case (because it is the only one of the directly affected markets in which a large proportion of consumers are Australians), and which is already supplied by a fringe of '5th Freedom' carriers plus the budget airline Virgin Blue (trading here as Pacific Blue), the airlines mounted surprisingly vigorous arguments to the effect that the cartel would not in fact result in an SLC.
It may turn out, when the reasons are at last published, that these arguments carried the day before the ACT, even though this was insufficient to upset the virtual veto power of the NZ High Court's judgement. Be this as it may, I will here broadly concur with the NZCC's finding, which was not overturned by the NZ High Court, that competition in all the affected markets would be substantially lessened, and will use a formal spreadsheet-based oligopoly model to produce quantitative estimates of the likely increases in market prices and the associated changes in outputs and market shares. These estimates will enable us to quantify and compare gains and losses resulting from the cartel, in the cause of carrying out the (net) public benefit test, which is the second major issue raised by the case, and the topic of this paper.
The application of the test in this case is affected by the rules used to define the test--in particular the choice between 'total surplus' and 'consumer surplus' standards--and by the trans-national extent of the parties affected, on both sides of the market. Price increases affect NZ, Australian and foreign consumers; the two 'Australian' airlines Qantas and Virgin Blue were each about one half foreign owned, and Qantas would have a substantial claim on the profit stream of Air New Zealand should the cartel go ahead. Broadly, most of the harm is suffered by New Zealand consumers; most of the gains go to Australian and foreign shareholders. Thus, the most favourable 'public' benefit test of the cartel results from the application of NZ or 'Kiwi Rules' to the Australian jurisdiction, and the most unfavourable from the application of these rules to NZ, whilst the consumer-oriented 'Aussie Rules' regime predicts harm in both countries. Section 2 of this paper presents these calculations.
These calculations are limited in two dimensions: firstly, to just the 'static' or allocative efficiency implications of price changes in terms of the (mis)allocation of resources generated by the raising of price (further) above marginal costs; and, secondly, to the direct impact on the air travel market. However, in this case much was made of possible 'productive' and/or 'dynamic' inefficiencies resulting from the impact on costs of a lessening of competitive pressure, as well as--in the other direction--claimed savings from the elimination of excess capacity which would be made possible by the cartel. And, in addition, claims were made and assessed of impacts in related upstream and downstream markets: for engineering and maintenance services and for inbound and outbound tourism. The cost savings and indirect effects will be assessed in section 3 of the paper. (4)
Section 4 asks if the NZCC and High Court may have misdirected their efforts because of failure to correctly interpret the 2001 'Purpose Clause' (s 1A) modification to the Commerce Act. Section 5 then evaluates the Purpose Clause itself- in particular, the justification for a consumer welfare focus to competition policy. Section 6 concludes the paper.
Note that this case generated thousands of pages of submissions, reports and Determinations, and many thousand more of transcripts of oral testimony. In the present quite brief paper I do not intend to re-argue the case, nor even to comprehensively summarise it. Rather, the paper uses quantitative and other analysis to highlight certain issues raised by the proposed cartel and its treatment by the authorities; in particular with respect to the use of public benefit tests.
2. Restricted Public Benefit Tests of the Lessening of Competition
We will in this section carry out some public benefit tests of the proposed airlines' cartel, restricting ourselves to its direct, short-run repercussions on pricing and costs felt in the markets immediately affected by the arrangement. That is, the analytical framework for these tests is as set out in Oliver Williamson's classic (1968) article 'Economies as an Antitrust Defense': the allocative inefficiency loss from the immediate restriction of output consequent on a lessening of competition is weighed against (any) efficiencies--in particular, scale economies--made possible by the merger or arrangement through reallocations of existing resources that would not be possible otherwise. Long-run implications are not considered.
One of the many interesting features of the airlines case is that the parties sought authorisation without calling on scale economies as benefits. In most such cases, including in particular the largest successful efficiencies defence of a merger that would admittedly raise prices--the Canadian Propane case--forecast scale economies are central to claims that there would be a net public benefit. But in this case the possibility of economies was raised in the original NECG (2002) submission, but not quantified or further considered on the grounds that they were not likely to be substantial and might be just about balanced by costs incurred in implementing the cartel, in particular costs of integrating the two airlines' IT systems.
Without scale economies or similar 'synergies' it might well be wondered how the applicants believed they had any chance of getting their cartel approved, and that is a good question. Most of the answer lies in claimed Good Things that would happen (additional tourism) or bad things that would not happen (a wasteful 'war of attrition') if the cartel went ahead. These are discussed below in section 3. But even within the confines of the standard tracking of the gains and losses in the immediate market a public benefit test has a chance of succeeding. This is because, under the competition law of both Australia and New Zealand (as in other countries), the nationality of those benefiting or suffering from an arrangement matters. Specifically, the interests of foreigners are not taken into account, and, as noted in the Introduction above, foreigners are present on both sides of the air travel market, as airline shareholders and as consumers. This can make a significant difference, as will be seen.
2.1 Predicting the impact of the cartel
In order to carry out a quantitative public benefit analysis we need estimates of the impact of the cartel on prices and outputs in the affected markets. We will here restrict attention to the three large markets most obviously affected because they are currently served by both Air New Zealand and Qantas competing independently: trans-Tasman ; main trunk domestic NZ, and the Auckland-USA run. Other markets, such as domestic Australia, routes between Australia and/or NZ and the Pacific Islands, and routes between Australia and Los Angeles (from which Air New Zealand exited in 2003) were also considered by the Commissions.
The estimates used here are generated by a fairly standard non-cooperative oligopoly model implemented on an Excel spreadsheet. This model is based on Haugh and Hazledine (1999) and was reworked for the present case in Hazledine (2003) and in subsequent submissions made on behalf of Gullivers Pacific Ltd. A version of the model was adopted by the NZ Commerce Commission, recalibrated on a city-pair route basis, and used to generate results for their 'revised modelling ' that was submitted to the NZ High Court in July 2004. A feature of these models applied to airline markets is that it is assumed--with some justification from econometric evidence--that 'normal' competition in small-number oligopolies is reasonably well represented by Cournot-Nash behaviour, which within the more general framework of oligopoly models corresponds to a 'conjectural variations' parameter, summarising each firm's expected output change in response to a unit change in other firms' output, equal to zero. Positive CV parameters represent more cooperative behaviour, with a value of 1 representing a fully colluding duopoly or cartel. Values less than zero capture more aggressive behaviour, with -1 being the extreme case of perfectly competitive conjectures, when a firm puts output on the market in the belief that other firms will accommodate it, so that market price will not change.
Although the results of this model would not be and indeed were not deemed satisfactory by the airlines and their consultants, they are quite mainstream and unsurprising in that the result of losing independent competition between Air New Zealand and Qantas is predicted to be a fairly substantial lessening of competition as measured by price increases in the 10-20% range for passenger air travel. Results do differ across the markets, however, and it is appropriate to discuss them. Table 1 shows, for each of the three markets, the 'without' and 'with' cartel scenarios, which are known in Australia and New Zealand as the 'counterfactual' and 'factual' situations (CF, F). In each market, my model is scaled to an actual 2004 situation with total market output = 1000 and market price = 1. For the domestic NZ and Auckland-Los Angeles market I judged that the counterfactual would be similar to the current observed situation--that is, it would be 'business as usual' for Air New Zealand and Qantas if the cartel is not permitted. (5) However, such may not be the case for the trans-Tasman routes, on which competition is currently believed to be unusually fierce. In my modeling, I assumed that, with the cartel definitely not a possibility, the two major competitors will allow prices and profit margins to drift up somewhat to restore 'normal' Cournot-Nash conditions. we look at each market in turn.
NZ main trunk market
The main trunk market between the three largest cities (plus Queenstown) is dominated by Air New Zealand and Qantas, who operate the only jet service. Air New Zealand has a much larger market share than Qantas, which admits to losing money on its New Zealand operations, and I have calibrated the situation as somewhat more competitive than Cournot-Nash, which is achieved in these models by specifying a conjectural variation (CV) parameter less than zero. I assume that Qantas will continue in New Zealand on these terms absent the cartel, though it must be a possibility that in the counterfactual it will choose to cut its losses and exit. (6) I assume that the cartel would hand over all its domestic NZ business to be flown by Air New Zealand as the lower cost carrier on these routes, and that Virgin (Pacific) Blue would then enter, and take around 20-25% of the market, despite which prices would rise by 18% with the elimination of competition between the incumbents. The extent of the price rise is affected by the market elasticity of demand, which on these routes is estimated to be relatively low because of the relatively high proportion of business travelers.
This market includes flights between the NZ cities of Auckland, Wellington and Christchurch, and the Australian cities of Melbourne, Sydney and Brisbane, as well as some other routes flown by Air New Zealand's subsidiary Freedom Air. Prices on these routes are currently extremely low, and have been modeled by assuming that the current 2004 benchmark price (= 1) is being generated by Air New Zealand and Qantas competing with a CV parameter equal to -0.5--halfway between Cournot and perfectly competitive behaviour. Such is probably not sustainable and in the simulations I assume that the counterfactual would see the restoration of normal or Cournot-Nash behaviour, which means that prices would be 15% higher than they are now.
Even so, the cartel would raise prices further by about 15-16%, assuming no change in the output of the 'entrant' (Pacific Blue), nor of the group of large international carriers which currently exercise '5th Freedom' rights to carry passengers across the Tasman on planes which are going on to or coming from long-haul flights.
Auckland-Los Angeles market
Since the exit of United Airlines, this route has been completely dominated by Air New Zealand and Qantas, who operate what may be described as a 'cosy duopoly' manifested in a diffidence about increasing capacity on the route, despite what seem to be quite high load factors. This is represented in the model by a positive CV parameter in the counterfactual, the implication of which is a smaller price increase under the cartel, of 12 1/2 %.
2.2 Short-run Public Benefits from the cartel; Results
Now we add up the gains and losses of the cartel's price-increasing effects from four different perspectives: NZ welfare, Australian welfare, Australasian welfare and 'world' welfare or total efficiency. We do the NZ and Australian cases under two sets of roles: the 'kiwi rules' under which only the nationality of those affected matters, and the 'aussie rules' under which both nationality and identity matter, specifically, it is consumer interests that are paramount.
The analysis requires calculations of changes in consumer and producer surplus, using data from the oligopoly market model as summarized on Table 1 above, with these weighted according to the nationality of those affected, given that the two Australian airlines were in fact about 50% foreign-owned, and that NZ, Australian and 'foreign' travelers consume the airlines' products in differing proportions, as shown on Table 1. The result are shown on Table 2.
Overall, the majority of consumers affected by the cartel on these three routes would be New Zealanders, and their total loss in consumer surplus is calculated to be $176.5 million/year, which is the measure of detriment to NZ under Australian rules. It would normally be expected that increases in profits to Air New Zealand would compensate for much of the loss in consumer surplus under kiwi rules, but such is not here calculated to be the case.
This is largely because of a little-analysed feature of the proposed alliance between the two airlines, which is the profit-sharing formula whereby, in effect, 20% of the profits earned from a flight operated by one airline are handed over to the other. Since this formula only applies to flights which take-off and/or land in New Zealand, of which a majority are operated by Air New Zealand, and since Air New Zealand is calculated to have somewhat lower operating costs and thus higher profit margins than Qantas on these routes (7), the formula must inevitably result in a net transfer from the NZ carrier to Qantas.
This transfer is in addition to any transfer resulting from Qantas taking a 22.5% shareholding in Air New Zealand, and thus a 22.5% claim on the profit stream, net of the value to Air New Zealand of the $550 million paid to secure that shareholding, which here I pro-rate to the different markets according to the revenues generated on those markets as a proportion of Air New Zealand's total revenues, and annualize by assuming a cost of capital of 10%. (8)
While the idea of a profit-sharing formula makes sense as an incentive for the airlines to make the most efficient use of their capacity in the cartel--notably, I expect, by pulling Qantas planes out of the loss-making domestic NZ market--it appears to have very substantial quantitative implications. Specifically, in my calculations, Air New Zealand comes out as a net loser from the cartel, to the extent of $76.7 million/year! This number is the net of an even larger transfer to Qantas from the long-haul Auckland-LAX route and two quite small numbers--one positive, one negative--for the change in Air New Zealand's retained profits on the Tasman and domestic NZ routes.
It is possible that the profit flows imputed by the model to the LAX route are too high, in that they may not, for example, allow for higher on-the-ground costs associated with operating this route. On the other hand, the calculations shown here do not count other routes affected by the formula, in particular the NZ-Asia routes on which Qantas does not fly, and so will obviously be a substantial gainer from the formula.
Profit transfers between Air New Zealand and Qantas must be considered potentially quantitatively significant in a public benefit calculus setting, simply because hundreds of millions of dollars are at stake. Yet my efforts in submissions to bring this issue into the debate on the cartel were totally unsuccessful. The NZCC in its Final Determination announced that
'Regarding the treatment of Qantas's profits from its 22.5% holding in Air NZ in the factual, the Commission has formed a view based on Professor Willig's submissions [for the applicants] that it is inconsistent to treat the profits due to Qantas from its equity stake in Air NZ as a detriment and yet to ignore the initial capital injection. Accordingly, the Commission's revised approach in this respect is that the share of Air NZ's profits due to the Qantas equity stake is not to be counted as a loss to New Zealand in welfare terms and that the initial capital injection is ignored.' (NZCC, para 1055)
Not unreasonably, then, given the lack of disagreement, the High Court did not challenge the NZCC on this matter when considering a related issue of consideration of fixed costs incurred in the process of implementing the cartel (High Court, para 316).
But it is difficult to see how the Commission could agree with the applicants on the profits transfers issue while disagreeing--as of course they did--on the extent of the lessening of competition that would result from the cartel. Higher price increases means larger flows of transfers from consumers to producers, and the identity of the ultimate recipients of those flows must matter to the national net public benefit calculation, just as, in their procedures, the identity of consumers matters.
That said, the prediction of the calculations shown in this paper that Air New Zealand would actually lose profits from the cartel, despite higher prices, because of the transfers to Qantas is worrying. Why would a firm voluntarily enter into a transaction from which it will make a loss? There are several possible reasons for the analysis to generate a prediction of loss to Air New Zealand:
* incorrect values for the parameters of the profit-sharing formula
* inadequate allowance for cost savings generated by the cartel
* mistakes made by the Air New Zealand negotiators
* the counterfactual would be even worse
While I would not be surprised if the skimpy information available from the airlines on the execution of profit sharing resulted in my calculations containing errors in parameters (perhaps especially on the Auckland-Los Angeles route), it seems almost inevitable that the process would have resulted in a net transfer of money from the New Zealand airline to Qantas, given the unequal market shares of the two carriers on the affected routes, and it must be asked what Air New Zealand expected to get in return, given that even as claimed--and a fortiori as determined--cost savings or efficiencies do not compensate.
Unfortunately, the possibility of the Air New Zealand management and directors being out-negotiated by their Australian counterparts cannot be ruled out, given that many of these people were involved in the near-disastrous buyout of Ansett Australia just a couple of years before.
As for the scenario that the counterfactual would involve even larger losses in profit flows for Air New Zealand: this might be linked to the threatened 'war of attrition' which was posited by the applicants. Given that the Commissions and the High Court did not find the plausibility of this scenario compelling (see Section 3 below), they, arguably, should have been more receptive to taking seriously the possibility of losses to Air New Zealand.
Now look at the cartel from the Australian perspective. If the NZ Commerce Commission were contracted to hear the case in Australia, then as far as the short-term public benefit test is concerned the cartel might well be authorised. Australian consumers lose, but there aren't so many of them, especially on the domestic NZ and Auckland-LAX routes, and the higher prices plus operation of the profit-sharing formula in Qantas's favour results in net benefits to Australian shareholders calculated here to be about double the loss in consumer surplus.
If anti-trust in Australia and New Zealand was harmonised, such that both countries adopted the total surplus approach and did not distinguish between each other's nationals in applying it, then the cartel would fail the public benefit test for authorisation: the combined losses in consumer surplus considerably outweigh gains in profits, once the transfers between Air New Zealand and Qantas are netted out.
Even an integrated trans-Tasman public benefit test is not a true efficiency test, because it is affected by transfers to foreigners. If these too are netted out, we are left with just the deadweight loss or allocative inefficiency cost of the cartel, which here is put at about $150 million/year. The reason this figure is smaller than the $187 million net detriment found for Australia and New Zealand combined is that proposed cartel would on balance benefit foreigners, who would gain more from their shareholdings in Qantas and Virgin Blue than their citizens would lose from higher airfares.
2. Extended public benefit calculations
A fully articulated public benefit test may involve extending the Williamson efficiencies defense framework in two dimensions, in an attempt to take account of all relevant benefits and detriments associated with the proposed arrangement. First, account may be taken of possible feedbacks over time on the market cost and/or demand curves as a result of the lessening of competition. And second, indirect impacts on other industries or markets may be considered.
The Air New Zealand/Qantas case was unusually rich in both types of extensions, claimed both for and against the proposed cartel. They included:
During the case, the applicants submitted that their cartel would enable the realisation of quite large 'cost savings', of around $100million/year. Unusually, these were not savings of currently incurred costs, such as are often claimed for scale economies or 'synergies', but rather were costs saved in the cartel 'factual' scenario relative to costs that it was claimed would be incurred in the 'counterfactual' should the cartel be disallowed.
Specifically, the airlines warned that a vicious 'war of attrition' would be unleashed, should their plans be thwarted. Qantas, in particular, would add considerable additional capacity to the Tasman and domestic NZ markets. The additional seats would not actually be used to carry passengers (no change in price was assumed), and so would represent wasteful excess capacity, the cost of which would be 'saved' if the cartel were approved.
The war-of-attrition scenario attracted considerable criticism--even, scorn--at the submission and hearing stages of the process, and was not given much weight by the Commissions. The basic implausibility--because lacking a sound commercial motive--of the scenario--is born out by the airlines' actual behaviour as the counterfactual has unfolded since the High Court's decision in September 2004, which appears to be quite relaxed and even cooperative to each other.
Productive and Dynamic Inefficiencies
On the other side of the ledger, the NZCC determined that the slackening off in competitive pressure permitted by the cartel would result in increases in operating costs, which it interpreted as inefficiencies (rather than as transfers to workers and management of 'quiet life' benefits). They distinguished between changes in the level of costs (productive or 'X'-inefficiencies linked to lower incentives to avoid waste) and changes in the rate of growth of costs (dynamic inefficiencies through loss of incentive to innovate), bravely attempted quantification of these and determined that they would be substantial.
Productivity analysts decompose changes in cost efficiency into shifts in the best-practice frontier and changes in the distance of actual practice from that frontier. In the long run, improvements in technology drive shifts in the frontier. In the case of the airline industry, such improvements will be generated by the efforts of the airlines themselves and their equipment suppliers. It would seem unlikely that the merger or alliance of just two of the world's several hundred scheduled airlines would have any significant impact on technological change (9), but it could indeed affect the keenness with which the two airlines press to keep their own practices on or close to the frontier.
However, there is a problem here with the notion of dynamic inefficiency, because it implies that the gap between actual and best practice will widen without limit over time. If this were so, for example, given the long decades of protection enjoyed by both Air New Zealand and Qantas before privatisation and deregulation, we might therefore expect the legacy of this to be that they were both now still flying Lockheed Electra prop planes! The NZCC deal with (or avoid) this problem by producing these and other numbers on 'year-3' basis, which is sensible in itself but perhaps would better justify not attempting to maintain a sharp distinction between 'productive' and 'dynamic' effects of lessenings of competition.
As for quantification, this is admittedly very difficult at the level of a particular market and industry. Of course it is one of the biggest ideas in Economics, and one that it is absolutely crucial to anti-trust, that competition fosters lower production costs and faster technical progress. This idea is supported by a mass of argument and evidence dating back at least to Adam Smith, and including many econometric studies, including some focusing on the airline industry which were brought forward as evidence in this case. However, one can believe, as I do, that the cumulative weight of this learning amply justifies competition-based anti-trust, whilst being sceptical that plucking particular case studies from the literature can always serve to reliably calibrate quantitative predictions of the impact on a particular firm or firms' costs of a particular structural change or arrangement.
The applicants claimed annual benefits in the $100 million range from increased tourism attracted by improved promotional efforts made possible by the cartel. These claims were heavily criticised in the proceedings. It seemed highly implausible to me and others that profitable opportunities to encourage inbound tourism to New Zealand were being neglected due to a market failure that would be corrected by monopolising air travel on most of the relevant routes. The NZCC mounted a particularly effective de-construction of the applicants' analysis, which revealed that numbers of foreign tourists were predicted to actually drop (because of higher prices), but that this would be more than made up by increased numbers of NZers choosing to holiday at home (again, because of the higher prices). The High Court broadly accepted the Commission's findings.
Overall, the treatment in the airlines case of long-run and related-market impacts of the proposed cartel does not seem to have generated any significant injustices. But should the exercise be carried out at all'? It is not being critical of Commissions and Courts to suggest that they really do not have the expertise to predict and quantify long-run and spillover effects from a lessening of competition in a particular market, because even the 'experts' may not have the competence to do this satisfactorily. If so, then this is an endemic problem of the full-blown public benefit test which of course a traditional SLC benchmark is not subject to.
3. The purpose of the Purpose Clause
In 2001 parliament modified the Commerce Act by adding a Section 1A:
"The purpose of this Act is to promote competition in markets for the long-term benefit of consumers within New Zealand."
John Land, in particular, has argued (2003 and before the High Court in this case) that parliament's intent here is to direct the authorities and courts to give more, perhaps full, weight to the interests of consumers rather than producers in implementing a Public Benefit test. However, in this case the airlines (not surprisingly) argued against such an interpretation, and so too did the NZ Commerce Commission and the Court.
It might seem natural--even, unavoidable--to interpret Section 1A as being based on the following two propositions:
1. That Parliament in this matter is concerned above all else with the welfare of consumers;
2. That consumer welfare is benefited by the protection or promotion of competition
But compare the NZCC's gloss on this:
"[T]he higher airfares result in a transfer of income from consumers, who pay the higher fares, to the airlines, who receive them as higher profits. These transfers are not generally counted as a detriment by the Commission. However, the purpose of the Commerce Act is to promote competition for the benefit of New Zealand consumers. Consequently [sic], where such transfers are from foreign consumers to a New Zealand supplier, such as Air NZ, they are considered to be a benefit. Similarly, where New Zealand travellers would pay a higher fare to an overseas-based supplier, such as Qantas, that is considered a detriment. (2003, ix, para 34).
That 'consequently' is breaktaking in its blitheness. Instead of the sentence reading, as one might expect or, perhaps, hope: 'Consequently, the Commission recognises that it is and has been wrong in not counting transfers from consumers as a detriment, and it is very sorry, and promises never to do this again', the Commission ignores the injunctions to promote competition and to focus on consumers, and picks up only on the modifier 'New Zealand' so as to justify its dismissal of foreign interests--thereby in effect undermining the sole uncontroversially honourable goal of a (transfer-neutral) public benefit test, which is to promote economic efficiency!
How did the High Court deal with this issue when it was forced to do so by the submission of John Land for Gullivers? After briefly paraphrasing the appellants' and the Commission's reasons for rejection of Land's position, the Court decided:
'We are satisfied that the introduction of s 1A should not disturb the Commission's established practice of treating as neutral any wealth transfers between New Zealand consumers and producers. Determination of authorisation applications under the Act are properly concerned with balancing any efficiency detriments associated with breaches of the statutory competitive standard, against any efficiency gains that may result from the business acquisition or contractual arrangement in question. It is the balancing of these real resource impacts on the economy that best serves the long-term interests of consumers. The inclusion of ad hoc [sic] wealth transfers, which are not losses to society, would distort [sic] the efficiency assessment by assuming additional economic harm to the public of New Zealand. In any event, consumers might well be the ultimate beneficiaries [eg as shareholders or taxpayers]. (2004, para 241)
Again, the impartial commentator might be taken aback. Note first the dismissive 'ad hoc' applied to what are in fact systematic transfers of income from consumers through higher prices resulting from monopolisation of markets. And then take the presumption that any consumer who would consider their interests harmed by such higher prices is myopically mistaken. Could this really be what Parliament believed when it added section 1A? And, even if it did, the Court is simply wrong in assuming the Commission's implementation of a public benefit test to be solely efficiency-based. This could only be generally correct if no foreigners were involved. Instead, in an export-oriented industry such as airlines, it is entirely possible that a price-increasing arrangement that was admitted to generate large inefficiencies could nevertheless be authorised under the Commission's procedures because the costs of these inefficiencies were exceeded by additional revenues extracted from foreign tourists and business travellers. (10)
The contrast with the Australian authorities' position is striking. Under what I dubbed 'aussie rules' in section 2, above, only the consumers' interests are counted. That is a bit simplistic, but not much. Here is the ACCC's determination:
"[A]ny weight to be attributed to the public benefits [from claimed cost efficiencies] is lessened by the fact that benefits accrue to the Applicants and their shareholders rather than consumers in an environment where there is reduced competition." (2003b, para 13.276)
And the list of possible benefits admitted by the ACCC does not include the transfers from consumers to producers through higher prices, which are, however, fully counted as detriments. The ACCC thus did not at all resile from the position that it stated quite bluntly in its Draft Determination:
"While the Commission is of the view that benefits to a particular group or segment of the community may be regarded as benefits to the public, consideration needs to be given as to whether the community [sic] has an interest in that group being benefited and whether that benefit is at the expense of others, for example, consumers through higher prices. Where benefits are not passed on to consumers they are likely to be accorded a lower weight by the Commission." (2003a, page C.12)
To be quite clear about this: the ACCC is not suggesting that benefits 'at the expense of others'--ie, transfers--should be netted out, as is the authorities' current practice in New Zealand. Rather, having elsewhere counted higher prices as detriments, they are reluctant to allow any compensating virtue in the higher profits thus pocketed by the firms. The ACCC in its Draft did note the airlines' consultants' arguments for the appropriateness and indeed legitimacy of the total surplus (net transfers) approach. But, and with a blitheness which in its own way is just as startling as the NZCC's esoteric interpretation of the s 1A Purpose Clause, they do not bother to even confront or argue against the total surplus argument, simply re-asserting in the final paragraph of their Draft Determination that:
"The distribution of welfare gains and losses is of relevance to the Commission's consideration of the Application." (p2003a, p. C. 15)
It is interesting that the ACCC believes it should maintain this position in a legislative environment which is (I believe) quite similar to New Zealand in terms of specifying public benefit authorisations, and which lacks the comfort (as I would interpret it) of the NZ purpose clause singling out the primacy of the consumer interest. Of course, it may be that the ACCC will be rebuked for its interpretation by the Australian Competition Tribunal, when the latter finally makes public its reasons for allowing the airlines' appeal in that country.
Overall, it seems reasonable to conclude that the New Zealand antitrust authorities have been intellectually 'captured' by the proponents of what I called (1998) the 'rationalist' (total surplus) public benefit test, to an extent which, at least since 2001, is arguably out of line with the wishes of the legislators. Be that as it may or may not, the important question remains as to the desirability of a total surplus-based authorisation of an anticompetitive practice: to that we turn in the next section.
5. The purpose of anti-trust
Bertram (2004), Hazledine (1998) and many others have argued for the continued legitimacy of the traditional view underpinning anti-trust, that the long-run health and integrity of capitalist market economies depends on the markets continually pressuring the capitalists to keep their prices close to their costs. Such remains the mainstream position of the commissions and courts if not the economists in just about all jurisdictions with the exception of New Zealand and, perhaps now, Australia.
There are two sides to a divergence of price and (marginal) cost: the customer pays more than the social opportunity cost of producing the good or service, and the firm receives more profit than is justified by the resources committed. These map into major issues of justice and efficiency. As Bertram emphasizes, the extraction of monopoly profits from consumers can be seen as confiscatory--in effect, a tax. Modern societies tend to reserve to the state the sole right of taxation without return and may also sustain a core belief that the 'weak'--consumers, households, small businesses--should be able to engage in the marketplace with the strong--the great corporations and oligopolies--on reasonably equal terms.
Such beliefs have the efficiency rationale that invention, innovation, creativity--the dynamism and vitality of an economy--depend largely on the contributions of diverse individuals and small enterprises, whose ability to participate in markets must therefore be safeguarded. More generally, economists worry about the incentive implications of non-competitive market structures. The point of keeping pressure on price-cost margins is so that the profit motive is guided towards reducing costs and/or producing more attractive products, rather than towards raising prices and investing in positions which make price-raising easier, as the path to increased profits. These are the 'productive' and 'dynamic' efficiency implications of lessenings of competition.
This Big Idea is operationalised at the policy level by economic theories and models of the drivers of the degree of pressure on profit margins, of which the most important--in this case as probably in most others--is the proposition that the numbers of actual--and to a lesser extent potential--capable competitors in a market is key to determining whether a proposed practice or merger between competitors would result in a substantial lessening of competition.
Now, the total surplus approach pushed by the New Zealand authorities obviously does not truck with any notions of justice. In essence, they replace 'one man one vote' with 'one dollar one vote'. But the long-run efficiency implications of lessening of competition have been taken very seriously. Indeed, they may in the end have determined the outcome of the case. The High Court concluded:
"We have accepted that the Commission was entitled to find that allocative, productive and dynamic inefficiencies would be likely to result from the proposed Alliance. We have rejected the quantification of losses it attributed to allocative inefficiencies and expressed reservations about the figure it put on productive inefficiencies. But we are satisfied that the Commission was entitled to find that substantial detriments are likely to arise in both categories and have been fairly quantified in the case of dynamic inefficiencies." (2004, para 427)
Having determined that offsetting public benefits are 'modest', the Court's last words were: 'The appeal is dismissed'.
In a mainstream competition-based competition policy, these issues do not arise. The determination of the likelihood or not of an SLC is all that is required for the authorisation process. No other benefits or detriments of the proposed arrangement need be quantified or even considered and indeed the harm done by the SLC need not itself be quantified. Note that this does not entail as proponents of the total surplus approach are prone to claim--making a value judgement favouring a dollar to one group (consumers) over a dollar to another group (capitalists). It is about where the dollar comes from, not who gets it.
SLC anti-trust is thus a rule- rather than discretion-based policy, like much of the legal and informal institutional/ethical system which civilised societies choose to let constrain their citizens' actions. As such it is not justified by a belief that in every case the application of the rule will necessarily lead to an 'optimal' outcome. Rather, the underpinning is the belief that application in every case--ie, without exception--results in an overall set of outcomes to be preferred to any other attainable set. As noted above, the underpinnings of this are ideas of natural justice dating back to the Enlightenment and beyond--that the strong (big firms) should not be free to prey on the weak (consumers, households, small business)--bolstered by newer learning in Economics illuminating the importance of incentives, rent seeking and transaction costs.
The airlines' application to form an anti-competitive cartel was turned down by the NZ Commerce Commission, supported on appeal by the High Court. The perspective of the present paper could be summarised as follows: The Commission and the Court made the right decision, for (in part) the right reason, but with the wrong reasoning.
The reason, in essence, is that a lessening of competition is likely to generate inefficiencies, in particular long-run or 'dynamic' inefficiencies caused by relaxation of the competitive pressure that keeps prices close to costs. This is a good reason, and together with the "non-economic" reason that monopolies are unjust forms the basic bedrock on which competition law and policy is and always has been built.
But the reasoning in support of the NZCC and the Court follows their special interpretation of the public benefit test approach to competition cases as requiring quantification (wherever remotely possible) of all the benefits and detriments of the proposed arrangement within the framework of a total surplus standard. Such involves making long-range forecasts of events in particular markets which may go beyond the limits of competence of professional expertise. In this case, the airlines' lawyers and economists might have reason to feel somewhat aggrieved that, having successfully on Appeal done the difficult work of muddying the waters around the issues of the lessening of competition and associated allocative inefficiencies, they were yet undone by the High Court's willingness to support the NZCC's determination of what from an economist's perspective might seem to have been the weakest link in the chain of quantifications: viz, the range of numbers assigned to long-run inefficiency costs.
The airlines case, with its international spread, also quite vividly demonstrates the awkwardness of the public benefit test when nationalities matter. Sincere advocates of the total surplus approach--that is not those who are just supporting a vested interest--invariably base their case on efficiency: specifically, that when all the 'transfers' are netted out on a dollar-for-dollar basis, the residual is the measure of the net efficiency cost or benefit of the proposal. But such is simply not true when foreign interests are given no weight. The implementation of the test can quite literally become a crude beggar-thy-neighbour exercise, the outcome of which can differ dramatically depending on the jurisdiction within which the test is carried out, as shown in Section 2 of this paper.
How might matters be arranged better? One path could be to make the test truly that of total surplus by adopting a supra-national definition of the 'public' whose interests are of concern. No doubt there would be supporters for such a reform, and it is not hard to predict who they would be. The problem would still remain that the range of quantification and prediction required may often or even usually exceed the competence of even the most able commissions, courts and experts.
A simpler process, which in most if not all cases would lead to the same result, would be to adopt a consumer surplus rather than total surplus standard in the public benefit calculations. Such indeed is the most common approach outside of New Zealand, and--to a lay observer--seems in fact to be required of the NZ authorities by the Section 1A Purpose Clause of the NZ Commerce Act.
A consumer surplus standard counts as a detriment all increases in price. The beauty of this is that it is robust to the raison d'etre of competition policy: that is whether the harm of a price increase is that it involves an unfair redistribution of income to monopolists, or whether the problem is perceived as being that the income will not be redistributed, but rather will be wasted in higher costs (or some combination of the two concerns).
But why not simplify further, and just strip back competition law to its original core, which is the test of a substantial lessening of competition. That is, abandon all the clumsy and contentious apparatus of the listing and measuring of costs and benefits, and base policy on the traditional faith that an economic system in which practices or arrangements that would substantially lessen competition are not, ever, permitted will deliver better outcomes overall than a system in which SLCs may or may not be authorised.
Admittedly, a simple SLC rule could lead to 'wrong' decisions in specific cases, when loss of competition really would be compensated by an improvement in efficiency not otherwise achievable. But such instances seem in practice to be truly rare. Perhaps the only plausibly documented example is the Canadian Propane case. It does not require unreasonable faith in the vigour and ingenuity of agents in the market system to believe that, if required, they will be able to find ways of realising genuine synergies or scale economy benefits without resorting to anti-competitive practices.
Finally, the transaction costs implications of a simpler system are not trivial. Consider the airlines case. It failed, but not before chewing through somewhere between $50 and $100 million in fees and other disbursements, when the costs of all parties are considered. Under an SLC regime, the case would surely never have been brought forward, and those resources would have been available for other purposes. Even $50 million would fill quite a few 'Harberger Triangles' of welfare losses.
Australian Competition and Consumer Commission (2003), Applications for Authorisation A30220, A30221, A30222, A90862 and A90863, 9 September.
Bertram, Geoff (2004), 'What's Wrong With New Zealand's Public Benefit Test?' New Zealand Economic Papers 38 (2), December, 265-278.
Evans, Lewis (2004), 'The Efficiency Test Under Competition Law and Regulation in the Small Distant Open Economy that is New Zealand', New Zealand Economic Papers 38 (2), December, 241-264.
Haugh, David and Tim Hazledine (1999) 'Oligopoly Behaviour in the TransTasman Air Travel Market: The Case of Kiwi International,' New Zealand Economic Papers 33 (1), June, 1-25.
Hazledine, Tim (1998), 'Rationalism Rebuffed? Lessons from Modern Canadian and New Zealand Competition Policy,' The Review of Industrial Organisation, 13, 243-64.
Hazledine, Tim (2003), 'Proposed Alliance between Qantas and Air New Zealand,' Submission to the NZ Commerce Commission and the Australian Competition and Consumer Commission, February 14.
Land, John (2003), 'The Proposed Cartel Between Qantas and Air New Zealand: Why the Commission was too Soft', presented to the 14th Annual Workshop, Competition Law and Policy Institute of New Zealand Inc, Auckland, 23-24 August
[Land, John: Submission on Behalf of Gullivers Pacific Ltd to the NZ High Court, July 2004]
[NECG] Network Economics Consulting Group, 'Report on the Competitive Effects and Public Benefits Arising from the Proposed Alliance Between Qantas and Air New Zealand,' December 8, 2002.
NZ Commerce Commission (2003), [Air New Zealand and Qantas Authorisation] Final Determination, 23 October.
New Zealand High Court (2004), Air New Zealand and Qantas Airways Ltd v.Commerce Commission and Others,  3 NZLR 550.
Williamson, O.E., (1968), "Economies as an Antitrust Defense: The Welfare Tradeoffs", American Economic Review 58:18-36.
(1) All monetary values are in NZ dollars
(2) The airlines have probably spent around $50 million on lawyers and consultants on this case.
(3) It is conceivable that the airlines (or Air New Zealand in particular) used the process as a means of 'buying time' to get their own financial and operational affairs in order in an environment in which the uncertainty (of the outcome of the application) discouraged entry from other airlines.
(4) The competition commissions also considered cartel impacts on the supply of air freight services and on the retail and wholesale travel agency industry. These impacts were not claimed or assessed to be substantial and will not be covered in this paper.
(5) In much of the analysis and adjudication, attention focused on 'year-3', meaning the situation three years after the implementation or abandonment of the cartel. Given expected annual market growth of around 2% or more, numbers shown here would need to be grossed up to be on a year-3 basis,
(6) Qantas has only operated its own aircraft in New Zealand since 2001.
(7) There is some independent evidence that Air New Zealand's cost structure is lower than Qantas', partly because of a lower wage structure and also because Air New Zealand's trans-Tasman operations include their Low Cost Carrier Freedom Air. However the cost figures used in this paper are generated by the model: having specified the nature of competition, the model solves for the marginal costs that would generate the actual 2004 market shares.
(8) I originally used 5% as the cost of capital, which seemed a reasonable real rate of interest given that the NZ government--a relatively low risk borrower--owns most of the airline. However the airlines' lawyers submitted that the appropriate figure is 15%. This seems implausibly high, and is, for example, out of line with the attitude taken by Air New Zealand's board to the raising of capital needed to finance its fleet expansion. To be 'conservative', in this paper I use 10% for the cost of capital.
(9) Though note Air New Zealand's invention and introduction under highly competitive conditions (in 2002) of its very innovative 'Express Fare' system, which was quickly matched by Qantas, and which now (January 2005) has in essence been copied in the United States by Delta Airways.
(10) Another problematic item in the Purpose clause is the qualifier 'long-term' applied to consumer benefits. An economist might read 'long-term' as meaning 'not necessarily short-term', and use it to excuse immediate price increases that might be compensated for by some future benefit (eg Evans, 2004, p247). The Deputy Chair of the Select Committee that framed the purpose statement section IA warned explicitly against this interpretation: 'The use of the term "long-term" also needs to be treated with caution as it is easy to see more immediate benefits for consumers traded off for some promised benefit that accrues for in the future' (Kevin Campbell, quoted by John Land in Submission to the High Court at para 14.23). The law is not usually drafted by or for economists and perhaps the appropriate interpretation of 'long term' here is its more everyday sense of 'long-lasting' or 'durable'.
Tim Hazledine, Department of Economics, The University of Auckland firstname.lastname@example.org
Table 1: Impact of Air New Zealand/Qantas Cartel NZ Main Trunk Trans-Tasman Auckland-LA Market Market Market CF F CF F CF F market price 1 1.18 1.15 1.33 1 1.125 market output 1000 885 895 724 1000 841 incumbent price 1 1.17 1.13 1.31 1 1.125 total incumbent output 1000 685 685 514 1000 841 Air New Zealand output 650 685 360 266 650 546 Qantas output 350 0 325 248 350 294 fringe legacy output 0 0.00 110 110 0 0 entrant price 1.09 1.09 1.22 entrant output 0 200 100 100 incumbent CV parameter -0.2 1 0 1 0.25 1 cournot demand elasticity -1.10 -1.3 -1.1 NZ passenger share 0.9 0.45 0.3 Australian passenger share 0.05 0.35 0.05 Foreign passenger share 0.05 0.20 0.65 Incumbent total revenue, $million 500.0 401.4 853.6 743.1 800.0 756.6 Table 2. Short-run Public Benefit Tests NZ Case Australian Case Kiwi Aussie Kiwi Aussie Rules Rules Rules Rules change in NZ consumer surplus, $m -176.5 -176.5 change in Australian consumer surplus, $m -65.4 -65.4 change in foreign consumer surplus, $m change in NZ profits, $m -76.7 change in Australian profits, $m 131.70 change in foreign profits, $m total net benefit, $m -253.2 -176.5 66.30 -65.4 Australasian World Welfare Welfare change in NZ consumer surplus, $m -176.5 -176.5 change in Australian consumer surplus, $m -65.4 -65.4 change in foreign consumer surplus, $m -96.4 change in NZ profits, $m -76.7 -76.7 change in Australian profits, $m 131.7 131.7 change in foreign profits, $m 131.7 total net benefit, $m -186.8 -151.5
|Printer friendly Cite/link Email Feedback|
|Publication:||New Zealand Economic Papers|
|Date:||Dec 1, 2004|
|Previous Article:||What's wrong with New Zealand's public benefit test?|
|Next Article:||"Globalization in Historical Perspective".|