By Alex Sundakov*

PRICES AND COSTS: GOOD ACCOUNTING IS BAD ECONOMICS

Introduction

A little bit of economics knowledge is a dangerous thing. A basic idea in economics is that in competitive markets prices tend to reflect marginal costs. In economics this concept provides an important means of calibrating the effects of real-life complexities and transaction costs. Theoretically perfect competition is an intellectual benchmark, not a realistic policy target.

However, in the political and policy debate, the idea of prices equating costs in well-functioning markets has jumped from its useful role as a high level concept into a mechanically applied policy target. Unfortunately, this will inevitably lead to bad policy.

What is Cost?

There is an obvious practical difficulty in converting the economics concept of cost into observable accounting data. Marginal cost is typically not observable, and would have to be proxied from the accounts. In any case, as long as fixed assets are employed – as is the case in just about every imaginable business – setting prices at marginal cost would lead to loss-making.

Economists use the concept of long-run marginal cost to try to capture the effects of fixed capital: when prices equal long-run marginal cost they cover the return on and the return of (depreciation) of the capital increment required to meet demand. But long-run marginal cost is also an artificial concept: in practice, capital is often lumpy, and it is extremely difficult to figure out exactly what capital increment is required to meet a particular level of demand. Moreover, even in rare circumstances when it is possible, observing whether prices are set at long-run marginal cost provides only partial information about the degree of competition in a market.

In practice, whenever policy-makers try to regulate price to cost, or try to derive information from the relationship between prices and costs, they focus on what is observable: the average cost of production. Even such a measure often requires many simplifying assumptions. Since many firms produce more than one product, fairly arbitrary rules have to be applied to the allocation of common costs, producing lots of happy hours for regulatory accountants.

But the real problem is that no matter how good the approximations are, the information about average costs does not provide us with any economically meaningful insights into what prices should be.

The reason for this is that, in the real world, markets are complex, dynamic and overwhelmingly characterised by the transactions costs. Pricing reflects reward for risk taking and is often designed to create specific incentives for customers so as to manage transactions costs.

Consider two examples to illustrate this point. First, think about the astronomically high charges for using telephones in hotel rooms. Most travellers would have experienced a moment of frustration, and wondered why hotels are stupid enough to force them to use their mobile phones, thus missing out on the revenue they would have received if they had set the price at a reasonable level. But there is a very good reason for what the hotels do. Long-distance telephone charges create credit risk for hotel operators. Guests can potentially incur high costs. If guests were encouraged to use hotel phones, hotels would have to devote more risk to credit risk analysis. They would be less willing to accept pre-paid reservations (where the traveller may not have further funds available), would be even less willing to deal with travellers without credit cards and would require security deposits. All of this would incur costs and create a potential for conflict with customers – something that hotels in a competitive market would prefer to avoid. The alternative would be not to provide telephones, but this would reduce the level of service for those who need them.

By setting prices for phone calls very high, hotels ensure that most people would not use them, and hence reduce their credit risk. They also provide those willing to pay, or requiring emergency access to a telephone, with the service they need. An analysis of the relationship between costs and prices would suggest that hotels are not competitive in this area (strangely, even though they may be competitive in other areas). A careful analysis of incentives shows that prices out of line with costs are a highly competitive response to the real world transactions cost: the fact that people do not always honour their credit.

The second example is the debate about the mining super tax. Clearly, many mining projects generate very high returns during a mining boom. But how does one know whether the returns are really above economic cost? This year’s profits reveal nothing, since the costs of the mining projects would have been incurred in the past, and there is no reason to expect that the accounting treatment of asset depreciation would fully reflect those costs. So, how far does one go back, and how does one calculate the actual costs?

More importantly, does it really make sense to attribute costs to a single mining project? A lot of mining exploration is a form of wildcatting: many holes are dug and occasionally something is found. The profits from the successful hole have to cover the costs of all the unsuccessful ones. So, how does one account for past failures? What if there were many failures and a few successes? Should the costs of failures be allocated evenly across the successes?

One can easily get tied in knots trying to force a dynamic and complex market into a static cost model. The problem is that any answer to the above questions would have to be arbitrary.

There may be a few fairly static utility applications where a link between average costs and prices may be useful. It has become conventional to regulate electricity and water distribution companies on the basis of recovery of average costs. However, this is a pragmatic compromise, based on the expectation that these are fairly static businesses with slowly changing technology and consumer demand, rather than a genuine improvement in efficiency. Things get very dangerous when trying to extend this compromise into more fluid and complex markets.

Exhibit One: Regulation of Mortgage Termination Fees

In late 2010, the Australian Securities and Investment Commission (ASIC) issued a regulatory guide on early termination fees for residential loans. The publication of the guide happened to come out smack in the middle of the political furore over the rising interest rates. This coincidence gave the guide an appearance of a timely response to an emerging problem. In fact, the guide had been in the making for some time, and represents a troubling development in regulatory thinking. ASIC’s approach to deciding which mortgage fees are unconscionable and unfair is based on bad economics, and establishes an unfortunate precedent which will reduce competition and disadvantage consumers.

In essence, ASIC takes the view that exit fees, if any, must reflect the administrative costs of processing an early termination. Anything else runs the risk of being viewed as unfair and unconscionable. In particular, ASIC emphasises that fees which seek to recover the forgone profits from a fixed-term mortgage contract are likely to fall foul of the regulator.

This approach, which narrowly ties one of a range of banking fees to the cost of the directly corresponding service component, stands market logic on its head. In effect, ASIC says that pricing by mortgage providers cannot be used to send incentive signals to consumers or to manage risks broadly across a portfolio of products and services. Rather, to be safe from the regulator, prices and services must be unbundled.

A key characteristic of dynamic and competitive markets is that they continually create innovative product and service bundles, where pricing bears little resemblance to the cost of each component. Hotel telephone and minibar charges are high not because the market is not competitive, but because hotels use them to send signals to consumers. In the highly competitive mobile telephony market, call charges are used to recover the costs of the hand-set. Caps and data charges are set to attract different categories of users to different service packages. Most importantly, in many competitive markets, consumers have a choice between a range of lock-in options – from pre-paid to multi-year contracts – where they can trade-off flexibility for usage costs.

If ASIC’s logic was applied to a wide range of dynamic and competitive markets, one would see a reduction in competition. Similarly, forced unbundling does nothing to promote banking competition. Banks compete over a portfolio of offerings. Customers choose between banks on the basis of the mortgage rate they charge, the transactions services they provide and various other products they offer, such as credit cards. In principle, there is no reason why customers could not buy different components of that bundle from different banks. But equally, there is no reason why they would not buy a comprehensive bundle from a single provider.

In a competitive market, one would expect banks to provide consumers with pricing options both for individual products and over the entire bundle of services. Pricing for bundles would be targeted to customers’ needs. Some customers may prefer to pay as they go, but to keep their options open. They may take up a bundle of residential loans and transaction accounts which had no exit fees, but charged all costs up-front through per-transactions fees. Others may prefer to sacrifice some flexibility in return for low payments up-front. These customers may prefer a bundle with no transaction fees, but significant exit fees.

The key point is that there is no reason at all why the pricing structure within a bundled offering should have any relationship to the costs of various service components. Customers think it is very competitive when a bank trying to sell them a large mortgage offers a low fee account to go with it. So, why is it not competitive to charge an exit fee which compensates for the low transaction fees by reducing the risk of them switching to another provider?

ASIC’s decision that loan exit fees cannot compensate banks for foregone profits further distorts the market. A fixed-rate mortgage is a bet between the bank and the customer about the shape of the interest rate cycle. To be viable, any bet has to have two sides to it. ASIC turns fixed-rate mortgages into a one-way bet: a customer can exit with no consequences if the rates decline, while the bank has to stick with the bet if the rates rise. This will undermine the incentive to offer fixed-rate products, making consumers worse off.

Competition in Dynamic Markets

The fundamental problem with ASIC’s regulatory guidance is that its analytical approach does not fit with the logic of competition. Cost-of-service regulation – the type of cost analysis ASIC is trying to apply – is used to regulate natural monopolies such as electricity networks or rail infrastructure. ASIC is trying to use the methodology for regulating prices in a competitive setting. The result is likely to be perverse: less competition and less fairness to consumers.

What is particularly unfortunate is that ASIC’s regulatory guide gives apparent intellectual cover to the broader trend towards forced unbundling and cost-plus regulation of bank charges. The class action against account penalty fees is another example of this anti-competitive trend. There too, the argument is essentially that banks should not be allowed to use overdue and other penalty fees to try to change consumer behaviour and to provide incentives within the overall bundle of services and prices they charge. The logic of the class action also firmly rests on the idea that each bank fee must directly reflect the cost of providing that particular component of service. It too will reduce the quality of choice and service available to customers.

A careful analysis of business costs is good accounting. It is typically very bad economics, which leads to bad decisions.

* Alex Sundakov, Executive Director, Castalia Strategic Advisors.

The citation for this article is (2011) 19 AJCCL 147.Â