Last week, a new working paper purports to provide evidence supporting the idea that overall market power exhibited by firms in the US has increased over the past 30-40 years. This was then picked up by a few blogs/media outlets. In short, the paper claims that there has been an increase in overall profit margins (or mark-ups) since 1980 and, therefore, this implies that overall competition between firms has decreased and that antitrust enforcement is not working.
The initial basis for the paper’s argument is that under 'perfect competition;, firms are supposed to charge a price equal to their cost. In standard economic theory, 'perfect competition' is a rather idealised scenario in which there are many buyers and sellers, each of which is a price-taker (i.e. no individual buyer or seller has any influence over the price of the product). Under these, and a few other conditions, in a 'perfectly competitive' market, the price of producing a unit ends up being equal to the cost of producing that unit – i.e. each producer in such a market makes zero economic profit (the 'economic' nature of the profit, as compared to accounting profit, is a crucial distinction). The paper then compares this to the standard economic theory of monopolistic pricing, in which one firm is the sole producer of a product and therefore can charge a price above cost – i.e. the monopolist obtains a positive profit margin on its sales. The paper then claims that the fact that its data indicate an increase in margins over time mean that the US economy has (in the aggregate) moved away from the 'perfectly competitive' scenario and towards the monopolistic scenario, thereby implying a reduction in the overall levels of competition in the US economy.
Unfortunately the paper fails to take into account a number of factors. For example, and at a rather basic level that the paper’s authors should really be getting right, the “costs” in the theoretical perfectly competitive market do not coincide with measures of costs that are calculated in companies’ accounts. In particular, economic costs include an 'opportunity cost' of using the resources for the next best option – i.e. economic costs include an additional element beyond the balance-sheet cost of using/purchasing the input that is not usually picked up in accounting measures of cost. Under the perfectly competitive model, although there is no difference between the price of a product and the economic cost of producing it, there is likely to be a difference between said price and an accounting measure of cost – in other words, even under the perfectly competitive model, individual firms are likely to make some positive accounting profits.
Despite this, the paper goes ahead and calculates margins (and makes inferences thereof) using accounting cost - the paper uses the accounts of publicly-traded firms in the US over the period 1950-2014. In other words, the paper automatically fails to measure the true margin as is relevant for economic theory. Hence, any attempts to link an increase in the margins in this paper to the competitive landscape as suggested by economic theory is flawed.
Even if the observation that margins have increased was valid (i.e. the margin was calculated appropriately including the economic cost of production), then that is still insufficient to support the paper’s claims that overall competition has decreased. In essence, by making such a claim based on the path of margins, the paper is claiming that the entirety (or, at least, the vast majority) of the increase in margins was due to a decrease in competition, thereby ignoring any other factors that could have resulted in an increase of margins over time. Although the paper looks at one other factor that could explain the change in margins (changes in the average size of firms), the paper ignores factors such as changes in the types and nature of industries over time (e.g. some industries might have much higher up-front costs and lower marginal costs, and if those types of industries grew over time, then that could explain the increase in average margin without any change in competition levels).
On a more technical note, margins can also be related to price elasticities of demand via the “Lerner Condition” – this is a mathematical relationship stating that a firm’s margin is inversely proportional to its own price elasticity of demand. Obviously, different industries can have different demand elasticities regardless of the level of competition in each industry and, as such, margins can differ due to that reason as well. This is particularly relevant if, as seems likely, the composition of the economy (in terms of which industries are most prevalent) has changed over time.
Worse still is that the paper’s result of margins increasing over time is likely to be affected by a “survivorship bias”. Specifically, as the paper tracks firms over time, clearly the more successful firms (the ones that survive and earn higher profits) will remain in business, while the less successful firms (the ones that go bust due to making lower profits) will exit the market. Consider a stylised example: suppose at the start, an industry consists of six firms of equal size in terms of revenues, but five firms each make a margin of 30%, and the sixth firm makes a margin of 1%. At the start, the average margin would be about 24%. Now suppose that the owner of the firm obtaining a margin of just 1% decides that they can do better in another industry, so decides to shut down – this means that the average margin would increase to 30% despite there not being any real decrease in the level of competition in the industry (as the five remaining firms would still compete against each other).
Hence, over time, one would expect the sample of firms over which the margin has been calculated to contain mainly successful firms and to lose the less successful firms, thereby resulting in an increase in average margin over time. The paper does not seem to have tried to account for this. (Note, too, that firms going bust is a sign of healthy competition – a more efficient firm is able to outcompete a less efficient firm such that the less efficient one stops trading.)
Overall, therefore, although the paper claims that 1) there has been an increase in margins over time; and 2) this implies that industries in the US have become more monopolistic over time, those claims do not stand up to scrutiny. Indeed, the paper’s approach to demonstrating such claims is flawed at the most basic level.