Are we kidding ourselves on competition?

The traditional textbook model of competition in an oligopoly goes likes this. Firms choose prices and other variables (like product quality, advertising and R&D) to maximise their own profits and disregard the impact of their actions on (a) competing firms and (b) consumers; although with the latter since they want them to buy products they aren’t completely immune to their welfare. This model is essentially unquestioned but, in reality, it relies on a view of firm ownership that is markedly different from corporate reality. In particular, large firms are owned by shareholders (who may also be their consumers) but, more importantly, may be the shareholders of their competitors as well.

Let’s start with that latter notion. What it means is that the presumption that firms maximise their profits independent of their consideration for the impact on profits of competing firms is surely suspect. And it is disturbingly suspect. Consider a situation where there are 10 firms in a market and they compete with one another. Now suppose that all shareholders — say because they are following the dicta of diversification — allocate their wealth in equal proportion across those 10 firms. That means that each owner of the firm — even if there are thousands of these — cares equally about each firm’s profits.

So ask yourself: when those shareholders vote on the composition of boards or the management of the firm, or, importantly how the management of the firm is compensated, are they going to vote for managers who will care only about the profits of the firm they manage or about the profits more broadly? The answer is obvious: they will look to managers who manage in the interest of shareholders and so that means they care about all firm profits and not just the one of their own firm.

In a world where shareholders can get what they want, we won’t have competition in this outcome but, more likely, a collusive outcome. What is more, the firms won’t have to go to all the difficulty of violating antitrust laws to obtain this outcome, they will do it unilaterally. There are no laws against that.

To see how that might arise, think about, oh I don’t know, Comcast and Time Warner. These are cable companies/ISPs who are the largest of their kind in the US. I suspect that there is lots of cross ownership of mutual funds of each of these. Why? Because they happen to each focus on different regions in the US and so any mutual fund worth their salt would want to diversify their portfolio to hedge geographic risk. And so ask yourself: how readily would the boards of those companies improve massive infrastructure investments to expand the reach of their networks into the territory of the other? With cross-ownership, they wouldn’t have a powerful incentive. (Actually, we can now just add that to the list).

Now this isn’t just speculation. Jose Azar, an economist now at Charles River Associates, did his Princeton PhD on this topic. His theory paper is here and it builds on others including Gordon (1990), Hansen and Lott (1995) and O’Brien and Salop (2000). Frank Wolak and I came up with a similar set of issues related to cross-ownership and hedging in electricity markets (for vertical ownership) and verified anti-competitive consequences arising from this. But Azar, along with Martin Schmalz and Isabel Tecu have demonstrated that cross-ownership has anti-competitive impacts on the US airline industry. They find that cross ownership increases US airline prices 3–5%. When they use the event whereby BlackRock acquired Barclays Global Investors (a merger changing the shares of common ownership in airlines), they found such ownership could indicate 10% bumps in pricing with US airline ticket prices rising by 0.6% as a result of that merger alone.

In Slate, Eric Posner and Glen Weyl took these results to their logical conclusion: that we should consider banning mutual funds or anyone from holding shares in competing companies. This confounded and shocked the Financial TimesMatthew Klein but he clearly was struggling to understand the basic industrial economics of the situation and also the power of Azar’s empirical findings.

In reality, the Posner-Weyl policy is on the risk averse side of the equation. Shareholders influence companies by voting and so it is the median voter who matters. In this regard, so long as 51% of shares in companies are owned by funds or people who do not have shares in competitors, we don’t need to be concerned. Of course, how we think about a policy that restricts competitive cross ownership remains an issue here but this is, at least, a softer place to start. My guess, however, is that we can come up with a formula for sufficient asymmetries in cross ownership to assure us that any anti-competitive hanky panky isn’t going on.

But there is another issue which I believe is likely to be more related and of importance for the wealth inequality and power story that Posner and Weyl concern themselves with: i.e., that wealthy shareholders are likely to exercise market power and harm consumers overall. That issue is that the relationship between the distribution of shareholders and the distribution of consumption across firms will be important. As Joe Farrell (1985) pointed out in 1985, if firms are owned by shareholders in equal proportion to their consumption levels, then shareholders would vote to self-regulate any firm’s market power (see also Mas-Colell and Silvestre (1991)). This would be a great outcome and would apply even if the shareholders had holdings across competing firms. Of course, the distribution of wealth is actually more uneven than the distribution of consumption, so this great outcome is unlikely to arise. However, thinking about shareholders more fully will lead us to consider these ownership issues and how they relate to market power more extensively.

The point here is that we cannot really ignore this issue as economists or as policy-makers. We have “known” about it for decades. Now’s the time to take it seriously.