Diseconomies of scale

Why isn't there just one company in the economy? That's the question you should ask whenever anyone tries to tell you that a proposed merger - such as that between NTL and Virgin Mobile - is a good idea. If mergers are a way of increasing efficiency, why don't all firms merge into one?
The answer isn't just that this would destroy competition. It's that there are diseconomies of scale, which cause larger firms to become inefficient.
And these really are important. In this paper (pdf), Staffan Canback shows that the U-shaped average cost curves we read about in the textbooks really do exist. Beyond a certain size, efficiency falls.
This helps explain a well-established finding - that the big gains which investors expect from mergers rarely happen.
In this recent review of the field (pdf) Robert Bruner concluded that "abnormal returns to buyer shareholders from M&A activity are essentially zero." M&A does pay, he says, for shareholders in the target firm - but it doesn't consistently raise profits.
His findings augur especially badly for  the NTL-Virgin deal. He estimates that M&A is unlikely to work for glamour stocks (those on high price-book ratios) rather than value stocks, and likely to fail when motivated by a desire to increase market power rather than cut costs; the rationale for the NTL-Virgin deal seems to be to create a "quadruple play", offering broadband, TV, mobile and landline phones.
I mention all this because I suspect the taste for mergers is an example of the salience heuristic - the gains to mergers seem obvious, whilst the costs of them (difficulties in managing larger firms) are more hidden.

Comments

[Paddy Carter]

One problem I have with all the research I have seen that demonstrates M&A tends to produce poor returns is the lack of counterfactual. As ever this is, in some respects, a pointless criticism, because we never do have counterfactuals in such cases, but nonetheless ... some M&A moves are defensive. The idea being that without them, the company would wither, sooner or later. So even though the calculated return is rubbish, the alternative (not-merging) may have been worse. No, hang on, that is a rubbish criticism, because we it works in both directions: we don't know that life might not have actually been better without the merger. That's the point. Duh.

It is, though, an argument that is sometimes put. For example, Microproceser company Arm recently bought a company called Artisan, in what looks like a text book example of a deal that will not create returns (over priced, no synergies). However, Arm executives argue that in the future, they will need to have the technology Artisan has given them, and there was no cheaper means of acquiring it. Or something like that. Or perhaps sometimes companies want to stop a rival getting their hands on an asset (Oracle recently bought a company called Retek to stop SAP getting it, and paid well over the odds). In a game theory way of looking at things, they'd argue that every alternative move was worse, even though the financial maths make the deal look like a bum move. Is this feasible, or hogwash? I'm just thinking aloud here.

[Robert Schwartz]

Mergers will remain popular as long as executives who promote them continue to get raises and bankers continue to collect fees.

[AJE]

"I mention all this because I suspect the taste for mergers is an example of the salience heuristic - the gains to mergers seem obvious, whilst the costs of them (difficulties in managing larger firms) are more hidden."

Great point. Would you also say that this is some kind of "planners bias" where people overestimate the extent to which they can plan economic activity? Solution being to read more Hayek!

[dearieme]

Put aside the general points: those of us who are NTL customers know that its management suffers from alcoholic conviviality in a brewing facility organisational dysfunction syndrome.

[Innocent Abroad]

Dearieme, I'm an NTL customer too and while I might have agreed with you six months ago they've got a lot better and frankly I expect things to get worse under the mad balloonist's régime... although if he can get Test Cricket on a proper PPV basis all will be forgiven...

[chris]

AJE - I certainly would say that. We cannot over-emphasize the parallels between centrally planned economies and centrally planned companies.

[Kevin Carson]

I totally agree with what you say about diseconomies of scale (with the research I'm doing on organizational behavior, btw, I've got stacks of Canback articles out the wazoo).

I would add this caveat, though. Bureaucratic diseconomies don't necessarily lead to poor market performance. If an industry is largely cartelized by state regulations, and made up of a handful of oligopoly firms that share the same pathological organizational culture, then they're largely insulated from the ill effects of inefficiency.

[Kevin Carson]

Very interesting post. I touched on some of the same issues on my blog today. It's nice to see Canback cited, because he's done some excellent work. His biggest influence was Oliver Williamson; he argued that the main reason for preferring internal hierarchy over market contract, despite bureaucratic diseconomies, was the need for a governance structure for dealing with asset specificity. One thing he neglected, though, is the fact that asset specificity is a dependent variable. Whether specialized or general-purpose capital equipment is more efficient, for example, depends on the volume of demand. Government policies that promote artificially large market areas also promote higher levels of asset specificity. In a decentralized economy of small-scale production for local markets, asset specificity would be a lot less severe.