Practical men, who believe themselves to be quite exempt from any intellectual influences, are usually slaves of some defunct economist. (Keynes) Many have recently said 1) US industries have become more concentrated lately, 2) this is a bad thing, and 3) inadequate antitrust enforcement is in part to blame. (See
That's why I qualified the claim with "if merger costs are small compared to barriers to entry". Another way to look at it is, companies use merger acquisitions to subsidize their competitors' exit.
If firm exit were directly subsidized by government, in an industry with barriers to entry, do we agree that *that* would cause excess concentration? And if so, what makes it different when the subsidy comes from a competitor?
This reply confused me greatly, until you wrote your next post with some examples. I think the difference is that you're focusing on vertical-integration acquisitions (which don't lead to excess concentration) while I'm thinking more of horizontal acquisitions (which do).
Missing from these models: mergers and acquisitions. If firms are free to merge with each other and merger costs are small compared to barriers to entry, then there will be too few firms.
"Optimal" there is defined specifically as maximizing consumer/producer surplus in that industry -- but many people would find it suboptimal for a different reason: because it concentrates too much power in one organization. (Power that could be used to influence politics or to "bully" other industries.)
Yes, the optimal number of firms with network effects can be low. Yes they may use that power to set prices that are too high. But as I said, that is mis-pricing, not excess concentration.
Is this because the model says that the optimal amount of firms in an industry with high network effects is few (or even one firm)?
In line with the models: social media users do, in a revealed preference sense, WANT there to be few (or one) big firms, because they are demanding to use the firm with the most users.
However, while efficient in that sense, such a situation still seems to leave that dominant firm with a lot of monopoly power. That concentration of power (which can be used politically, or to push into other industries) may be what people most object to.
Network effects are a real thing, but models with network effects don't predict excess concentration. The might predict mispricing, but that's a different story.
The biggest thing I see is network effects. Especially relevant for modern US / big tech companies.
If almost everyone prefers to use the biggest social media company because the largest network is on there, then intuitively the largest such company will have a degree of monopoly power.
Regulatory costs are fixed costs, which favor bigger firms. But models that include fixed costs don't predict too few firms. Of course all else equal it would be good for all kinds of costs to be lower, including regulatory costs.
> Including these factors in models does not typically predict excess concentration.
Anecdata: when I was working for a large bank, a top executive told me that they would complain about regulation but in fact they love regulatory costs, especially those that imposed his costs on going from 0 to 1 customer, because it limited competition.
I am interested in this in the context of occupational licensing. I see examples everywhere of this causing higher prices (and worse service).
Is there a citation for the claim that including these factors in models does not typically predict excess concentration?
In you inquiries have you seen or looked into such aspect as firm size and when it stops being a market participant and becomes, at least in part, a market itself?
I'm certainly not surprised that some will see concentration X and jump to "that's bad" when in fact we really are all better off. Still, I wonder if the analysis you're referencing keeps a clean separation between market and market participant. Or do you largely reject the idea that at some size a corporation starts displacing the market.
Thanks (i meant more:-). And I guess it's situational and a matter of judgement how far the analysis must extend beyond industry X to decently assess total welfare impact.
That's why I qualified the claim with "if merger costs are small compared to barriers to entry". Another way to look at it is, companies use merger acquisitions to subsidize their competitors' exit.
If firm exit were directly subsidized by government, in an industry with barriers to entry, do we agree that *that* would cause excess concentration? And if so, what makes it different when the subsidy comes from a competitor?
Horizontal mergers only temporarily increase concentration, they don't create a new equilibrium with a different concentration.
This reply confused me greatly, until you wrote your next post with some examples. I think the difference is that you're focusing on vertical-integration acquisitions (which don't lead to excess concentration) while I'm thinking more of horizontal acquisitions (which do).
I think I remember reading something by Julian Simon about research that showed a few firms produced enough competition to keep prices down.
We have plenty of models of mergers; they don't predict excess concentration.
Missing from these models: mergers and acquisitions. If firms are free to merge with each other and merger costs are small compared to barriers to entry, then there will be too few firms.
"Optimal" there is defined specifically as maximizing consumer/producer surplus in that industry -- but many people would find it suboptimal for a different reason: because it concentrates too much power in one organization. (Power that could be used to influence politics or to "bully" other industries.)
Yes, the optimal number of firms with network effects can be low. Yes they may use that power to set prices that are too high. But as I said, that is mis-pricing, not excess concentration.
Is this because the model says that the optimal amount of firms in an industry with high network effects is few (or even one firm)?
In line with the models: social media users do, in a revealed preference sense, WANT there to be few (or one) big firms, because they are demanding to use the firm with the most users.
However, while efficient in that sense, such a situation still seems to leave that dominant firm with a lot of monopoly power. That concentration of power (which can be used politically, or to push into other industries) may be what people most object to.
Network effects are a real thing, but models with network effects don't predict excess concentration. The might predict mispricing, but that's a different story.
The biggest thing I see is network effects. Especially relevant for modern US / big tech companies.
If almost everyone prefers to use the biggest social media company because the largest network is on there, then intuitively the largest such company will have a degree of monopoly power.
Regulatory costs are fixed costs, which favor bigger firms. But models that include fixed costs don't predict too few firms. Of course all else equal it would be good for all kinds of costs to be lower, including regulatory costs.
> Including these factors in models does not typically predict excess concentration.
Anecdata: when I was working for a large bank, a top executive told me that they would complain about regulation but in fact they love regulatory costs, especially those that imposed his costs on going from 0 to 1 customer, because it limited competition.
I am interested in this in the context of occupational licensing. I see examples everywhere of this causing higher prices (and worse service).
Is there a citation for the claim that including these factors in models does not typically predict excess concentration?
I don't understand you. ALL firms are market participants, both as buyers and as sellers.
In you inquiries have you seen or looked into such aspect as firm size and when it stops being a market participant and becomes, at least in part, a market itself?
I'm certainly not surprised that some will see concentration X and jump to "that's bad" when in fact we really are all better off. Still, I wonder if the analysis you're referencing keeps a clean separation between market and market participant. Or do you largely reject the idea that at some size a corporation starts displacing the market.
Thanks (i meant more:-). And I guess it's situational and a matter of judgement how far the analysis must extend beyond industry X to decently assess total welfare impact.