DoJ to AT&T: No! Street to Network Vendors: No!

The Justice Department has filed an anti-trust suit to block the AT&T acquisition of T-Mobile, a move that is raising all manner of comment on both sides of the issue.  With all deals like this, the question is whether the consolidation is bad for the consumer to the point that justifies blocking it.  Is this one of those deals?  I’m not sure.

The telecom industry is suffering from low ROI, and has been suffering for years now.  Consolidation can definitely help with that because competitive overbuilding of infrastructure raises the capex and opex for everyone involved.  Would four operators have a lower total network cost than five?  Sure would.  Similarly they’d have lower marketing costs, and three would present lower costs than four, and so forth.  The limiting case here is a regulated monopoly, which we had at one time in the Bell System.  While the issue here is hardly whether we return to the  Bell System, it’s useful to look at the limiting case to frame the problem.

If we collapsed the industry into a single player there would be no consumer choice and no mechanism to control pricing other than regulation.  That would work in theory despite what many have argued; it worked for a century in fact.  The difficulty with regulation comes not from its inability to protect consumers but from its inability to manage innovation.  How do you get a regulated monopoly to invest in NGN?  You’d have to presume government bureaucrats knew when and how to do that, and all you need to do in order to disprove that presumption is look at the recent debt ceiling fight.  National communications can’t be a slave to partisanism.  A monopoly won’t work any more, but we’re not monopoly just because we approve the AT&T/T-Mobile deal.

I don’t think there’s any convincing proof of consumer harm here, and in fact I think that forcing the industry to try to sustain more players than the market could naturally support has the effect of raising base costs for everyone.  It’s more likely that this is a political move.  The average voter knows little enough about basic, but highly important, issues.  They know nothing about telecom, but they are easily swayed by the vision of Their Government On the Rampage, riding out against the forces of anti-competitive evil to assure them lower wireless prices.  That’s an easy image to sell, where reasoned debate on the merits of the deal would be hard to capitalize on.  It’s an election year.  We have the Party of the People and the Party of Business.  Guess who’s running DoJ!  Wrong decision, in my view.

The same issues of ROI pressure are also hitting the equipment vendors, many of whom will be happy if the merger doesn’t go through.  The big Wall Street research houses are split on 2H11 capex, but the general view is that it will be better than the first half but not up to par.  As a result, vendors in the space are likely to remain under pressure.  The Street seems to think that the carrier Ethernet space will be the most pressured, and thus has been preferencing players with limited exposure there.  My own view is that it’s not helpful to look at the prospects of vendors based on OSI layer.  Spending today is focused on revenue and competition and tends to be more “vertical”.  Everyone knows mobile is hotter than wireline (which is why Cisco did its NEC partnership).  The issue is that outside mobile it’s hard to identify easy vertical categories because there’s no convincing picture of what’s on top of the heap.  The players with mobile dominance are likely to do well, but not only in the RAN.  Juniper, who was dropped by UBS to its “least-preferred vendor” list, has no mobile/RAN position and so is particularly vulnerable.  That’s odd to me, or at least unnecessary, because Juniper has very strong service-layer assets and has simply not been able to exploit them fully.  Errors like that are easily corrected, or at least more easily than fundamental product-line omissions.

UBS also downgraded F5, and that seems to establish a broader negative positioning for switching, which I happen to agree with.  But it’s not because switching is less valuable; it is in fact at the heart of the virtualization and cloud revolution.  The difficulty is that the vendors have not been able to make buyers understand that their products have any distinctive value in supporting those revolutions, and thus are facing “feature shock”, something that happens when you deluge a prospect with features that have no business context to validate them.  Again, this is a problem that should be easily corrected, but somehow we’ve become incapable of dealing with our own benefit case in the industry!  IT giants like HP, IBM, and even Dell are moving data center network iron because it’s increasingly seen as something as undifferentiated as raised flooring.