Wednesday, September 25, 2013

Myths of competition

Everyone knows prices are lower with competition - right?  Well - actually they aren't. There is absolutely nothing in the theory of competition that says they must be in all circumstances.

But that doesn't stop people from continually asserting it when talking about telecommunications.

The latest horrendous examples occurred on the 7:30 report last night. The short excerpt from a report by Greg Hoy;

IAN MARTIN: Look, I think it's a fantastic opportunity to get more competition going in at least a large part of the market where companies like TPG and Optus will be able to use their own infrastructure to provide services, broadband services, high speed broadband services directly to consumers.

DAVID KENEDY: The Coalition already envisages that it will allow Telstra and Optus, if it wishes, to continue to operate their cable broadband networks which reach about 25 per cent of the Australian market.

GREG HOY: Competition is a good thing, it can drive down prices. But this is sure to trigger a debate as to whether those who haven't already
been connected to the NBN will be saddled with an inferior technology for the foreseeable future.

So let's look at the theory and reality of competition.

The assertion that prices are lower under competition comes down to a simple conclusion of orthodox economics.  The first conclusion is that a profit maximising firm will set price equal to marginal cost. (see note)  The second conclusion is that where the market output is produced by a number of firms (competition) then firms will enter or leave the market till marginal cost equals average cost - so price equals cost and there is no profit.  (This is economists profit which includes the need to get a return of and for capital in the cost).

The monopoly scenario is built around a market where there is no entry or exit - and as a consequence the firm's profit maximising position is where the mark-up (that is (price-cost)/price) is equal to the inverse of the own price elasticity of demand.

This also highlights another myth - hat monopolists charge whatever price they want.  This is not true.  If costs go up or decline then price will too.  How much by depends upon the shape of the demand curve and hence how elasticity changes from one point to another (elasticity varies the whole length of a linear demand curve).

If the cost doubles because there are competing firms and even assuming this creates the competitive outcome (see note) for prices to be lower under competition the monopoly mark-up would need to have been over 100%.  While that is possible, whether it is in fact the case is an empirical not a theoretical question (that is, what is the actual demand). I am not personally aware of any studies that have indicated this has ever been the mark-up in telecommunications access monopolies.

Which is all good so far.  The question is "why one firm?"  That can be because of regulatory barriers.  But economists recognise it can be because of a thing called "natural monopoly" - which is formally defined as one where the cost function is sub-additive.  That is all possible values of industry output can be generated more cheaply by one firm rather than two.

This condition clearly applies to distribution networks.  I have said it before - no one who promotes competitive access infrastructure in telecommunications advocates duplicating water pipes, electricity wires or gas pipes.

We have had three types of competition in fixed line telecommunications in Australia. The first between HFC and copper pairs.  The second between two HFC networks. The third between competing providers over copper.

The competition between HFC and copper should have been intense because Optus's reason for building HFC  was for telephony competition.  But it didn't work very well, and the price point for accessing Telstra's copper was cranked down by the ACCC to the point where HFC wasn't really effective - unless used in a bundle.  The ability of HFC and copper to co-exist is very similar to the co-existence of gas and electricity.  The latter both deliver energy, but the former is better for application that need heat (household heating, cooking, hot water) while the latter is essential for driving everything else.  Pay TV was built to deliver linear TV, copper to deliver voice, and both were tweaked to provide two-way data.  Both have limitations for data - HFC has poor uploads and shared access capacity, while copper has distance limitations and is severely affected by loop quality.

The issue is that FttH is actually superior to both HFC and copper for all the communications applications.  The rationale for competing platforms expires.

The competition between HFC networks was a disaster.  The two networks wound up competing on content, and once one got a sustainable lead the market fundamentally tipped.  The Optus HFC network still exists but is a poor second cousin to Telstra.  It is also important not to get fooled by the accounting treatment of the Optus asset.  Over a billion dollars in value has already been written off - an accountant will measure profit against the written off base - for an economist the written off cost never truly goes away.

Finally we had competition on DSL over Telstra copper.  Competition cannot be credited with any price declines that came about through the reduction in the ULL price by the ACCC.  Where competition has been effective has been in the upstream component - by taking responsibility for the engineering of the service beyond access firms make decisions about Quality of Service.  Others gain benefits from scale efficiencies on backhaul (from access to network core) and transit (from core to other places on the internet).  Really clever firms understand price elasticity and drop prices in anticipation that lower prices will add scale and reduce average cost to justify the price decline.  All of that happens actually in a world just like NBN Co - there is a monopoly access provider.

So really the idea that the access network should be a monopoly is a no-brainer.  Costs and prices will be lower.

The question then is why a Government owned one.  And there are two simple reasons for that.  The first is that as it is a monopoly it needs to be regulated - and one of the best ways of doing that is Government ownership, rather than tension between interests of shareholders and interests of Government.  And the second is the ability of Government to be a patient investor, which enables it to make the big call to build the whole network relatively quickly.

Actually the Australian Government has a good record here.  The copper telecommunications network was built by Government.  When Telecom Australia came to be in 1975 only 62% of households had a phone.  The Post Office had been entirely funded through revenues and commonwealth borrowings since at least 1959.  When Telecom was created it assumed all $4.5 billion of that debt.  Between then and 1989 Telecom paid interest on that debt and repaid $1.5 billion of the principal.  When it was corporatized in 1989 the remaining $3 billion was converted to equity.  The Government received dividends from Telstra in return.  Immediately prior to the first tranche of Telstra being sold it paid a $3 billion "special dividend" to the Government. I can't recall how much the Government got over the three floats - but it was over $50 billion - for an asset with a carrying value of zero!

And an interesting story dredged up from the archives:

Deputy Prime Minister John Anderson says the federal government should consider spending part of the AU$34 billion it will reap from Telstra's sale to build a national fibre-optic network.

Treasurer Peter Costello has said he wants to see all of the money from the Telstra sale tipped into the government's new Future Fund set up to cover public servants' superannuation payouts.

Anderson said a plan by the Nationals think-tank The Page Research Centre, for the government to build an AU$7 billion fibre optic network across Australia, deserved serious investigation.

He also called on the government to have a close look at the true state of Telstra's existing copper wire phone network, its maintenance cost and how much it would cost to replace it if necessary.

Anderson said the issue of how a shortfall of telecommunications services, particularly in the bush, might be met in the future must be addressed.

"It is my view that if it emerges that there are circumstances where the normal commercial forces are not going to drive critical investment in infrastructure and if it becomes apparent that government intervention is needed...the logical first option to fund it would be from the sale proceeds," Anderson told a telecommunications conference in Canberra.

"It's hard not to observe...(that if the technology) was optical fibre rather than copper a lot of the concerns of regional Australians about Internet speed simply would not have arisen."

Later, Anderson told reporters he thought it would be unlikely the government would have to dip into the Telstra sale proceeds to build telecommunications infrastructure.

But he said such a move would be logical if needed.

Note: The orthodox view has been challenged.  Steve Keen has demonstrated that the orthodox approach underestimates the actual price in real markets.  Also if the market is not fully competitive the mark-up doesn't drop to zero even in theory - the mark-up becomes the HHI divided by elasticity where the HHI (or Herfindahl–Hirschman Index) is the sum of the squares of market shares of firms. The lowest that can be under duopoly is merely half - that is the mark-up is reduced by 50%. 

1 comment:

Listohan said...

The same concepts apply to roads (and increasingly rail). In both cases, the government provides the infrastructure and the freight companies compete to carry freight on that infrastructure.

Since it favours (and presumably understands) the concept of roads, perhaps those examples will carry more weight to convince the government it does not really understand how competition can best deliver the advantages the government claims at every opportunity.