Monday, March 9, 2015

You Can't Forecast Telco Revenues and Costs 20 Years from Now

Projecting telco revenues and network costs 20 years from today is not useless, but likely to be highly inaccurate. 

As early as the 1990s, a rational executive could have claimed that the majority of firm revenue in a decade or two would be generated by products that had not been invented yet. That has largely proven to be the case.

Also, anyone in the forecasting business can look back on projections made 20 years ago and confirm the wisdom of not being too certain about such forecasts.

In the early 1990s, before the passage of the Telecommunications Act of 1996, it likely would have seemed inconceivable that the biggest U.S. telecommunications companies would, with a couple of decades, be minority suppliers of the key fixed network telecommunications products.

But that has happened. In the strategic high speed access business, the largest U.S. telcos have just about 41 percent market share.

Perhaps significantly, the incumbent telcos seem to be slipping further behind. In 2014, cable TV companies added 89 percent of the net high speed access additions, building on the 82 percent net gains cable TV companies made in 2013.

In the voice business that once drove 70 percent of industry revenue, Verizon Communications recently sold $10.5 billion in fixed network assets, to support its capital investment in mobile spectrum, the one area where the tier one telcos still dominate.

From 2000 to year-end 2013, telcos will have lost nearly 62 percent of all traditional phone lines  and 70 percent of traditional residential voice lines, USTelecom says.

For the twelve-month period from mid-2011 through mid-2012, residential and business consumers dropped 10.1 million ILEC switched voice lines, a twelve-month decline of 10.7 percent, according to Federal Communications Commission data..

From 2000 to mid-2012, the number of ILEC switched lines fell from 186 million to 84 million, or a decline of 55 percent. Straight-line trends suggest ILECs will have lost approximately 62 percent of these lines by the end of this year, according to the USTelecom.

Telco switched line losses have been greatest in the residential market, where the annual rate of decline from mid-2011 to mid-2012 was 13.6 percent, USTelecom says.

In 2000, some 120 million consumer voice lines were in service. As of mid-2012, there were approximately 45 million consumer telco lines being purchased, a decline of 63 percent.

Newer services such as linear video entertainment represent a smallish percentage of telco revenues, as both AT&T and Verizon have about six percent market share each, of the traditional subscription TV business.

AT&T’s planned acquisition of DirecTV notwithstanding, entertainment video has been--and always has been--a tough business proposition for any telco, of any size.

One study conducted for an independent U.S. telco in the early 1990s found the video subscription business would reach positive net present value only after eight full years of operation, even with a “first installed cost” of just $800, at $365 annual revenue for a subscriber.

The analysis by our consulting team assumed at 20 percent net income and a 12-percent discount rate. In other words, the venture would make money, but only barely, with breakeven in year nine of the 10-year investment lifecycle.

Even that analysis assumed use of fixed wireless access, as the first installed cost of asymmetrical subscriber line in the early 1990s ranged as high as $8,000 per line, at a time when fiber to the home cost $4,500 per line and fiber to the curb cost about $2500 per line.

Much has changed since the early 1990s. Access costs have fallen. Video revenues have skyrocketed. Significantly, the early 1990s analysis included zero revenue contribution from what we now call “Internet access.”

The biggest conclusion might be the near futility of forecasting network costs and revenue streams as they will develop over a couple of decades.

No comments:

Many Winners and Losers from Generative AI

Perhaps there is no contradiction between low historical total factor annual productivity gains and high expected generative artificial inte...