Bandwidth price collapse: Will it ever end?
Bandwidth as a Commodity
Several factors combined to drive down prices. With construction at a halt, consolidation imminent, and swaps under scrutiny, the return of a cash-based market may offer a silver lining.
When the network construction boom was at full tilt, most industry observers and participants expected bandwidth prices would fall rapidly as new capacity became available. Many bandwidth providers even built assumptions of annual double-digit price declines into their business plans.
However, few observers expected that price competition would rapidly escalate into a headlong race to the bottom. Many bandwidth providers are finding that wholesale capacity is now trading at prices that are 99% lower than when they built their networks two or three years ago.
Supply and demand
The network construction boom spawned the infamous bandwidth glut, which played a central role in depressing prices. Some cities are connected by literally thousands of (mostly dark) fiber pairs. However, the interaction of supply and demand no longer plays a role in ongoing price changes in the long-haul capacity market. The amount of bandwidth on terrestrial networks has so completely outstripped the market's ability to absorb it that demand is likely to play little role in determining prices, with the exception of a small number of thinly supplied routes.
The impact of the supply overhang is compounded by the fact that the available capacity is distributed widely-among multiple carriers. Only four years ago, bandwidth buyers in Europe had little choice but to purchase capacity from the monopoly incumbents. Today, several European (and American) cities are served by as many as 20 distinct wholesale bandwidth providers. Even small cities like Geneva house more than 10 bandwidth suppliers.
Recent technology advances sharply reduced the unit cost (dollars per bit per mile or kilometer) of bandwidth. Indeed, the newer the network, the lower the unit costs tend to be. This reduced cost base prompted new market entrants to aggressively underprice rivals that were earlier to market. Many late market entrants touted this "late mover advantage" as key to their business strategy. However, many early entrants proved willing to match their newer rivals price cut for price cut.
The cost of constructing a network is directly proportional to the length of the route. Longer routes require more amplifiers, more regenerators, more fiber, more trenching, and frequently, more complex rights-of-way negotiations. Consequently, all things being equal, longer routes need to cost more than short routes.
Per-circuit prices for STM-1/OC-3 (155.52-Mbit/sec) annual leases in the United States and Western Europe are generally comparable. STM-1 lease prices in Europe range from about $50,000 to $175,000 per year; OC-3 prices in the U.S. range from $20,000 for shorter routes to $190,000 per year for New York to Los Angeles, the longest major route in the United States.
Costs may be overridden by other considerations, however, most notably by the financial health of the bandwidth provider. A company in dire financial condition may be forced to sell capacity below cost to meet short-term debt payments. There is ample evidence of this trend on undersea routes. In October 2001, a carrier offered to resell transpacific STM-1 indefeasible rights of use (IRUs) on PC-1 for about $1 million-$1.5 million, well below the cable's estimated costs.
Capacity swaps between operators helped to increase long-term competitive pressures on network operators by enabling them to enter one another's markets in short order. While a swap allowed an ambitious network operator to enter a rival's market quickly and cost-efficiently, the transaction afforded the same courtesy to its rival. Many operators' network maps appear strikingly similar. This similarity is no illusion: Many competitors' network links are built on strands of fiber only centimeters apart and buried in the same conduits.
More insidiously, in the past 18 months, some carriers have used superfluous asset swaps to inflate their top-line revenue growth. While these transactions may have staved off the inevitable day of reckoning by a few months, they also saddled carriers with completely useless network assets. For example, the creditors of one bankrupt pan-European network operator recently discovered that they now also own metropolitan-network assets in Canada. Similarly, on the last day of the third quarter of 2001, Enron and Qwest concluded an asset swap valued at more than $500 million, whose centerpiece was dark fiber between Salt Lake City and New Orleans-a route on which both carriers already had ample capacity.
As financial pressures increased, many of these swapped assets ended up back on the market, frequently at prices far below their book value or the prevailing market price. These sales of "off-net" capacity often made it even more difficult for other providers seeking to bid for long-term business on a route to obtain a fair price for their network assets.
Short- and long-term outlook
How are these factors likely to play out in the months (and years) ahead? There appears to be little scope for further price reductions. In light of the relatively low fill rates on most networks, it is likely that prices are at or below cost for even the most competitive operators. In recent months, many industry observers have suggested that this reduced scope for further price reductions has led to a slowing of price erosion. There is some evidence, largely anecdotal, that the rate of decline on some highly competitive routes has in fact slowed. In the past six months, prices on the hyper-competitive New York to Los Angeles route have been relatively steady-at least by the standards of the bandwidth market.
However, research published in TeleGeography's "International Bandwidth 2001" report (April 2001) indicated that European "prices plunged rapidly but have since leveled as the European market has matured." It turns out that prices only held steady for a few more months before plunging further by 60-70%.
Thus, the current respite may simply be the calm before the next storm. Several carriers recently introduced substantially more aggressive pricing. Whether others will follow suit remains to be seen. Nevertheless, several factors indicate pricing stability may come to the market:
- Rate of decline in previous years. Prices have fallen by more than 90% in only two to three years on most routes. At these price levels, which are at or even below the sellers' costs, little scope remains for further reductions.
- End of network expansion. Plans for new carriers or market entrants have all but disappeared.
- Market consolidation. As carriers curtail their network expansion plans, others may seek to acquire and consolidate the assets of bankrupt carriers, reducing the number of active competitors on the market.
- Demise of the swap market. The swap market has essentially evaporated. The return to a cash-based market should allow far more realistic pricing practices to emerge.
- Flight to quality. In light of the current market turmoil, buyers are favoring stable, financially sound carriers-or at least that's the perception. If that is indeed the case, it hasn't translated into additional leverage in terms of pricing. Frequently, financially stable incumbent carriers are among the lowest-priced bidders on a route. Conversely, some carriers expected to file for bankruptcy are on the most expensive end of the market.
One wild card remains in the deck: bankruptcy. Several major terrestrial and undersea network operators, including Global Crossing, 360networks, KPNQwest, and Williams, have filed for bankruptcy. Dozens more are teetering on the brink of insolvency and expected to file for protection from their creditors within the coming 12 months.
In the short term, the bankruptcy of a large carrier is likely to provide a shot in the arm for its more financially sound rivals, as customers abandon the bankrupt carrier for a more stable service provider. One of Global Crossing's marquee accounts, SWIFT, announced in April that the company would seek to purchase services from other carriers.
However, bankruptcy will have little impact on the oversupply that persists in the industry. Investors will typically lose all of their investment in the company, and many employees will find themselves out of work. But the network assets will remain in one form or another. In some cases, the bankrupt company's creditors (who find themselves the reluctant owners) will agree to restructure the company's debt, allowing it to return to the market with a substantially lower cost structure.
Operators that emerge from bankruptcy may end up operating networks acquired at 1/10th the cost of building their own network and would then be able to set prices far below those of rival operators still bound by the discipline of debt payments. The chief executive of one network operator recently argued that bankruptcy "will almost become a business necessity in order to compete." If his pessimistic prognosis proves to be correct, prices may continue to tumble throughout the coming year.
Stephan Beckert is a research director at Washington, DC-based market researcher TeleGeography Inc. He can be reached at 202-741-0020 or email@example.com.