Friday, June 21, 2013

"Smart" or "Dumb," Value is the Issue for Networks

The chief business reason communications executives hate the term “dumb pipe” is that it has the connotation of “low gross revenue” or “low margin” earned by services without distinctiveness. The term implies that communications services are “commodities.”

The real problem therefore is not whether the “pipes” (networks) are dumb or smart, but whether the profit margin is high or low.

So the problem with competitive markets and new technology often is that profit margin gets wrung out of a business.

That appears now to be the case for high-frequency trading, which relies on ultra-low latency communications connections. At least some communications service providers care about high-frequency trading because they supply the ultra-low latency connections that help provide the trading advantages.

Fast, powerful computers and algorithms are the primary driver of high-frequency trading. But acting on what the algorithms suggest is where the low-latency connections come in.

Trading firms have spent millions to maintain millisecond advantages by constantly updating their computers, collocating in data centers and connecting distant computers using low-latency networks.

Of course, once the trading exchanges saw how valuable thousandths of a second were, they raised fees to collocate, and hiked the prices of their data feeds.

“Speed has been commoditized,” says Bernie Dan, CEO of Chicago-based Sun Trading, one of the largest high-frequency market-making trading firms.

And that is precisely the problem: when real advantage is seen, a competitive market tends to reduce the value of the advantages when all competitors adopt the latest technology and approaches.

But the economic downturn is a factor. Overall trading has declined since the 2008 Great Recession, and high-frequency trading might now represent about half of all U.S. trades, according to the Wall Street Journal.

At one point high-frequency trading represented more than 80 percent of transactions, according to the Financial Times.

But U.S. stock-trading volumes declined since at least 2010 and in 2013 are running 35 percent below the industry's peak in 2009, when an average 9.8 billion shares changed hands a day, according to Sandler O'Neill + Partners.

Precisely how much additional value service providers can create in their networks is a legitimate issue. Ultra-low latency networks to link exchanges are one example of how even “dumb networks” can add value.

The problem isn’t whether the networks are dumb or smart: the low latency networks are no smarter than the “normal” networks. They simply use the shortest routes.

Value is the issue.

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...