Larry Lessig Bets His Job

Larry is so sure of this solution, he's now betting his job on it. He wrote in his blog if a national law such as this is passed and it fails to significantly reduce the spam levels, then he'd quit his job.

The problem with Larry's proposal is that, at best, it papers over the spam problem, and at worst, spam levels will actually get worse. If this labeling law is passed, then I hope Larry has some other career skills to fall back on, because his teaching days may be over.

Larry's proposal initially seems appealing, and you may think it's counterintuitive to claim it would fail. Let me explain why.

Spam levels have been rising dramatically over the past year. In a recent article I documented nearly a doubling of spam levels through 2002. This increase generally is attributed to two things. The first is the economy. Spam is advertising with costs shifted onto the recipient, and in these tight times that's an attractive lure to desperate people. The second reason for the increase is the improved effectiveness of spam filtering.

As filtering tools get more effective at identifying and labeling spam, spam levels go up, not down. Spammers have to increase their sending levels to assure continued delivery of their message. The marginal cost of sending additional spam is nearly negligible, so they can keep ratcheting up the sending levels to ensure some amount of their spam gets through.

Larry's proposal would allow nearly optimal filtering. Larry is betting that this will squeeze the incentive out of spam, and so spammers will stop. A rational conclusion--but wrong. The problem is that spammers are greedy, stupid, sons-of-bitches. For goodness sake, we are talking about morons that believe chain letters are effective and legal. Or that a pill will make your penis bigger.

Most spammers don't care about efficacy, only cost. Data show that spam levels go up as filtering improves, so under Larry's proposal spammers would simply raise their spamming rates to levels previously unimaginable. Spammers won't mind doing that, because Larry's proposal provides no penalty for flooding mail servers, only for omitting the spam label.

The problem with spam filtering is that it is a salve for the symptoms of spam. It may prevent spam from landing in your inbox, but it does not relieve you of the cost burdens of spam. This is illustrated by the network utilization graphs for my mail server. The blue plot shows outbound traffic and the green plot shows incoming data. Notice how that green plot is a solid mass of noise. That's spam. Most of it is being filtered away, but I still have to pay the price of bandwidth and processing. Under Larry's proposal the area under that green line will explode to volumes unimaginable, and so too will my costs.

For those spammers that do care about efficacy, I question whether $10,000 is a suitable deterrent. National ad campaigns probably spend more money on ink. Many spammers will eat this as the cost of business. Do you think the infamous fax.com junk faxers would have been deterred if they were slapped on the wrist for ten grand instead of reamed for five million smackeroos? The FTC was able to do this because junk faxes have been outlawed--just as junk email should be.

If spam labeling becomes law you may see less spam in your mailbox, but you'll also find your ISP bills skyrocketing to pay for the massive infrastructure to handle the hidden--and filtered away--flood of spam. I hope a spam labeling law isn't passed, because I fear the spam flood that will result. And I don't want Larry to lose his job.