Do Spammers Deserve Death?
Spam—messages sent in bulk to people who did not ask for them—imposes several costs on its recipients. First, the messages have to be transmitted, processed, and stored. Second, because there is so much spam, email providers have to build and operate filters to remove it. Users ultimately pay for those systems, directly or indirectly. Third, the filters are probabilistic: some legitimate messages are inevitably classified as spam. Finally, whatever gets through still costs the recipient time to recognize and delete.
The cost of any one message is tiny. The problem is scale. In 2012, Justin Rao and David Reiley estimated that spam was costing American firms and consumers almost $20 billion a year (see here). Of that, American users were losing \$12–20 billion a year in time to spam, based on five seconds per message and a value of \$25 an hour. hat is 480–800 million hours a year—roughly 55,000–91,000 person-years, or about 700–1,150 American lives at 2012 life expectancy.
Against these costs, spammers and spam-advertised merchants earned only about $200 million worldwide, an externality ratio of roughly 100 to one.
Yet recipients’ lost time plays remarkably little role in criminal punishment. Jeffrey Kilbride received 72 months, the longest surviving federal sentence I found, after he and a partner sent millions of pornographic spam messages using falsified headers and domain registrations. But his convictions also included fraud, money laundering, and obscenity, and his sentence was increased for obstruction. Robert Soloway, the “Spam King," sent tens of millions of messages using forged headers and botnets and falsely sold customers supposedly opt-in address lists. He received 47 months for mail fraud, criminal spam, and failing to file a tax return.
Why does the largest cost of spam—the time taken from recipients—figure so weakly in punishment?
One answer might be that the loss to each recipient is too small to count for much. That would make sense if law generally assessed harm one victim at a time. In fraud cases, losses are added across victims, and the Sentencing Guidelines have long allowed judges to estimate total loss from the approximate number of victims and the average loss to each. A scheme that takes $100 from a million people is treated as a \$100 million loss.
A second possibility is that time is harder to value than money. An hour does not have the same value to every person or in every circumstance. But economists and governments deal with this problem routinely. Rao and Reiley valued recipients’ time at \$25 an hour; the Department of Transportation assigns dollar values to travel time and delay, including time spent waiting for flights.
Another explanation lies in how the crime itself is defined. Federal criminal spam law, 18 U.S.C. §1037, does not criminalize unsolicited bulk email as such. It targets unauthorized access, falsified headers, concealed origins, false registrations, and related conduct. The basic maximum is one year, rising to three years for aggravating factors such as very high volume or at least \$5,000 in aggregate loss, and to five years when the offense furthers another felony or follows certain prior convictions. The law is built mainly around how the spam was sent, not around the aggregate amount of recipients’ time it consumed.
That suggests an obvious fix: define the offense around unsolicited bulk email itself, rather than around forged headers, unauthorized access, and related conduct. But that creates a different problem. A broad ban on unsolicited bulk email can reach political, religious, and other protected speech. Virginia discovered this in Jaynes v. Commonwealth. Its Supreme Court overturned Jeremy Jaynes’s nine-year conviction because the state’s anti-spam law was broad enough to prohibit protected anonymous political and religious email.
But protected speech does not imply an unlimited right to make unwilling people receive it. In Rowan v. Post Office Department, the U.S. Supreme Court upheld a law allowing a householder to stop further mail from a sender. The speech could remain protected while the recipient retained the right to say no. Protecting what someone may say is not the same as giving the speaker an unlimited right to impose the costs of delivering it on everyone else.
Thus, none of the explanations quite fit. We have a potential hole in how we prosecute shaving a few seconds from millions of people.