External links

Full text: PDF

Information about this paper

Abstract

Spam is commonly defined as unsolicited email messages, and the goal of spam filtering is to eliminate these messages. In this paper, we argue that not all unsolicited email messages are useless to recipients, and dropping all unsolicited messages is not optimal for either the recipient or the sender. Increasing the sender s cost (with stamps, puzzles or tarpits) has been another approach aimed at discouraging senders from generating large volumes of spam. But the optimal level of cost to be imposed upon the senders, the most important factor in this approach, has yet to be determined. In this paper, we propose a combination of cost and filtering approaches: charging senders of massive unknown messages or unsolicited messages based on aggregated filter information.1 We present both an economic model to calculate optimal surcharges under different conditions and the results of a simulation meant to study the consequences of different charging mechanisms. We use theory and simulation to show that our differentiated surcharge mechanism can improve social welfare compared with a stamp or a perfect filter. Our work can also shed light on how much extra cost we should impose on senders to regulate incoming email messages, which is a fundamental factor neglected in previous research.