Teddy Sherrill CS199 Privacy and Technology A Crisis of Identity Anonymity and Decision Making in a Pseudonymous Economy Introduction The global media has been quick to label Second Life Linden Labs immersive virtual social environment the Next Big Thing with respect to electronic commerce and business communication However due in part to the inability of Second Life s users called Residents to verify each others real life identities when conducting commercial transactions counterparty risk remains high and many users express reluctance to engage in in world commerce worth more than several dollars at a time1 While some including several large business interests2 believe that Linden Labs must take an active role in verifying at least some users identities these elements are often met with passionate opposition from those who value the high degree of anonymity afforded to them while in character within Second Life This paper will first present a theoretical framework for approaching the anonymity accountability tradeoff3 in an online setting weighing various possible solutions by viewing them through the lens of economic modeling It will then consider the unique applicability of this framework to the metaverse or the collection of three dimensional virtual roleplaying worlds including Second Life and others Finally it will treat Second Life as a case study in psuedonymous commerce offering two key suggestions for reducing in world counterparty risk based on a combination of original analysis discussions with two prominent Second Life analysts and interview 1 Second Life interviews conducted as avatar GeorgeBataille Streeter May 9th 11th 2 Linda Zimmerman interview CEO of Business Communicators of Second Life 3 See Trading Privacy for Trust by J Seigneur and C Damsgaard Jensen for additional theory work conducted with Residents within Second Life itself Identity and Anonymity An Economic Approach In order to objectively consider competing options for identity management in the realm of online commerce it is important to develop a consistent theoretical perspective from which to approach the problem Our method here based primarily on the game theoretic perspectives offered in Friedman and Resnick s The Social Cost of Cheap Pseudonyms and Peter Kollock s The Production of Online Trust seeks to predict the effects of a given identity management scheme by assessing the economic incentives associated with each possible course of action by each agent in the system4 5 The goal of each analysis is to maximize the overall efficiency of a given closed pseudonymous economy by a minimizing transaction costs and deadweight loss due to the fixed costs of operation and b minimize monetary and participation losses due to malicious or unreliable behavior While the variables chosen for the papers quantitative modeling sections vary slightly they reach a similar set of conclusions about identity decision making and reputation The fundamental problem plaguing pseudonymous economies is that of serial pseudonymity This phenomenon which refers to a user s unchecked ability to constantly create and assume new identifiers renders pseudonymous functionally equivalent to perfect anonymity by preventing a given user from being trackable over time In game theoretic terms this means that each progressive stage of a closed simulation of the system must be treated as a new initial stage in which no user has any reliable information about any other user Since it will always be in a violator s interest to assume a new identity immediately following fraudulent or disruptive behavior sanctions and punishments quickly become meaningless flagged or tainted identities will simply stop being used and each remaining active user will necessarily have a clean or no history 4 Resnick and Friedman 1 15 5 Kollock 1 19 This means that the character of a given pseudonymous economy will be determined in large part by the generally accepted policy with respect to newcomers Since it is impossible to tell whether a reputation less identity is a true first time user or a criminal who has recently re spawned to avoid punishment the decision to transact with a new identity represents a considerable counterparty risk If users choose to distrust newcomers they prevent themselves from being defrauded but necessarily constrict their trading options and discourage new users from joining the economy and thereby growing the system as a whole If users choose to trust newcomers they increase their total number of trading partners but do so at the cost of an increased risk of being defrauded and potentially encouraging further misbehavior Needless to say both policies are inefficient loss heavy and far from ideal Several solutions have been proposed in order to alleviate the inefficiencies introduced by the newcomer problem The first of these discussed in some detail in The Social Cost of Cheap Pseudonyms is that of entry fees for the creation of a new identity While a required fee for identity creation almost certainly discourages the the frivolous creation of malicious accounts it also drives down participation as fewer casual users will be willing to try any system if it requires a monetary commitment before any use or benefit has been derived Again a lesser of two evils choice is imposed though this time on the fee setting authority higher fees lead to lower overall numbers and a smaller variety of goods and services in the economy whereas lower fees may be insufficient to deter malicious users from registering new identities That said it is important to note that this inefficiency may be minimized by the fact that a a user s willingness to pay a given entry fee depends on his expectations for the use or misuse of his account This is because in a healthily functioning pseudonymous economy it may be the case that a small entry fee improves the reliability and quality of the economy enough that over the course of an identity lifespan of some duration t the user can expect to recoup his entry fee in increased profitability This means that users who expect their identity lifespan to fall below t users who expect to misbehave will not recoup their entry fee whereas honest users will eventually do so This provides a user with an extra disincentive to pay the fee only if he plans to misuse his account and reduces the effect of the participation problem A variation on this concept is the use of entry bonds rather than fees which pay out after a