Skip to main content

When do you need anonymity?

Do we want anonymity in payment systems or not? I think not, but who knows. Even those who are actually thinking about this sort of thing are not sure. A few years ago, the US Government Accountability Office published a report on “Emerging Regulatory, Law Enforcement, and Consumer Protection Challenges” (May 2014) and the first of its conclusions was that virtual currency systems “may” provide greater anonymity than traditional payment systems. Well, as was famously observed in one of the greatest British films of all time, Withnail & I, they may provide a pork pie. It’s a question of design.

The design I want is strong pseudonymity, essentially.

I am always interested in the issues around anonymity and pseudonymity in payments and I try hard to collect a variety of different use cases so that I can test proposals for new payment systems to examine their behaviour in very different circumstances. While I was mulling over some issues to do with different kinds of pseudonymity and who might be responsible for delivering them into the marketplace, I came across a very interesting circumstance where “strong” pseudonymity might be needed: lottery winnings.

First, the basic, revolting, economics of American lotteries (ours are, I’m sure, no different). Around half of all Americans buy at least one lottery ticket every month, but there’s the usual 80/20 rule in the distribution. In other words, most tickets are purchased by the same small group of people. These people tend to be poor and uneducated, which is why I think it reasonable to call these lotteries a despicable form of regressive taxation that falls mainly on the mathematically-challenged (in a survey a few years ago, a third of players who had not completed high school identified buying lottery tickets as a “wealth building” strategy - which it is, but only for the people running the lotteries). How big is that regressive tax? Huge. One study found that US households that make less than $12,400 per annum  spend 5% of their income on lotteries, another found that households earning under $13,000 per annum spend 9%.

Rant over. But why am I ranting about lotteries at all? Well, because they are an interesting locus for a discussion of anonymity with a central use case that I hadn’t thought about before. I was pointed in this direction because of Don Thibeau’s tweet concerning the winner of a HALF A BILLION DOLLAR lottery prize in the United States who was involved in a  court case (as Jane Doe) to remain anonymous because she doesn’t want everyone to know about it.

Now for me this wouldn’t be a problem. If I won half a billion dollars on the lottery I would collect the money and then very publicly donate all of it to the Monster Raving Loony Party in return for becoming their leader and then dedicate the rest of my life to getting more MPs than the Greens or UKIP (ie, one) so that I could be on Question Time. But other people might decide to keep the money, and this can lead to problems.

In November 2015, Craigory Burch Jr. matched all five numbers in the Georgia Fantasy 5 drawing and won a $434,272 jackpot, The Washington Post’s Lindsey Bever reported. Two months later, police said, Burch was killed in his home by seven masked men who kicked in his front door.

From All she has to do to collect a $560 million lotto jackpot is make her name public. She refuses.

So you can see why, apart from trying to avoid a stream of huggers, chuggers and muggers, the woman in question would like to keep her good fortune to herself. Would she really be anonymous though? After all, the money would have to go into a bank account, so not only would lottery officials know who she is but people at the bank would know who she is, and so on. Being anonymous is really difficult in an infrastructure that has no anonymity. Which leads on to an interesting question: if we are designing the identity system of the future, should it allow for this kind of anonymity? It turns out that New Hampshire actually allows people to form anonymous trusts and these trusts can buy lottery tickets. Again, though, would the trust members really be anonymous?

You could of course construct the lottery to be completely anonymous from the beginning by using a variant of the cryptographic blinding invented by David Chaum for Digicash. That is, you buy a lottery ticket, fill out the numbers and add your ZCash, Monero or whatever address and then submit it with a blinding factor. The lottery signs the ticket to confirm your numbers and sends it back, at which point you divided out the blinding factor to give yourself a completely anonymous, but completely valid, lottery ticket. If you win the lottery, you can present the ticket by sending it to the lottery owner and the money can be sent to your cryptocurrency address without the lottery owner having the slightest idea who you are.

Here’s the thing though. Would you want lotteries winnings to go to anonymous people? How would you know that the lottery is fair? How would you know that the lottery operator isn’t channeling winnings to their family? There must be a way to audit, and this of course points towards pseudonymity as the only realistic way to manage privacy at scale in an online world.

Comments

Popular posts from this blog

Euro area card payments double in a decade

xxx "The number of card payments in the euro area have more than doubled in a decade as consumers increasingly dispense with the hassle of carrying notes and coins, according to the latest statistics from the European Central Bank. In 2018, card payments accounted for almost half of the total number of non-cash payments across the single-currency area. Credit transfers and direct debits were the second and third most common non-cash payment methods, accounting for approximately 23% each, while e-money and cheques together made up around seven percent. However, the relative popularity of each type of payment service still varies widely across euro area countries. In 2018 card payments accounted for just over 70% of all non‑cash payments in Portugal, compared with around 23% in Germany. The stats show that the number of card payments made by consumers and businesses has more than doubled in the last decade, with an average of 121 card payments per capita in 2018, compared with