Filed Under: People, Social Media and Organisations

Identity might work better bottom up

Leave a Comment

Dgwb blog white border

One way to help people obtain financial capital is by helping them to build up social capital.

Forum friend Kosta Peric from the Bill and Melinda Gates Foundation (BMGF) Financial Services for the Poor programme recently picked out four technologies as being of particular interest right now. He pointed us to:

  • bitcoin-inspired distributed systems,
  • open APIs (application programming interface) as a new way to consume business services on the internet,
  • crowd-sourced identity schemes, and
  • open source hardware and applications.

[From Four Technologies That Will Revolutionize Financial Services | copernicc]

I’m sure we’d all agree with his views on blockchain technologies and the “Amazonisation” of financial services organisation through APIs (I don’t know enough about open-sourced hardware to comment) but I think his point about crowd-sourced identity is especially interesting, as it points to a shift in the way that identities are created, managed and used. And, since I’m rather obsessed with identity right now (as our clients should be and, in some cases, are too) I thought I’d take the time to explain why I agree with him.

First, look at what the conventional, top-down notion of identity means. It means someone (the government, generally speaking) must find some way to assign an identity to everyone who needs one, record who those identities have been assigned to, and check that when the identities are presented they are genuine.

It is almost certain that the government is having difficulties establishing who is a genuine citizen purely on the basis of identification papers produced by the existing system. Some of the illegal immigrants caught in the current security swoop have Kenyan ID cards and passports but their details are not in the national database.

It has been claimed that immigration and provincial administration officials at all levels have enriched themselves by selling these sensitive documents while compromising national security.

[From KENYA: Kenyans to apply for digital identity cards, says Ruto]

There are problems with this top down approach. Apart from being expensive, it is also vulnerable. One a false identity has been entered into the system, it is no longer false (if you see what I mean). As a consequence, obtaining such an identity becomes an essential precursor to crime as well as legitimate use and therefore the identities are obtained by all sorts of people who are not supposed to have them and the system is subverted. Managing and protecting the database at the heart of this scheme is complicated and difficult. In a great many emerging markets, in particular, the national identity scheme is soon degraded: sometimes because of corruption, sometimes because of carelessness, sometimes because of errors in the concept and design. This is precisely what has happened with the Aadhar scheme in India.

What was supposed to be a unique identification number providing identification and access to a host of government benefits and services, ‘Aadhaar’ has almost unvaryingly been extended to anybody residing within Indian territories. Almost anyone, be it Indian or an illegal immigrant can get an Aadhaar Card made without any proof of identity. More importantly, they get a numbered identity.

[From Sting reveals Aadhaar documents forged for Nepal, Bangladesh citizens – IBNLive]

So how does the alternative, crowdsourced version of identity take us forward? Well, if national identity schemes don’t provide “real” security then why bother with them? Save the money. At a basic level crowdsourced identity means asking everyone who you are rather asking anyone (e.g., the government) who you are. Whereas we are used to the idea of identity as something that is granted to us by a third party, such as the government or a bank, and the idea of an identity based on reputation that grows up through our networks and long-term relationships seems rather different.

Compare the two kinds of identity and their functionality in practice. Crowdsourced identity may seem a poor substitute for national identity at first glance, but it seems to me that Kosta is onto something here for two specific reasons that I have touched on before. The first is that this kind of reputational identity is actually better than conventional national identity because it is much harder to forge or counterfeit. A good friend of mine told me a story about an industry event he attended recently where he ran into a chap late at night when he was going back to his hotel. The guy was in the hotel lobby and recognised my friend as he had been a speaker at the event. The man explained that he had been tricked by a woman in a bar into following her back to hotel room where he had been drugged and robbed. He had no money and was too embarrassed to call his wife and asked if my friend might loan him some money so that he could get home and would report his wallet lost on the train or something. My friend had never met the man before but asked him his name and who he worked for and then looked him up on LinkedIn. Having established that not only did the fellow have a full LinkedIn profile but was actually connected to my friend via several different people, my friend loaned him the money which was, of course, gratefully returned a couple of days later. Now imagine that the unfortunate chap had instead presented my friend with his Portuguese fishing licence: how would my friend evaluate that and assess the strangers plausibility from that official document?

The second reason is that these crowdsourced identities may well be far cheaper to establish and this is especially true, and especially valuable, in the developing world where official infrastructure may be unreliable at best and non-existent at worst. Here the particular combination of mobile phones and social networks is especially powerful, because mobile phones tend to deliver not only unique identity but transactional history to go with it and this can be linked through social networking in powerful ways. You might have listened to the podcast I recorded earlier this year with Shivani Siroya and and heard a very good example of this where the transactional histories from mobile payment accounts are slurped up by organisations who provide alternatives to conventional kind of credit reference agencies that we are used to in the developed world.

Shivani Siroya is currently the CEO and Founder of InVenture. InVenture facilitates financial access by providing simple mobile accounting and credit scoring tools for offline and unbanked individuals, the subject of this podcast.

[From Media – Consult Hyperion]

Taken together, I think these provide compelling support to Kosta’s intuition and it strikes a that, to use Jaron Lanier’s term in “Who owns the future?”, the “economic avatars” that arise at the intersection of the mobile phone and the social network may well prove to be more useful to a great majority of the world’s population than their “official” identities even if they have them and indispensable to them if they do not. And, by the way, if you regard the whole idea of giving people credit on the basis of social capital as ridiculous and fanciful, I guess you didn’t see this:

[Bogota] where Lenddo introduced a “social network” Visa card to 100,000 of its customers yesterday afternoon. By 4:00 p.m. today in New York, where the online lender for developing countries is based, more than 1,000 Colombians had applied for the card. Lenddo CEO and co-founder Jeff Stewart calls it the first time ever, anywhere, that approval for a credit card is based on applicants’ reputations on Facebook, Google, LinkedIn, and Twitter.

[From This Emerging Markets Credit Card Is Backed by Facebook Friends » Techonomy]

Identity is the new money, as they say. Well, as I say.

3 thoughts on “Identity might work better bottom up”

  1. iang@iang.org' Iang says:

    On identity schemes. While the central point that government identity schemes are not all they are cracked up to be is well made, it doesn’t mean that we can do better with some random crowd-sourced scheme. We have to look at why there is a problem with government schemes, and without this understanding, there isn’t much expectation that we won’t then make the same mistakes.

    Identity schemes typically stop at the point of the name; they leave the problems to (fast windmilling hands seen here) another party. A well crafted identity scheme includes a way to establish the value at risk and mitigate harm to it, and this is needed for finance, because there is … value at risk. Without this attention to contract and theft, the schemes touted are pretty by not really strong enough to support finance.

    And, as a specific conclusion, I’d say that anything that is crowd-sourced cannot be strong enough, because it is the crowd by definition. So why bother? We need more precision, such as is found in a tightly designed social network. For specific meat and a template, don’t look at Facebook, rather look at CAcert’s Assurance Programme which can be adjusted to meet the needs of a financial network.

  2. mike@joyce.net' Michael Joyce says:

    In looking at top-down vs bottom-up identity, I think it’s best to think about the reason for wanting to verify an ID. If the reason for KYC is for creditworthiness, protection from default, then crowdsourcing is probably a great approach. Similarly for fraud, the provider would need to look at the risk-reward equation of creating a fraudulent profile to determine if top-down or bottom-up verification works better. Is it easier to make a fake ID, or to create a robust-looking friend network on facebook?
    However, if the reason for the Customer Due Diligence is for counter-terrorist-financing or anti-money-laundering, the risk-reward calculation is very different as it starts to become a compliance rather than a risk decision. Using bottom-up or crowd-sourced identities here will only work if your regulator truly understands “risk based assessments”, and you are confident you have sufficient information to provide law enforcement agencies in the event of an investigation.

  3. sy@sytaylor.net' Simon Taylor says:

    and of course, to create a true bottom up identity service you need a technology that works by consensus, is owned by nobody and maintained by everybody…

    Wonder what that could be? 🙂

Leave a Reply

Your email address will not be published. Required fields are marked *

Tags: , ,