Power Arrangements in Identity Systems

Since the launch of the Sovrin Foundation, Phil Windley has been blogging a lot (no, reallya lot and more, more, more, more, and more) about how self-sovereign identity works and can be used. His most interesting and accessible post for a liberty-minded identity-layperson might be “On Sovereignty,” in which he briefly lays out what it means to have a “self-sovereign” identity.

Sovereignty over your identity doesn’t mean having complete control over information about yourself, but it puts you in a peer relationship with others, including the larger organizations we deal with, such as governments. “The beauty of sovereignty,” Phil emphasizes, is the “balance of power that leads to negotiations about the nature of the relationships between various entities in the system.” I want to expand on this notion that there are power arrangements in identity systems.

In a centralized identity system, the identity provider (such as your Department of Motor Vehicles) determines whether you can assert information and what you can assert. Centralized systems also often share information about you, or facilitate such sharing, whether you want them to or not. Implementation of the REAL ID Act would essentially move these powers from state governments to the U.S. Department of Homeland Security.

A self-sovereign identity system, on the other hand, gives you power to assert information about yourself, which others may accept or reject. It also better positions you to decline to share information about yourself. Those powers are important.

“Power” is an elusive concept. We’re more familiar with talking about power in terms of political and legal arrangements, such as how the Constitution gives certain powers to the U.S. federal government or denies all U.S. governments other powers. But absent these rules, “pre-political” power is simply the ability to do something or act in a particular way, or the capacity to direct or influence the behavior of others or the course of events. Power comes down to what resources you can bring to bear in going after what you want.

To illustrate pre-political power arrangements, look at physical power in the “state of nature.” If a tall, muscular man meets a short, skinny man on a trail in a desolate forest, the arrangement of power between the two is pretty clear: the bigger man has more ability to use force, so he can take things from the smaller man, do him in, or whatever he pleases. Give the skinny man a gun, though, and the power allocation reverses. The gun’s capacity to defeat human musculature gives the small man more power to defend himself, to take from the larger man, or to do the larger man in. Physics and technology allocate power before legal rules come into play.

Pre-political power arrangements seem to strongly influence what political and legal rules can achieve. It’s taken hundreds of years to bring the power of physical force possessed by feudal lords and governments relatively under control. This is probably because it’s pretty great to have power, and pretty awful to lose it.

It follows from all this that liberty can be advanced not just by fettering governments’ powers, but by allocating power more evenly in the first place. There aren’t going to be “do-overs” about how physical arrangements allocate power, but the ways that information arrangements allocate power have yet to gel. Balanced power in identity and information systems are important to pursue now.

Instead of the large man and the small man, imagine two scenarios where a government agent meets a local resident on that same trail. It’s a post-political world, so the government agent has to live within decent rules, but there can be very different allocations of power based on information arrangements.

In the first scenario, the government agent knows nothing more than what he sees: an ordinary resident making her way up the path. Without information to justify anything else, he bids the resident good day and they pass one another.

In the second scenario, the government agent can identify the resident and pull up records about her. Instead of sharing brief courtesies, the government agent notes debts owed on some parking tickets and a tax bill in arrears. With information about where the resident and her friends live, and data about them, the agent might infer that the resident is on her way to visit people with criminal histories. Financial details in hand, the government agent finds that the quality of clothing and jewelry she wears suggests untaxed or illicit profits on her person. In these ways or countless others, identity and biography empowers the government agent to interfere with the life of the hapless local resident by questioning, detaining, or arresting her.

Think also of the power imbalance when the government can decide when to allow someone to identify him- or herself and when to deny it. In the not-too-distant future, identity systems will all be digital and electronic. This means your centralized identity provider, the DMV, will be able to shut off recognition of identities that have unpaid parking tickets, tax liabilities, or other defects in the eyes of official policy.

Many people like this power when used to deny licenses to illegal immigrants today, but the use of the power is likely to broaden. Linking the vote to a national ID, as the Carter-Baker Commission proposed some years ago, could be one direction. Conditioning access to health care, financial services, housing, and more on the showing of government-issued ID have all been raised seriously in recent years. The power held by identity providers—now mostly governments—are only going to grow.

The solution is to disperse that power—not through rules but through technical infrastructure design. Let people assert their own identities without requiring it to be done through intermediaries, governmental or otherwise. There is plenty of capacity in a distributed identity system to prove the facts that make identities reliable, even while they minimize the data that needs to be shared in doing so.

The Sovrin Foundation has a good library of documents describing how they would do this, using a public, permissioned blockchain. And for those who are just learning the importance of identity—say, through this blog post—one entry-point to the basic concepts might be my book, Identity Crisis: How Identification is Overused and Misunderstood.