Data is Capital (and why it matters)
Continuing posts on the economics of data, here we'll discuss the distinction between "capital" and "labor" and why it matters for understanding the economics of data. We'll see a notable paper that seems to argue the opposite, why I actually mostly agree with its message -- and how the "capital" perspective suggests alternative market-based solutions.
(If you're into this stuff, try not to get pedantic on me -- we just need some rough definitions!)
In economic terms, labor refers to work you do for payment, where your employer owns the results of that work. Examples: most jobs, whether hourly or for a yearly salary. A journalist who is paid a fixed amount to write an article.
Meanwhile, capital is a resource that can be used to create economic value. Examples: a factory, a truck, tools. The article itself, that was written by the journalist, could be considered capital: the company monetizes it in various ways. A house or apartment can be considered capital.
How Capitalism Works (and why it matters)
"Capitalism" refers to a system where capital is privately owned by individuals or e.g. businesses. (Popular discourse generally mixes capitalism up with free or competitive markets. However, in theory you can have either without the other, as demonstrated in Posner and Weyl's recent book "Radical Markets".)
The point is that owning capital allows one to accrue relatively unbounded wealth in relatively passive ways, while providing labor typically doesn't. The landlord, by virtue of owning the building, keeps the profit: (rent paid) - (cost of upkeep/taxes). Owners of a company, or shareholders, keep the profit between market price of labor, which is paid to employees, and market price of the products sold.
In other words, capital is where the game is at. This will be important shortly for how we think about the value of data.
(Not to get into it, but for background: socialism might argue this is bad because value is created by workers regardless of who nominally owns the factory, so they should get all the surplus; capitalism might respond that the owner took on the risk and that markets in private capital encourage investment; etc.)
Data is Definitely Capital (and it definitely matters)
We've seen recently that data can be hugely empowering for solving problems or accomplishing tasks. Products like Google Maps rely on tons of data, as does Google search, for that matter (including feedback data from user interactions). Personal data is used by a huge advertisement and recommendations industry. Etc.
So, data functions as capital. It is a resource that is used to accomplish economically valuable tasks.
(It is true that sometimes data is collected or otherwise produced by labor. For example, in crowdsourcing, humans may be paid a wage to label images to help train learning algorithms. Google Maps might employ people to drive around with cameras, etc. But the data produced is owned by the employer, and functions as capital.)
This matters because, as discussed above, capital = wealth. If personal data -- location data, interests, friends, demographics, shopping habits, etc -- if all of this data is valuable capital, that sharpens the question of whom people should give this data to, and in exchange for what. If we think we can create better economic systems (in the sense of creating more value or sharing it more fairly), this matters to the design of those systems.
So Why Do Smart People Say Data is Labor?
Last year, the aforementioned Glen Weyl (a collaborator of mine) wrote an American Economic Association paper on this topic with Imanol Arriet-Ibarra, Leonard Goff, Diego Jiménez-Hernández, and Jaron Lanier. (Read a version here.) Entitled Should We Treat Data as Labor? Moving beyond "Free", it argues that data should be thought of in terms of labor rather than capital.
Let's explore where the authors are coming from.
In political science, labor and capital have socioeconomic connotations that I haven't explored above. In particular, we can think of the large labor class that mostly makes their money from wages, versus the small elite capital-owning class that has significant wealth. Traditionally, the labor class gains power, rights, and money through collective action: forming unions, imposing regulation around safe working conditions and limited working hours, etc.
So the authors are aiming to leverage this history to motivate similar action around data rights. The thinking goes that the broad use of personal data is a modern example of exploitation of the labor class by the capitalist class, and labor should respond as it usually does to such exploitation, e.g. by passing the GDPR. A second component is that people should stop viewing data as a byproduct of their life, and start seeing it as a product with economic value that they can sell.
This part of the argument is fine (maybe depending on where your sympathies lie). However, from an economic standpoint, the phrasing of the message is still wrong: data itself does not function as labor economically, any more than a hard hat, hammer, or sickle.
So I agree with many of the paper's points: people's personal data creates a ton of economic value and they currently don't have economic control or ability to capture that value financially. And in line with the authors, this suggests to me that alternative systems can do a better job. But because I recognize that data functions as capital economically, my perspective on those alternatives looks a bit different.
Economics of Data as Labor vs Capital
To understand how my argument relates to that paper, it helps to find-and-replace their terms "Data as Capital (DaC)" with "current world's status quo", and the terms "Data as Labor (DaL)" with "proposal for individuals to make money from their personal data". Then we see two things:
- The issues the paper lists as consequences of Data-As-Capital, like how people's data is currently harvested en masse for free, are just the current status quo. They aren't inevitable consequences of recognizing that data is capital economically; in fact, we'll discuss alternatives below.
- The possibilities for people to be compensated for personal data aren't limited to treating data as labor economically, e.g. being paid a wage for producing data. Again, we will suggest alternatives next.
My point is that viewing data as capital, economically speaking, sheds light on how people might be compensated for their personal data.
In my previous post on data ownership, I suggested thinking about data rights similarly to copyright: a person has legal control over use of her personal data, and may be able to rent or sell the rights to use that data for specific purposes for a specific length of time.
If you think about it, this is exactly how capital functions in the economy. The owner of an apartment collects rents from its use; the holder of a copyright collects rents from others for use of the copyrighted work.
For example, it would be quite a blow against workers' rights if we argued that individuals can't own copyrights, and that they should just be paid a wage to produce creative works, which are then owned by a company. In the same way, failing to treat personal data as capital could be extremely disenfranchising. It sets up a transaction where people receive a fixed amount of money in return for a piece of capital (their data) whose value keeps on giving and increasing.
Owning capital makes people rich, usually by collecting rents. Since personal data functions as capital, creating legal rights around its ownership could be extremely empowering. Ordinary people could rent this valuable resource to corporations. This suggests care around systems where companies pay a one-time fee to gain perpetual ownership and use of personal information. The information can be useful over and over again, so it is unlikely for the individual to get the better end of that deal.
What I find most interesting about this topic is that it can be supported from many political and economic perspectives, but also attacked from many as well.
One can view this as an uber-capitalist suggestion: convert everything possible, even data, into private property, and allow for (regulated) markets in it. But this is a sort of subversive use of capital ownership in favor of a populist movement, i.e. to empower and enrich the "labor class". So I'm not sure what kinds of political viewpoints would criticize it.
Similarly, one can view this as an individual-liberties movement in the sense of "my data should be under my control, not a company's." But these data rights laws are literally limitations of rights (i.e. other people's rights to use my data). So it's quite anti-libertarian -- a similar expansion of private control to copyright. One can worry from the history of copyright and capital in general that such systems end up enriching and empowering the already rich and powerful. This is one reason I make the suggestion in the previous post that certain types of control over one's personal data would be legally impossible to surrender, e.g. all data cannot be rented for longer than one year at a time, and it cannot be resold at all.
 Some broad strokes being painted here: I'm not distinguishing between capital and natural resources such as land; for my microeconomic perspective, the difference does not matter. When I say capital, it is probably most accurate to think of means of production.
 http://radicalmarkets.com/ In particular, under the COST proposal, capital is owned collectively, but it is perpetually rented by individuals, who can buy and sell the right to rent from each other.
 A more blunt way to put it is that people who pay income tax aren't wealthy; wealthy people pay capital gains tax (if any). Anyone in the U.S. who gets really rich off of labor is most likely famous, an athlete or celebrity.
 This is why regular folks -- laborers -- are advised to invest in the stock market or owning a house: to become owners of capital.
 I like that the "Radical Markets" perspective draws from both of these, yet critiques both as well.
 By the way, my thinking and I believe everyone's thinking around this issue is evolving rapidly. And this paper is part of that evolution, so I'm really grateful to the authors for writing it and helping shape my perspective.
 By the way, that post was focused partly around privacy and the notion of the right to privacy, but here we separate out the economic question only.
 Or become weapons of censorship.