Hidden Costs of Verification

Within the blockchain community many identify its key technological advantage as its ability to create a consensus regarding the veracity of a set of transparent records.

This benefit was expressed in economic terms by Christian Catalini (MIT) and Joshua Gans (Univ. of Toronto) in their working paper “Some Simple Economics of Blockchain” as a reduction in the cost of verification.

However, as Catalini and Gans point out, blockchain does not reduce the costs of verification equally for all types of transactions. It is important to distinguish between practical applications which can be executed using today’s technology and those for which we must still develop workable solutions. In applying some key economic concepts, we can better understand these distinctions and current limitations in order to help us move this technology forward and expand its practical cases of use.

Information can be considered verified when all relevant parties agree on it. This is why consensus is a pivotal concept in blockchain. With that said, the most important step in the blockchain verification process is making its essential information publicly available. At first glance, this might seem like a one solution fits all problem. However, not all information is equally easy to make public: some types of information are more excludable than others. Excludability is the degree to which access to a piece of information may be impeded, and it is important to understand as it distinguishes types of information for which validation costs and procedures will differ.

An example of a piece of information that is virtually non-excludable is weather. If we bet on whether it will rain tomorrow in Central Park, you and I and any number of agents can go to Central Park and observe if it rains or not. While you do need to be present to directly observe this information, because Central Park is a public park, everyone does have access. Additionally, no one needs any special instruments to determine whether or not it is raining, even a child could do it.

Contrast this with a potential deal that depends on whether a building is structurally sound. Private property rights mean that the owner has the right to refuse entry, implying that this information cannot be observed by the public. Additionally, special measurement tools and expertise are necessary to make a determination on soundness based on what can be observed in the building. This means that even if access was permitted to all, the relevant information may still be privately held by those with the right tools and expertise, such as building inspectors. Those individuals could, then, exclude others from obtaining that information.

Even further along the spectrum of excludability is private information such as intent, which can be relevant for enforcing insurance and other types of contracts, as well as the law. While we may attempt to glean clues that are indicative of intent, we cannot directly observe a person’s intent (or do not currently have the technology to do so) unless that person chooses to truthfully reveal it to us.

It is no surprise that thus far, the most successful implementations of blockchain-facilitated transactions involve non-excludable information such as blockchain-native information and information that can be accessed by smart contracts through an API, since the cost of verifying this type of information is near zero.

The key to pushing the boundaries on where and how blockchain technology can be used will be to identify what type of information needs to be verified for particular types of transactions, and what institutions will be necessary in order to make that information accessible and verifiable.

When it comes to excludable information, there are two challenges to making that information public so that it can be verified on a blockchain. The first is self-censorship, or choosing not to divulge information that would be useful. Take used car sales, for example. Economists have long recognized buyers’ lack of credible information about quality as an impediment to market efficiency. While services such as Carfax have arisen to cope with this problem, the problem continues to persist because Carfax can only incorporate information made available by its sources. A car owner who has damaged their vehicle could get it repaired at a shop that does not report to Carfax, creating holes in the available information and reducing the overall value of a Carfax report.

Moving used car sales onto the blockchain would not eliminate, or even mitigate, this verification issue. It may help make any reported information more accessible, rather than held by a central authority such as Carfax; however, even that change could have an impact on an individual car owner’s willingness to report this vital information, and cause even less relevant information to be available in the first place. In this way, verification of this type of information on a blockchain platform could be more costly or difficult than verification using existing intermediaries.

In order to deal with an issue like self-censorship, platforms will have to encourage participation by users and trusted information sources, noting that gaps in participation can reduce information quality. If a blockchain platform for used car sales somehow managed to get all repair shops to use the platform, it could provide near complete information to potential buyers. This strategy might work better than relying on self-reporting of vehicle owners, both because there are fewer shops than owners and because shops have less conflict of interest than owners.

The second challenge to attaining quality public information is what economists call cheap talk. Anyone can say that their building, or one that they inspected, is structurally sound; anyone can say that they are a reliable seller: talk is cheap, and it is possible that information that is disclosed is not accurate. For example, a platform wanting to facilitate the sale of art will have to ensure that the pieces for sale are genuine, not fakes. The existing institutions that facilitate such sales rely on qualified experts as well as authoritative intermediaries, such as Christie’s and Sotheby’s, who select those qualified experts and maintain a reputation in which buyers can instill their trust; trust being the paramount factor to ensuring market participation.

Just as with self-censorship, a distributed ledger itself will not have the capability of eliminating the cheap talk problem. The information in this case is excludable due to required expertise, which means that simply letting market participants determine the quality of appraisers will not solve the problem as most of these participants themselves lack the expertise to determine whether or not the appraiser is being truthful. Therefore, in a market like fine art, if a platform does not select information contributors based on expertise, it may be unable to build a reputation for trust.

All of this means that for some applications, identifying and differentiating between different categories of platform participants and granting different permissions to different groups will be required in order to build participation sustainably.

Blockchain startups wanting to disrupt existing markets or improve on incumbent platforms via better public information must recognize that the cost of verification differs depending on the context in which blockchain is being applied. Each project will encounter distinct challenges in creating structures that will facilitate information flows.

Some useful questions to ask as you design the verification processes in your ecosystem include:

What kind of information is needed to facilitate activity on the platform?

  • Are there publicly available sources?
  • Is this excludable information?

Who should be allowed to contribute this information?

  • Who would be in a position to have this information?
  • Is expertise required?
  • Is there a potential for bias or conflict of interest?

How can we ensure that contributors tell the whole truth and nothing but the truth?

  • What requirements will be put on contributors?
  • How will information regarding contributors be shared with market participants?
  • What are the consequences if the rules are not followed?