Smart Contracts at the World Bank

The concept of smart contracts has been around since the 1990s. The basic idea is that contracting parties would reflect some part of their obligations in computer code. This code would be able to recognize (or be told) when conditions for action had been met, or not, and then perform the obligations (execute the contract) or impose penalties for failure to meet the conditions.

This seemed to promise significant reduction in transaction costs, notably of monitoring to see if conditions were satisfied and of execution of the obligations. (Other transaction costs of business, like finding potential deals and negotiating their terms, were less subject to automation, though being able to offer modular transactions, computer-ready rather than shovel-ready, did offer some economies.)

The goal of cutting out expensive administrative intermediaries seemed to be brought closer by the advent of distributed ledger technology (most commonly, the blockchain). If the rules of enforcement could be entrusted to an impersonal and independent network of nodes, they could be removed from the risk of human judgment and delay.

Not all smart contracts must be on a blockchain, and certainly not all blockchains need to feature elements of smart contracts, but the two phenomena are frequently associated.

One large class of intermediaries that advocates of smart contracts and the blockchain target for elimination is financial institutions. If money makes the world go round, the bankers have traditionally had a big hand in the spinning – for a profit. (And, as noted in my previous column on Smart Contracts on Slaw, financial institutions have a significant share of blockchain-related patents. They are not sitting idly by waiting to be made obsolete.)

As the world’s economy goes digital, financial technology (fintech) grows quickly, inside and in competition with banks. Can fintech ride the blockchain and smart contracts to displace the banks and make the economy more economical? And if it is more economical, may more interests participate?

Making opportunities to participate in the money-driven economy is known as financial inclusion – bringing in participants who did not use to have much of a role, and certainly not much of a voice, in financial decisions.

Such matters are of interest to the World Bank, a key institution promoting (sustainable) development around the globe. A recent publication, “Smart Contract Technology and Financial Inclusion” reviews the state of knowledge and poses the key technical and policy questions we all face in making the economy work better and more fairly.

The study is an excellent review of the issues, in plain language, concise but thorough.

The study says in leading off that “smart contract usage is largely conditional on blockchain adoption.” It goes on to say, however, “[a]lthough blockchain optimism is strong, applications are incipient.” (p. 1) That is a very diplomatic way of saying that the blockchain is not yet all that it has been made out to be. Its promises remain to be kept, despite years of heavy investment in many places. In fact, fewer businesses had blockchain projects under development in 2019 than 2018.

However, the focus of the publication is on what might be, not what is. It is a very competent description of how smart contracts can be integrated into blockchains and how the resulting transactions can fit into the digital economy. It raises a series of “implications” for financial inclusion and also policy considerations for “responsible and smart deployment” of smart contracts.

The basic notion of a contract and how it is made smart via the blockchain is set out in four basic stages:

  1. making a deal in the first place and coding its terms for performance;
  2. having an agreed-upon party “validate” the contract and record it in a blockchain/distributed ledger;
  3. the smart contract collects data on operating conditions – data feeds and so on (the source often being a so-called “oracle”);
  4. the contract executes or expires (with or without penalty.)

It is noted that different types of blockchain, notably “permissioned” or “permissionless”, will affect how validation is done, and the expense to the system of this stage.

A discussion of transaction costs underlines that enforcement costs are the main ones to be saved by smart contracts , through making performance more likely and breach more costly, by automating monitoring and verification, and by making monitoring transparent to all parties in real time. It shows how these savings can benefit lenders, suppliers and their customers.

A chapter on legal considerations notes that smart contracts can meet civil law and common law rules on formation of contracts, though requirements for jurisdiction and some enforceability rules may be more problematic. Questions of admissibility of evidence of the codes used can be difficult and “will vary across jurisdictions.” (p. 14)

Cutting out intermediaries may make amendment of contracts difficult when circumstances change.

“Smart contracts will not prevent fraud, illegality, or unconscionability in the formation process. [they] will not eliminate the need for dispute resolution and third-party legal intervention.” Thus “the combination of smart contracts and blockchain represents a ‘garbage-in-garbage-out’ system.”(p 14) Having secure content in a blockchain is little help if the content is false.

If tainted smart contracts cannot be changed before they self-execute, the legal system may need more “post-performance” remedies than in today’s system, when remedies tend to be sought to compel performance or compensate for its absence.

In the short term, smart contracts are most likely to complement rather than replace natural-language contracts, serving as a payment mechanism and possibly setting up to execute unambiguous if/then statements. These provisions could be written in code, while areas requiring human judgment would be expressed in natural language.(p. 15) It will be difficult, i.e. time-consuming and expensive, to put into computer-readable form all possible contingencies in performing contracts. “Ironically, ambiguity can in some cases enhance efficiency by allowing parties to contract now and dispute later if something goes wrong.”(p. 15)

The study goes on to review in this context financial consumer protection, customer due diligence, how to make “foundational legal determinations” (like the elements of valid and binding contracts, challenges to validity, addressing deficiencies and mistakes, and other topics), standardization and vetting of smart contract terms by relevant authorities and stakeholders (a way of making them acceptable, especially as contracts of adhesion), and the automation of data sources (oracles again).

The operation of smart contracts in supply chain finance and insurance is said to be likely to increase financial inclusion, but its benefits for consumer credit transactions are less obvious. The work concludes with an annex on “how select microfinance products would change if smart contracts are used in their deployment.” The annex is a fascinating analysis of risks and opportunities from the point of view of different participants in international microfinance operations.

It is clear from this relatively brief but well-informed publication that aside altogether from the speculative aspects of the blockchain, the deployment of smart contracts raises a number of issues on the financing side of transactions and even more on the legal side. Pages 3 and 4 of the document set out the “key takeaways” on each of these aspects and are worth a look even for those who do not have the time or motivation to digest the whole publication.

Brave new worlds often take time to materialize and may turn out to be less brave and less new than advertised. No doubt we will have smart contracts in blockchains in due course, but the way there is not a smooth one.


  1. Thanks for this overview. I’ve heard my share of grand proclamations that smart contracts will be the end of contracts/contract disputes especially in my field, insurance. But I think what is not fully appreciated is the oracle challenge: do you have a single source of truth you can build a product around. You have a challenge for most products to find an oracle just for triggering coverage. But even if you can solve that, you have an additional challenge of an oracle for determining loss. It’s this second one that is often glossed over. Absent a good loss oracle, to make it work, you have to either over- or under-compensate people for their loss. People probably won’t accept being under-compensated; practically speaking you’d have to over compensate them. This would raise the cost, in most cases dramatically such that it doesn’t outweigh the efficiency gains. Today I think the universe of suitable products where this could work would be small.

  2. Being sure that a natural language contract has the effects that the parties intend can be difficult and often requires legal skills.

    Being sure that a computer program will operate in the way that the parties intend can be difficult and almost always requires technical skills.

    Thus the making of smart contracts carries an overhead which its gains must exceed. Is there convincing evidence that the game is worth the candle?