In this edited version of her recent paper, Book-Smart, Not Street-Smart: Blockchain-Based Contracts and the Social Workings of Law (which can be linked to below), Karen Levy examines blockchain-based contracts and the social workings of law
There has lately been much excitement about blockchain technology, and one of the chief avenues of optimism concerns its application to so-called ‘smart contracts.’ Smart contracts are agreements that utilize the blockchain––a digital ledger, distributed across a network, that securely records transactions between parties––to automatically and securely execute obligations when certain conditions are met; for example, a smart contract might automatically transfer ownership of an item when it is fully paid for, upon the passage of a set period of time, or upon the meeting of any other predetermined condition that can be instantiated in computer code. Because they are based on code, smart contracts can be immediately and automatically effectuated, without reliance on manual transfer or the intervention of institutions like courts. Smart contracts have recently been heralded as potentially transformative legal instruments that can do away with problematic ambiguity and expensive enforcement.
Like other blockchain-based technologies, the smart contract is designed to function without reliance on a centralized authority. In financial applications of blockchain technology (like cryptocurrency), the institution obviated is the system of state-issued currency and centralized banking infrastructures; smart contracts are designed to self-enforce without recourse to courts. In this sense, the smart contract can be understood alongside a number of other emergent technologies that aim to automate the enforcement of law —a vein that, read broadly, includes such seemingly disparate tools as digital rights management systems and traffic ticketing cameras. These systems are frequently received with much fanfare by policymakers, and heralded as a way to make the application of legal rules and agreements more consistent and efficient.
Though smart contracts may serve these goals in some contexts, I suggest that we should temper our enthusiasm about the transformation of modern contracting practice. This is because smart contract boosters tend to understand contracts chiefly as technical artefacts, rather than as social resources. Under this view, contractual agreements are bare transactions that can, and should, be optimized through code; term ambiguity and enforcement costs are understood as inefficiencies that plague the system of exchange. Understanding contracting practices within broader social and relational contexts, by contrast, reveals that contracts are in fact much more than this: they ‘work’ in many ways and accomplish many aims that are unaccounted for by the smart contract framework.
Blockchain-based Contracts and the Promise of Code-based Enforcement
What makes a smart contract ‘smart’? The underlying technology enabling smart contracts is the blockchain, a database distributed across multiple computers. When transactions occur, records of them are aggregated into blocks of code that are appended onto the chain (the ledger of transactions). The chief benefit of this structure—which is public, and which uses sophisticated cryptography to ensure the security and trustworthiness of the record—is that it removes the need for a centralized authority to facilitate or mediate the transaction in any way.
Smart contracts are, to many, the most promising application of blockchain infrastructure. The core idea is this: contracting parties agree to a set of terms that are instantiated in computer code (and which, therefore, can include any conditions which are legible to computers). When those conditions are met, transfers can be made automatically. Some of the real power of this model comes from the increasing digital interconnection of ‘smart’ devices–– imagine, for example, that a smart lock on your home could effectuate a short-term rental agreement (and then automatically lock the tenant out at the end of the lease term), or that payment for goods would transfer immediately upon arrival of those goods at a particular warehouse within a given timeframe, detectable by location-tracking technologies. In a sense, smart contracts aim to collapse contract formation and enforcement into a single instrument.
This stands in contrast to the way traditional contracts are enforced. Typically, contract terms are written, acceded to, and—if need be—enforced, in distinct phases: if a contract is breached, the wronged party must take action to recognize and document a harm, establish the other party’s responsibility for the harm, (perhaps) initiate a legal proceeding through which to establish the breach, and ensure that any damages are paid. Importantly, the enforcement phase formally depends on centralized institutions—courts—to intermediate disputes, and on a plaintiff taking affirmative steps that require access to multiple kinds of resources, including knowledge, money, and time. Traditional contract enforcement is messy and resource-intensive—and it is this perceived inefficiency that motivates much of the excitement about smart contracts.
To be sure, some of this enthusiasm is well placed. There are good reasons to avoid reliance on centralized authorities for governance: an authority may be corrupt; it may impose undue delay or bureaucratic inefficiency in resolving a conflict; it may fail altogether. And goals of social justice, equality, and fair outcomes may be served by decentralization, as well. A self-executing agreement could, in theory, obviate some institutional barriers that keep poor people from being able to make property claims. Well-resourced institutional players consistently obtain better outcomes in court than smaller players do, often as a result of procedural rules and institutional structures. By distributing enforcement outside court systems, smart contracts might, in some contexts, be seen as removing some of the barriers that keep the disadvantaged from attaining their legal due.
However, scholars have raised salient critiques of the idea that smart contracts might transform the current legal regime around contractual enforcement (that is, of legal-institution-backed ‘dumb’ contracts). James Grimmelmann and Arvind Narayanan suggest that blockchain boosterism might be tempered by fuller consideration of the human side of business dealings. What if you lose your blockchain private key (a secret code which authenticates your ownership of resources on the blockchain) in a fire? What if you buy a defective product, and need to show your ownership to invoke state consumer protection? These sorts of inevitable ‘hiccups’ are poor fits for blockchain-based enforcement, without the attendant state institutions that step in in cases of exception.
In addition, there are a number of questions about how smart contracts will deal technically with issues that are fundamental to contract law, but which are difficult to instantiate in code: issues dealing with temporality (such as mutual mistake or recission), or standards that apply to contracts and contractors across the board, even if not represented on the face of the document (such as the duty of good faith). These issues represent important challenges to the smart contract framework, and there are some extant efforts to address them technically and legally. But I contrast my present inquiry from these concerns, which have largely to do with translation and implementation of existing legal standards to a new technological form. I also bracket from my analysis critiques of the fundamental viability of blockchain infrastructure—a critique raised most notably in the wake of the theft of millions of dollars’ worth of cryptocurrency on the Ethereum network, and the subsequent ‘hard fork’ of the code that resulted, believed by some to indicate the fundamental nonviability of the model.
My aim here is to raise a separate critique of smart contracts—one premised on the latent social functions of contracting. Just as the security of property depends on both technological artefacts and human institutions in the real world, so do contracts depend on both the technology of the contract—its form and provisions—and the social behaviours that surround, support, and sometimes subvert contracting. Smart contract technology depends on a thin conceptualization of what law does, and how it does it— by focusing on the technical form of contract, to the exclusion of the social contexts within which contracts operate, and the complex ways in which people use them.
How Does Law Work?
Law works through all kinds of avenues other than formal adjudication. Contracting is a deeply social practice in which parties engage for all sorts of purposes, and the effects of contract negotiation reverberate outside of the ‘four corners’ of a formal agreement. In legal studies, the legal realist perspective has long focused on how law unfolds in in vivo practice, and emphasized that these practices may diverge significantly from ‘on the books’ agreements. This perspective suggests that parties engage with legal instruments and negotiations in order to achieve myriad socially productive aims—including many that may not be apparent from the face of an agreement. The social context of law as lived is key for understanding what law is ‘doing’ in any situation.
People use legal instruments to negotiate their social and organizational lives in ways that are both complex and counterintuitive. Contractual obligations are enforced through all kinds of social mechanisms other than the legal system proper; concomitantly, contracts serve many functions that are not explicitly legal in nature, or even designed to be formally enforced. This insight is important for understanding the real-life effects of automated contracts.
In what follows, I describe three contracting practices in which people engage to manage their relations. I suggest that each of these practices, while serving important social aims, are uncontemplated by the automated enforcement framework underlying so-called ‘smart’ contracts. The practices I describe (summarized in the table below) are by no means an exhaustive list, but serve to illustrate how contracting practices operate in the wild—in ways that belie smart contracts’ assumptions about how contracts ‘work’.
Writing or acceding to unenforceable terms
Writing or acceding to purposefully vague terms
Wilful nonenforcement of enforceable terms
Set expectations for future behaviour, including behaviour outside the formal purview of the contract
Facilitate stable and flexible long-term relations
Provide a strategic resource for operating ‘in the shadow of the law’
The Behavioural Force of Unenforceability
First, contracting parties may enter into contracts that contain provisions they know, or suspect, are legally unenforceable. Though courts generally give parties significant latitude in determining what they may agree to in a contract, some provisions cannot be enforced for a variety of reasons, largely based on social policy considerations. Parties cannot, for instance, legally enforce a contract based on unlawful subject matter, like an agreement to murder.
More conventionally, contracts often contain clauses that their authors know to be unenforceably overbroad. This arises commonly, but not exclusively, in contracts of adhesion (or ‘form contracts’), in which one party has more or less exclusive ability to specify the terms without negotiation. For instance, employment contracts frequently contain unenforceable post-employment non-compete clauses.
Why might sophisticated parties include clauses in their contracts that they know to be invalid? The most reasonable interpretation as to why is that the party who chooses to include the clause assumes that the other party believes it to be valid, or at least will act as though it is. If I don’t know that a non-compete clause is unenforceable, I am likely to follow its edicts anyway. In this way, unenforceable clauses are performative: they ‘work’ functionally, even though they don’t ‘work’ legally. In a sense, these clauses work because of chilling effects—because of uncertainty and imperfect knowledge, they deter conduct beyond that which is actually legally proscribed.
Some unenforceable clauses are included even when the parties are on a fairly even playing field in terms of resources and knowledge. In these cases, the inclusion of an unenforceable contract term can serve both parties’ interests, by operating to influence and communicate norms for each of their future behaviour. Consider, for example, the inclusion of ‘infidelity clauses’ in prenuptial and postnuptial agreements, which contractually bind spouses to a (usually financial) consequence if they cheat; though most courts refuse to enforce these clauses, they still ‘work’ by communicating expectations between the parties, and may act to precommit behaviour in a socially desirable way.
To sum up: parties knowingly include unenforceable terms in contracts in order to set expectations for one another’s future behaviour—including behaviour legally beyond the reach of the formal contract. Though this effect of contracting may serve to intensify power asymmetries in some cases, it might also be used to prosocial effect, to mutually precommit behaviour, or to express allegiance to certain norms.
Vagueness and Long-term Relationships
A second set of contracting practices involves the inclusion of terms that are purposefully vague. Terms might be left uncertain for a variety of reasons. It might be unduly cumbersome to prespecify terms if the context of the agreement is likely to shift; parties may have a variety of motivations for conducting some aspects of their business ‘unofficially’; the pecuniary or social costs of negotiating terms might be great enough that parties would rather take a chance, based on their perceptions of costs and probability of breach, than expend resources delineating terms up front. As a legal matter, contracts are not enforceable if their material terms are left vague or open-ended; an ‘agreement to agree,’ though undertaken commonly in practice, lacks legal weight.
The seminal sociolegal scholarship on this phenomenon is Stewart Macaulay’s 1963 study about what he termed ‘non-contractual relations’ in business, summarised in my paper. He found that, to a surprising degree, parties left many aspects of their contractual relationships vaguely planned, incompletely defined, or imperfectly resolved. This was not for lack of sophistication; in fact, it seemed to be a deliberately undertaken and widespread practice in spite of the parties ‘knowing better,’ even when significant financial risks were on the line.
Why might parties act in such a way? In part, Macaulay’s businessmen were unworried about the eventual enforceability of a contract in court because alternative mechanisms, like reputation and internal sanctions, served in its place. Most significantly, parties may choose not to insist on a perfect ‘meeting of the minds’ on every term because they anticipate (and hope) that their relationship will extend beyond the contract at immediate issue. Being exacting about every aspect of an agreement might endanger the long-term relationship of the parties. Vagueness, then, preserves operational flexibility for parties to deal with newly arising circumstances, and sets the stage for social stability in an ongoing relationship.
This way of understanding contracting behaviour has been described as a relational theory of contract. Relationalism pays close attention to the interactions among norms, practices, and doctrines in transactions, as opposed to a more formalistic, classical theory of exchange that concentrates principally on legal decision rules. In this view, contracting is often more like marriage than like bare one-off exchange; the stability of a relation is more important than preserving one’s legal rights in each and every situation.
Strategic Non-enforcement and the Shadow of the Law
A third set of contracting behaviours entails parties’ decisions not to pursue enforcement of enforceable agreements. Legal obligations of all types often go unenforced. Sometimes, parties’ lack of resources or awareness may limit them from pursuing others’ legal obligations in court. But, in other cases, failure to pursue a legal remedy serves social purposes for the non-enforcer, and may be part of a considered behavioural strategy. A good deal of socio-legal scholarship considers how individuals use law—and non-enforcement specifically— as a strategic resource for the management of institutional and interpersonal relationships.
This dynamic is often called ‘bargaining in the shadow of the law’. The basic premise is that formal enforcement—‘taking someone to court’—is only one way in which people choose to enforce their legal obligations to one another, but is often not the best way. In part, this is due to the high costs of formal adjudication, in terms of both time and money. Even when not formally used, the spectre of the court has important effects, because it provides a backdrop and framework against which parties make bargains among themselves. These self-constructed bargains are often mutually advantageous: they avoid the ‘all-or-nothing’ consequences likely to attend litigation in terms of more granular, cooperative negotiations. A good deal of socio-legal scholarship (discussed in more depth in my paper) examines conditions in which such bargaining is likely to be effective; for instance, it tends to be a greater resource when strong community norms or networks provide viable alternative forms of sanction (eg, reputational harms). Enforceable contractual obligations, rather than operating as steadfast if-then consequences of noncompliance, create a backdrop of social expectations against which extra-legal negotiations can occur. Non-enforcement, then, can be as much a ‘use’ of contract as is enforcement.
Smart Contracts Aren’t So Smart
Contracts are deeply social tools as well as legal ones. We have demonstrated how contracts operate as social resources, through and against which people manage their relationships, and within which legal obligations and social expectations are intricately interwoven and mutually constitutive.
What are the implications for blockchain-based contracts? I assert that ‘smart’ contracts fail to take the social complexities of contracting into account; indeed, they are not designed to. We might think of smart contracts as book-smart, not street-smart: while they may facilitate technically perfect and seamless implementation of agreements, they fail to understand or integrate the social world. The social friction required to negotiate and enforce a ‘dumb’ contract can be highly functional—by preserving flexibility, facilitating stability, signaling norms, and the like. Blockchain-based contracts rely on careful prespecification of terms and automated enforcement of obligations. These contracts thus impose a degree of inflexibility on contractors’ relations that might short-circuit a number of alternative uses to which law is put. Smart contract boosterism demonstrates the appeal of bright-line enforcement rules; rules of this sort often garner strong support from the tech community due to their computability and perceived efficiency. But when these rules hit the ground of real practice, custom, and human behaviour, their effects may be nonobvious or counterintuitive.
Notably, the social effects of contract—which do not require recourse to courts for enforcement—represent a form of self-execution, as well, though a different sort than that contemplated by smart contract proponents. We might think of contracts as sometimes having the capacity to socially self-execute—they ‘work’ via their effects on parties’ behaviour—rather than through technical self-execution, in which enforcement relies on code. Social self-execution relies on the communicative functions contracting serves between parties, which may be wholly distinct from their formal legal effect.
Of course, this isn’t to say that there is no role for automated contracts: indeed, they might introduce some desirable social consequences with respect to equitable access to justice. A good deal of non-enforcement of contractual obligations results from the high transaction costs of taking someone to court. To the extent that smart contracts would reduce these transaction costs, they might beneficially remove an impediment to underresourced parties’ contractual rights. At the same time, automated enforcement of contractual obligations—particularly when such obligations are instantiated via interconnected objects— seems likely to have detrimental effects on underresourced people, who often depend on transaction costs to grant them a ‘buffer’ within which to operate. Consider, for example, several recent cases in which subprime lenders have used starter interruptor devices to remotely, and immediately, disable vehicles whenever a borrower falls behind on payments, to the great detriment of borrowers.
As a policy matter, we ought to think carefully about the features of the social setting in which smart contracts are permitted to operate, the degree to which they might preclude other means of enforcing obligations, and the social and normative implications of this foreclosure. To be sure, the subject of the agreement is one consideration here: a stock trade, for instance, is likely to be a better candidate for smart contracting than performance of a personal service, due in part to the computational verifiability of events surrounding the exchange. To realise their promise while mitigating against their potential disadvantages, ensuing discussion, development, and regulation of smart contracts ought to attend explicitly to the relational contexts in which such contracts are deployed.
The foregoing discussion suggests that alternative enforcement mechanisms might work best when certain social resources are present in the contracting relationship that can be drawn upon by the parties. For instance, are the parties likely to want to engage in continued relations beyond the immediate agreement, suggesting that reputation and stability might be desirable alternatives to automated enforcement and strict prespecification of terms? Are the parties motivated to preserve flexibility in their contract, to account for a changing environment, or to reduce the transaction costs of meeting an initial agreement? Are they chiefly trying to communicate particular precommitments or behavioural norms to one another by way of contracting? Do they trust one another? Are they part of the same community, and if so, how durable are relationships in that community? Do alternative forms of signaling (ratings, custom, reputational signals, gossip) exist to incentivize behaviour? Questions like these will help us to understand the social contexts in which contracts are likely to ‘work’ in a certain way, and how technological interventions might operate if deployed there.
Karen Levy is an assistant professor of Information Science at Cornell University and associated faculty at Cornell Law School.
Karen’s full paper, Book-Smart, Not Street-Smart: Blockchain-Based Contracts and the Social Workings of Law was published in Engaging Science, Technology and Society.