Trust, Transparency, and Complexity

Richard T. Thakor is Assistant Professor of Finance at the University of Minnesota Carlson School of Management, and Robert Merton is Professor of Finance at the MIT Sloan School of Management. This post is based on their recent paper, forthcoming in The Review of Financial Studies. 

In recent years, there has been increasing complexity in both physical and financial products, due to higher demand for customization and financial innovation. Customers—who may have trouble understanding all relevant product attributes—have to trust the producers in order to buy their products, and investors have to trust these producers in order to invest in them. This has elevated the role of trust in enabling the adoption of these products. The importance of trust has led to significant discussion about how it is built, and transparency (disclosure) is viewed as a common tool for building trust (e.g. Offermann and Rosh (2012)). This has first-order policy relevance, as laws and policies are frequently proposed and put into place to mandate disclosure and verification.

Despite this common view that transparency can help build trust in product sellers, thereby helping buyers cope with complexity, the empirical evidence on the issue is mixed—there is little direct causal evidence on the relationship between transparency and trust in the cross-section. Indeed, anecdotal evidence suggests that greater transparency need not be associated with greater trust: some of the most trusted institutions often disclose less information than their less-trusted counterparts.

In our paper, “Trust, Transparency, and Complexity” (forthcoming in The Review of Financial Studies), we examine the interaction between trust, transparency, and verification theoretically in a model of endogenous product complexity and transparency. In contrast to the common view, we show that transparency does not build trust per se—in the sense that becoming more transparent does not directly lead to the firm becoming more trusted—but rather it substitutes for trust in that if a product is fully transparent to its user, there is no need for trust. There are however limits to the degree of complexity a product can have in order for it to be transparent.

In our model, a firm makes a choice of project that it finances with outside equity, where a project is the development and marketing of a product. What consumers and investors want is a good investment or a good product, but the firm may not make such an investment. The measure we use for “better/preferable” is the probability that the project selected by the firm is a good one for either its investors or its customers. Users will only use the product if the probability that it is a good project/product/service is above a threshold. Reputation or trust in a firm is measured quantitively by the probability investors attach to the firm investing in a good project; the more trust the firm has the higher that probability.

Firms can disclose information, to increase transparency, which has two opposing effects. The functional aspect of disclosure is that it increases the information investors receive about the firm’s project choice, which reduces the firm’s cost of capital. Transparency thus does not increase trust but more of it will increase the probability of a good project/product, and thus it is a substitute for trust. The dysfunctional aspect is that it is directly costly, for example due to “disclosure processing costs” or competitive reactions by rival firms triggered by the disclosure of proprietary information. Optimal disclosure balances the functional and dysfunctional aspects.

Firms can differ in terms of the nature of the projects that they may undertake, or the products that they sell, with some being more complex than others—e.g. a firm that produces annual calendars or appointment books versus one that sells highly customized swaps or engages in complex derivative transactions. Complex products are associated with more complex information. Some information can be easily communicated at a low cost by firms, while other information may be very costly for the sender to communicate or costly for the receiver to process. We focus on “essential” complexity and assume that the expected payoff on a good project is increasing in its information complexity, leaving out of the analysis the prospect of unnecessary complexity created for the purpose of obfuscation by these firms to exploit their customers.

Thus, for a given amount of trust, to increase the probability of being a good product to the user, one can increase transparency. To make it more transparent to the user, it may however be necessary to simplify the complexity of the product, and if that reduction in complexity reduces how well the product or service performs, that is a cost. We show that for high product complexity, trust is the only way a product or service can be used. Thus, high product complexity can be pursued only by the most trusted producers, and they choose less transparency (opacity). Our analysis thus provides an economic rationale for the evidence that transparency and trust are sometimes negatively associated.

There are many high-value products, particularly in finance and healthcare, that are “inherently opaque” and so, using disclosure to achieve transparency is either infeasible or too costly relative to alternative mechanisms. Examples of such products abound. The development of high-value medicines or financial products/services have such complexity (i.e. they need very specialized expertise or scientific knowledge to understand how they work) that the details of the product or service cannot be made transparent to the consumer. Another example is fintech payment systems like Alipay and WeChat Pay, which are clearly not transparent to users. However, such products or systems are widely used. Why? The answer is verification: you experiment and verify it works or a friend, co-worker, or family member tries it and it works, and thus you are willing to use it. To accommodate this, we include ex ante third-party verification. Such verification is commonplace; examples include inspections with real estate transactions, ratings issued by credit rating agencies, and public audits of firms’ financial statements. In practice, such third-party verification requires trust in the monitors that perform the verification; we also account for the possibility that such monitors may shirk in their duties, and show that contracts can be written to overcome this potential inefficiency. Verification is also a substitute for trust. For a given amount of trust, to increase the probability of being a good product to the user, one can create verifiability of the provider. To do so, it may be necessary to simplify the complexity of the product, and if that reduction in essential complexity reduces the how well the product or service performs, that is a cost.

Trust is the only way a product or service can be used, if it is not transparent enough to the user and it cannot be verified by the user. We find that verifiability substitutes for disclosure-induced transparency for low levels of project complexity, but is dominated by disclosure for higher levels of complexity. The analysis thus generates a trust-based hierarchy of firms with transparency. At the lowest level of trust, firms choose the least complexity but remain opaque and use ex ante third-party verification. For intermediate levels of trust, firms choose greater complexity with transparency, but transparency declines as complexity increases. For high levels of trust, firms choose maximum complexity but return to being opaque, disclosing nothing. When the product/service is quite simple, having information verified by a third party may involve such a low cost that it is even cheaper than the cost of disclosure. So when complexity is very low, the firm remains opaque and relies on verification, leaving disclosure-based transparency to operate only for intermediate levels of complexity.

In summary, by treating complexity, trust and transparency as endogenous, our paper can make predictions about the relationship among trust, transparency, verifiability and complexity. Our paper also demonstrates how trust is an asset of the provide (e.g. a firm, institution, or government).

Both comments and trackbacks are currently closed.