Financial Disclosure with Limited Processing Capacity

Abstract

The received wisdom about financial disclosure is that increasing public information in security markets raises their liquidity. We show that, while true if disclosure concerns information whose interpretation does not require sophisticated processing, this principle does not apply to information about the structure of security payoffs, which instead heightens the advantage of informed investors vis-a-vis market makers and therefore reduces liquidity. The reason is that this type of disclosure helps informed investors to deploy their “attention” more profitably, by focusing it on the asset’s sensitivity to the most price-relevant risk factors and thereby increasing their informational rents. Thus, the traditional insight that transparency enhances liquidity only applies when applied to information that is easily processed by all market participants, including market makers – but is reversed when applied to information about asset structure. Despite its simplicity, the model is rich enough for assets to differ along two distinct dimensions: (i) “complexity”, which is captured by the number of states to which the asset payoff is sensitive, and (ii) its “opacity”, as measured by the marginal cost of learning the asset’s sensitivity to each factor. Both of these dimensions affect the potential damage arising from disclosure about the asset payoff structure. The larger the number of risk factors to which the asset’s value responds, the thinner market participants have to spread their information processing capacity, and the greater the informational gap between informed investors and market makers. Up to a point, an increase in an asset’s opacity has the same effect, since it increases the attention that market participants must allocate to understand the asset’s sensitivity to each value-relevant factor. However, the effect is non-monotonic: beyond a critical point even the information-processing capacity of the most sophisticated investors will fall short of the required minimum, so that for very opaque assets all market participants become equally uninformed, and liquidity is restored.In our setting, disclosure regulation is warranted only when the objective of the asset issuer is not aligned to those of the regulator. For instance, the regulator may care more about price discovery and less about trading costs than the issuer, because price discovery has beneficial effects on the allocation of investment or on the liquidity of other assets that are not internalized by the issuer. In this case, regulation may mandate also some disclosure about asset structure, despite its negative effects on market liquidity, so as to speed up price discovery. But since the issuer may respond by increasing the asset’s opacity, disclosure regulation may be ineffective unless supplemented by a constraint on the opacity of assets.The model also has the interesting prediction that informed investors will tend to specialize, in the sense of concentrating their information processing ability on different risk factors and on different assets. Finally, we conjecture that if investors and market makers are allowed to choose the level of their financial sophistication, their investment in information processing capacity will be strategic complements, and both will tend to increase in the complexity and opacity of the assets being traded.