Three concepts of performance-based regulation
Concept 1: Using existing data — what have customers chosen in the past?
Description o f the concept
Customers continuously choose among financial service provider product offers. These choices will depend both on their specific needs and wants (i.e., their preferences) and on the way they are served by employees of the institution.
Concept 1 relies on a systematic analysis of the choices a specific institution s customers make. In particular, the aim is to measure how often consumers choose a product from this provider that is dominated either in terms of the cost to the consumer without offering additional benefits, or in terms of the benefit — without any additional costs — by another option the financial provider offers. By restricting the offers made by an individual provider, it cannot be argued that sale channels, other services of this provider, or the location of branches make a difference for customers. The only difference is that some consumers may not be aware of or find a better option. It can be expected that good performance of financial services would ensure that consumers are able to choose the option that is most financially advantageous to them.
More formally, consumers may purchase a financial product — which may be a savings product or a credit card - and have several options on offer that differ in a limited number of characteristics. They might be fees, interests owed or earned, and limitations on access to funds. In most cases these characteristics can be clearly ordered: higher interest earned on savings is better than low interest; higher interest owed on debt is worse than low interest owed on debt; low fees are better than higher fees. A rational or reasonable person will always prefer the better option. In terms of economic theory, if they choose the opposite they violate the assumption of transitivity as they prefer less over more. Concept 1 counts the cases of customers choosing dominated options among the portfolio offer by a financial provider in a specific class. Looking at all consumers making a choice, the ratio of consumers making suboptimal to optimal choices can be used as a confusion index.
Natural experiments: how would “confusion audits” work?
Financial service providers would be asked to report on consumer decisions among their product classes as well as to report on what options the financial service provider had available at the moment in time the consumer made the decision. This data would be available to financial service providers given the concept does not rely on the set of options from which consumers chose. Rather, it relies on the set of all options that would have been available from the provider at the time. Financial reporting data is likely to be available to satisfy regulatory requirements and to provide appropriate support to customers.These decisions can then be aggregated and the confusion ratio can be calculated and compared across institutions.
The control and treatment group in this case is stylized. Thus it can be said that a neutral environment - for example a consumer picking his or her product via a well-organized market survey such as that provided by third party ratings - allows for optimal choices. Against this benchmark the identified ratio will show the level of deviation due to the financial providers particular strategy used for the majority of consumers and employing the services of the institution. Alternatively, recent developments using methods of artificial intelligence and empirical approaches to so called Big Data could be employed to identify dominance relationships in existing administrative data of financial service providers.
What would a “confusion auditor” look like?
The auditor could be internal or external to a financial market regulator. The index or ratio depends purely on truthfully reported data.The auditor needs to aggregate this data (most likely treat it as commercially in confidence) and calculate the ratios, and only the aggregated data would need to be made public.
To be able to calculate the ratios, the auditor needs to classify products and identify dominance relationships. This should be doable for those with a high level of financial literacy. A university-qualified auditor with some experience in empirical research will be able to conduct such an analysis. Given that the methods and data are well defined, the audit could be easily verified by regulated entities as well as by the interested public. As for academic research, anonymous data and the code used in the analysis can be published to ensure credibility and transparency.
