Know-your-client (KYC) questionnaires should lead to an accurate determination of client risk tolerance. At least, that’s what a group of Canadian academics expected to confirm when they compared anonymized transaction data and KYC information from a securities dealer with more than 50,000 accounts and 23,000 clients.
Instead, their two research papers point to flaws in the KYC process and to new ways of assessing risk.
In the first paper, the authors adopted a recency, frequency and monetary value (RFM) model commonly used by marketers to segment customers. They examined the number of days since a client’s last trade (recency), total trades, average number of days between trades (frequency), and buy and sell totals (monetary value).
After grouping clients into five clusters based on their trading behaviour, the authors compared the data against KYC profiles. They found the vast majority of clients in each cluster were classified as medium risk despite demonstrating very different trading behaviours. The authors concluded that transaction frequency and volume can tell advisors more about risk tolerance than standard questionnaires.
In a follow-up study released in May using data from the same firm, the authors compared KYC profile risk allocations to investment portfolios using a value-at-risk (VaR) discrepancy methodology. The authors used VaR to measure clients’ elicited (or stated) and revealed risk individually. Elicited risk came from the KYC questionnaire and client profile; revealed risk was demonstrated in trading behaviour, and took utility of wealth (or the idea that a client’s total wealth affects how they perceive a loss) into account.
By quantifying clients’ elicited and revealed risk, the authors used the difference between the two to measure how appropriate their client risk profiles were.
“We examine how clients’ trading behaviour and cash flow impact their portfolios and use VaR to quantify how much of each client’s portfolio is genuinely ‘at risk’ and how far clients are from their stated risk preferences,” the paper stated.
The authors found that advisors generally put clients in portfolios less risky than their stated profiles. KYC is often used as a “guardrail” against which advisors build clients’ portfolios, and advisors are generally conservative. That’s good for clients trying to preserve capital, they wrote; less so for those wanting to maximize growth but who aren’t taking sufficient risk to meet their goals.
“Under-risked accounts are just as problematic as over-risked, but easier for advisors to defend to regulators and lawyers given there is no significant loss of funds,” the authors wrote. “An example might be older clients who can preserve their capital but are unable to achieve investment incomes that allow them to maintain their lifestyles.”
VaR is a useful communication tool to help clients understand risk in dollar amounts or portfolio percentage and to conceptualize how different asset mixes match with their stated risk preferences, the authors said.
“Using the same measure for both profile risk and revealed risk — particularly the difference between the two — in real time, we believe VaR could become an important tool in helping clients, advisors, dealers, and regulators monitor a careful balance between all the variables and stakeholders.”
“Know Your Clients’ Behaviours: A Cluster Analysis of Financial Transactions” by John R. J. Thompson, Longlong Feng and R. Mark Reesor from Wilfrid Laurier University and Chuck Grace from the Ivey Business School, published in the Journal of Risk and Financial Management, 14(2), 2021
“Measuring Financial Advice: Aligning Client Elicited and Revealed Risk” (same authors and Adam Metzler from Wilfrid Laurier University), published on the SSRN network, May 21, 2021