Advising Sophisticated Customers: Evidence from Health Insurance Brokers

57 Pages Posted: 7 Nov 2023

See all articles by (Joy) Tianjiao Tong

(Joy) Tianjiao Tong

Ivey Business School, Western University

Date Written: October 9, 2022


This paper investigates how brokers advise sophisticated customers by looking into the employer-sponsored health insurance market. The commonly held belief is that kickbacks are detrimental to customers, but this may not be true for sophisticated customers. Using detailed health plan level data, I first observe that employers frequently use brokers in purchasing health insurance plans. I then explore an event in 2004 in which three large insurance brokers were banned from receiving contingent commissions. The results show that banning such kickbacks can have negative consequences for broker-intermediated plans, leading to higher premiums and worse plan qualities. This is likely due to brokers reducing their effort to acquire information and compensating for their loss via premium-based compensation. Additionally, this paper finds that unionized firms exhibit lower price increases, while well-governed firms can mitigate premium increases by discontinuing the brokerage service. Furthermore, the adverse effects on premiums are more pronounced in public firms than in private firms, which are generally less sophisticated and have been receiving lower-quality advice. In more competitive brokerage markets, the distortion of premiums is less severe. Overall, regulations on brokers' compensation can have unintended consequences on sophisticated customers.

Keywords: Brokerage, advice, consumer sophistication, healthcare, insurance

JEL Classification: G22, G24, G30, G50, I11, I13, D18

Suggested Citation

Tong, (Joy) Tianjiao, Advising Sophisticated Customers: Evidence from Health Insurance Brokers (October 9, 2022). Available at SSRN: or

(Joy) Tianjiao Tong (Contact Author)

Ivey Business School, Western University ( email )

1151 Richmond Street
London, Ontario N6A 5B8

Do you have negative results from your research you’d like to share?

Paper statistics

Abstract Views
PlumX Metrics