Stationarity and stability of underwriting profits in property‐liability insurance

Purpose - To examine the existence of underwriting cycles for the property‐liability insurance industry as a whole, and by line of business. Specifically to consider whether the combined ratio is stationary and stable. Design/methodology/approach - The augmented Dickey‐Fuller (ADF) test is employed for unit roots, while dummy variable methods, the Chow test, and switching regression are used for stability. Findings - Underwriting profits of most lines of business and all lines combined are not stationary and have structural changes. For the whole property‐liability industry, a structural change occurred in 1981. Before the change, underwriting cycles existed since combined ratios followed an AR(2) process. After the change, combined ratios are non‐stationary. Practical implications - Without clear underwriting cycles, there is more difficulty for the insurance industry in pricing and reserving, for regulators in monitoring the financial strength of insurers, and for customers in terms of the affordability and availability of insurance. Originality/value - The paper recognizes the non‐stationarity of combined‐ratio series, years of structural changes in the insurance industry and specific lines of business, and the possibility that underwriting profit is cointegrated with investment income.