The Relative Asset Pricing Model Incorporating Liabilities and Delegation to CIOs: Version 0.1
37 Pages Posted: 22 Sep 2012
Date Written: June 1, 2012
This paper makes a simple but very bold argument that because MVO and CAPM were derived from a theoretical construct of the investment world rather from investment reality, the resulting theory is a very specialized case of a much more general theory. Therefore, rather than using CAPM to derive solutions for liability-driven investing (LDI) and delegation to CIOs or external managers (principal-agent issues), we suggest that maybe the lens should be inverted and the theory start with a much broader view of a liability to be serviced for an institutional/retail investor with delegated decision-making. Starting from this perspective, all decisions are relative decisions and hence we make the case for a Relative Asset Pricing Model (RAPM) as the true starting point for any asset pricing theory. Turning off many of these features (e.g., no delegation and hence no concern for skill of agents or choosing the portfolio with the highest absolute return for level of risk and ignoring liabilities) gives us the CAPM and hence the claim that CAPM is a very stylized model of a more general theory. This paper presents a very simple model to show how this framework changes the asset pricing paradigm. The key point is that since current theory was derived from Expected Utility Theory, which arbitrarily forces a choice between just two variables (e.g., work vs. leisure, consumption of good 1 vs. good 2 or trade-off between mean and variance), it may have missed the complexities of real-life investment decision making. If a decision-maker had to choose among three (independent) variables and was forced to make a two-dimensional choice, one would get a limited result. This is exactly what has happened with investment theory. A typical investor, whether institutional or retail, sets aside money to service some future consumption (liability) and typically lacks the skill to make the decisions themselves and hires agents who may be lucky or skillful. Therefore, the investor has to worry about the relative return of the portfolio (relative mean) and relative risk (composed of two independent variables – relative variance and correlation) and the investor is not agnostic about how the relative variances and correlation are chosen. The goal of this paper is lay the first stone in this direction to add a correlation dimension, which may help shed light on the important yet unobserved “risk aversion” parameter in current theory, and to hopefully encourage other more talented academics to develop the RAPM framework as it will provide better recommendations for asset pricing, asset allocation, rebalancing and manager compensation.
Keywords: relative asset pricing, liabilities, agency, skill
JEL Classification: G11, G12
Suggested Citation: Suggested Citation