Why Estimating Investor Risk Aversion Requires a Structural Choice Model
Summary
The note considers whether an institution’s risk-aversion parameter can be inferred from observed trade profit and loss under a proposed power utility function. Its response cautions that utility is generally specified over wealth or consumption, rather than trade P&L, and that the proposed functional form needs a defensible economic foundation before its parameter can be estimated.
It points toward structural estimation using observed choices and an explicit model. For households, cited research estimates risk preferences and intertemporal substitution; for institutional portfolios, the response points to demand-system work that begins with a mean-variance framework and relates asset holdings to characteristics. These references suggest approaches rather than a ready-made estimator. The question lacks details about the institution, decision process, and data, so trade P&L alone is not presented as sufficient to identify risk aversion, and holding period is not developed as an estimation method.
Key ideas
- Risk preferences are typically modeled over wealth or consumption rather than isolated trade P&L.
- An estimated utility parameter depends on the functional form and economic model chosen.
- Structural estimation can infer preferences from observed household choices under an explicit framework.
- Institutional portfolio demand can be studied with a mean-variance model and asset characteristics.
- The supplied question lacks enough detail to specify an identification strategy.
Tags
Full text
# How estimate utility function and risk aversion level for an investor from empirical data?
# How estimate utility function and risk aversion level for an investor from empirical data?
Suppose we are interested in estimating utility function for an institutions. We have collected P&L data from different trades, and are interested to estimate some $u(P\&L;\alpha)=(P\&L)^{\alpha}$ utility function's parameter, namely $\alpha$. As we do not observe both $u(P\&L;\alpha)$ (values of utility function) and the parameter $\alpha$, this is kind of problematic problem (seems to me). Do you have any thoughts how to handle with this problem?
P.S. I think one may use holding period to estimate $\alpha$, though not sure.
## Answer by phdstudent (score 1)
https://quant.stackexchange.com/a/80441
Utility parameters usually depends on the utility function you specify which is usually a function of wealth or better consumption and not P&L. Also your utility function is weirdly defined and has no decent microfoundation. I would need some more details.
From a structural approach these are two very good references in estimating risk aversion and elasticity of intertemporal substitution:
- Calvet et al. The Cross-Section of Household Preferences (2022)
- Choukhmane & de Silva What Drives Investors’ Portfolio Choices? Separating Risk Preferences from Frictions (2024)
If you are looking to estimate specifically for institutions, then you should look into the work of Koijen & Yogo, who start from a mean-variance problem and estimate how certain institutions "like" certain characteristics. Here's the reference:
- Koijen & Yogo A Demand System Approach to Asset Pricing (2019)
Either way your question needs more details.Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.