Goal-based Wealth Management by Reinforcement Learning

Investors always face the dilemma of risk and return: risky investments offer better returns on average, but also dismal returns once in a while. Goal-based Wealth management is a rather natural way of communicating the “risk vs return” to an investor. Essentially, he or she is asked to define a financial goal over a time horizon (for example, having 1M$ in 10 years), and the probability of attaining this goal is optimized on the basis of historical data, yielding for example an 83% chance of success. If the investor is unhappy with this probability, the goal can be downsized. In all cases, the program optimizes the asset allocation to provide the best chance of success.

Faculty Supervisor:

Michel Denault;Jean-Guy Simonato

Student:

Partner:

Inovestor

Discipline:

Business

Sector:

Information and cultural industries

University:

HEC Montréal

Program: