Bucketing and Monte Carlo Solve Different Problems

There is often confusion about the purpose of bucketing because it is frequently discussed alongside Monte Carlo analysis. The two, however, solve fundamentally different problems.

Bucketing is an asset allocation methodology. Like a traditional fixed asset allocation (FAA), its purpose is to determine how a retirement portfolio should be invested. Different retirees and advisors may prefer different asset allocation methods, and there is no universal answer as to which is best. The objective of any asset allocation process is to build a portfolio that gives the retiree a reasonable chance of achieving their goals while maintaining a level of risk they are comfortable living with.

An asset allocation decision, however, is only one factor affecting retirement success. Withdrawal rates, market returns, inflation, retirement timing, longevity, and many other variables also influence outcomes.

Monte Carlo analysis serves a completely different purpose. Rather than deciding how the portfolio should be invested, Monte Carlo evaluates whether a given retirement plan is likely to achieve its objectives. It is largely asset allocation agnostic—it can evaluate a portfolio constructed using a bucketing methodology just as easily as one built using a fixed asset allocation. In fact, Monte Carlo cannot even begin until an asset allocation has already been selected.

Monte Carlo therefore does not replace the asset allocation decision. Instead, it evaluates the consequences of that decision, together with all the other assumptions in the financial plan.

Another important distinction is behavioral. Asset allocation is concerned not only with expected returns but also with ensuring that the retiree is comfortable with the amount of investment risk being taken. Monte Carlo, by contrast, generally assumes the investor will stay invested according to the prescribed strategy throughout retirement. It does not determine whether that strategy is psychologically sustainable for the client.

For this reason, a strong Monte Carlo result does not necessarily imply a suitable asset allocation, just as an asset allocation that perfectly matches a retiree's comfort level may still produce an unfavorable Monte Carlo outcome if the withdrawal rate is too high or the starting conditions are unfavorable. The two tools answer different questions.

A useful analogy is that asset allocation is like a thermostat, while Monte Carlo is like a thermometer. A thermostat determines how the system is set up and operated. A thermometer simply measures the resulting temperature. Both are essential, but neither performs the other's function.