Retire in Monte Carlo? Simulating retirement outcomes

• Goals-based planning in our simulations may have helped advisers better analyze a target-date fund’s potential to meet a client’s retirement goals.
• According to our model, adding an equity multifactor allocation, or transitioning to a low-volatility allocation near retirement, improved performance without altering the fund’s glidepath.
• We integrated ESG views into the equity allocation without sacrificing retirement prospects in our example.

An image of dozy days along the French Riviera undoubtedly appeal to many future retirees. Yet, despite a decadelong bull market in global equity markets, the retirement picture for many participants in U.S. defined-contribution (DC) plans isn’t so sunny.

One well-chronicled reason for this is America’s low household retirement-savings balances.¹ Another is the possibility that the principal DC savings vehicles — balanced and target-date funds — have not generated the returns needed to fund retirements that can span multiple decades. We found that integrating an equity multi-factor or environmental, social and governance (ESG) allocation improved a hypothetical plan participant’s funded status, based on Monte Carlo simulations.

A goals-based planning approach

We started with a hypothetical couple: Both 40 years old and planning to retire at 65. They have $500,000 in retirement assets and each contribute the maximum pre-tax amount to their retirement account annually. Their goal is to have $100,000 in annual income throughout retirement.

Using MSCI’s WealthBench’s goals-based planning approach, we made projections for a range of possible retirement outcomes using Monte Carlo simulations.² We used the asset mix, return assumptions and glidepath from a major provider of target-date funds as a base case. Importantly, we added our own conservative return estimates (relative to their simulated history) for the factor and ESG allocations.³ Lastly, we used asset-class correlations and volatilities from the extra-long-horizon variant of the MSCI Multi-Asset Class (MAC) Factor Model.

Testing four scenarios

The four scenarios we tested are shown in the table below, starting from the Base scenario described above. The Multifactor, Multifactor with low-volatility transition and ESG integration scenarios involve replacing the Base scenario’s equity allocation only. We did not change the fixed-income allocations or the glidepath itself.

There were advantages and disadvantages to each. Multifactor indexes have historically outpaced market-cap indexes over long horizons, yet they have come with higher volatilities and correlations to other asset classes. Low-volatility indexes have historically had lower drawdown risk and a higher correlation to a retiree’s fixed-income-like liabilities, but underperformed during equity-market rallies.

Finally, some exclusion-based ESG indexes have taken large sector and country bets, resulting in a high tracking error relative to the broad market. ESG integration indexes, on the other hand, are designed to more finely control for risk and return similar to those of the broad market.

Living in a simulation

The first exhibit below shows the simulated return and risk for each scenario over the course of the simulation period. To bookend the scenarios, we also included allocations to 100% domestic U.S. equity and 100% domestic fixed income.

Risk and return over simulation period

Simulated returns are in USD and extend to a joint life expectancy of 85. U.S. fixed income is proxied using the Bloomberg Barclays US Aggregate Index. U.S. equity is proxied using the MSCI USA Index. Expected returns, volatilities and correlations are fixed throughout the simulation period. Volatilities and correlations for each asset class are taken from the MSCI Multi-Asset Class (MAC) Extra-long Horizon Factor Model as of June 30, 2019.

While it may appear, at first blush, as if U.S. equities’ or our Multifactor scenario outperformed, risk and return tell only part of the story. How close did each scenario come to meeting the couple’s funding needs in retirement? The next exhibit shows the retirement’s goal-funding status.

Funded status of retirement goal

Funded status refers to proportion of the goal that is met by current and future assets. Annual contributions and retirement income grow with inflation. We assume there are no other sources of income, such as social security or inheritance, during retirement.

Two observations are noteworthy. First, both the Multifactor and Multifactor with low volatility transition scenarios increased the funded status over the Base by almost 25%. Second, the ESG integration scenario, though not indicative of future results, maintained a similar success rate as the Base.

The results for the Multifactor with low volatility transition scenarios might come as a surprise. We assigned it lower returns than the Multifactor scenario, yet it delivered similar outcomes. To examine this further, we can again use Monte Carlo simulations to trace each scenario in a critical period — the five years leading up to retirement.

Watching the tails

Asset allocators are familiar with the balancing act between equity exposure and drawdown risk. The higher equity allocations during a participant’s accumulation phase could also lead to large drawdowns just as retirement nears.

Transitioning at that time from a market-cap-weighted (or multifactor) equity allocation to a low-volatility equity allocation is one method of maintaining equity exposure while working to mitigate the risk of a drawdown.

The exhibit below shows the worst-case outcome and maximum drawdown for each scenario in the five-year period preceding retirement, from our Monte Carlo modeling. The Base, Multifactor and ESG integration scenarios all had similar tail characteristics. The Multifactor with low-volatility transition scenario, in contrast, avoided ruinous losses. This helps explain why its success rate was on par with the higher-return Multifactor scenario, as shown in the previous exhibit.

Multifactor with low-volatility transition had lowest tail risk of all equity scenarios

Bottom 5% in cumulative return values refer to the average of the worst simulations for each scenario for the five-year period. Maximum drawdown refers to the average of the maximum drawdown of each simulation for each scenario.

Here comes the sun?

Although nascent, target-date funds have gained wide adoption over a short period. Potential opportunities may still exist in portfolio construction. In our analysis, incorporating recent equity-index innovations improved outcomes, helped avoid losses and implemented ESG views without losing equity exposure during the study period.

The author thanks Dan Schneider for his contribution to this post.

¹“How America Saves 2019.” Vanguard, June 11, 2019.

²MSCI’s WealthBench’s goals-based approach determines the funded status for a participant. The funded status is the ratio of the value of available assets to the value of assets required to meet a goal with a desired level of confidence. We use a 75% confidence level for the simulations.

³ The hypothetical portfolios replicate indexes. We project 150, 100 and 30 bps of annual active return for the MSCI ACWI Diversified Multifactor, MSCI ACWI Minimum Volatility and MSCI ACWI ESG Focus indexes, respectively. These are approximately half of the simulated, historical active returns for each index relative to the MSCI ACWI index.

Further Reading

Same target, different destinations

Goals-based Asset Allocation in WealthBench