David Blanchett: Regret Optimized Portfolios, and Optimal Retirement Income (EP.254)

Episode Summary

Executive Summary: Benjamin Felix and Cameron Passmore interview David Blanchett on behavioral portfolio design, retirement income, advisor incentives, and global equity exposure. The core thread is that real investors are not utility-maximizing robots: regret/FOMO, flexible retirement spending, and product/channel frictions materially affect optimal decisions, so planning tools should model human behavior more realistically.

Main Topics: Regret aversion and 'FOMO' in portfolio construction (Priority: 5/5): Blanchett explains a regret term in portfolio optimization that captures the emotional pain of missing out on assets that perform well. The goal is not to prescribe speculation, but to rationalize why investors may want a small allocation to volatile assets they care about. Dynamic retirement income and spending flexibility (Priority: 5/5): He argues that treating retirement spending as a fixed inflation-adjusted liability is unrealistic. Retiree spending is partly elastic, so optimal withdrawal strategies should adapt over time using funded ratios and evolving household needs. Utility-based retirement planning vs binary success rates (Priority: 5/5): Blanchett criticizes one-number Monte Carlo 'success rates' and instead advocates utility-based, graded outcomes that reflect how badly a shortfall matters. This usually supports higher spending and sometimes more aggressive portfolios than traditional LDI-style plans. Advisor compensation, channel effects, and product choice (Priority: 4/5): The conversation covers evidence that advisor pay models and channels can influence fund selection and client outcomes. Blanchett stresses transparency and due diligence, while noting commissions may partly signal advisor quality rather than only create incentives. Default design and participant behavior in DC plans (Priority: 4/5): The hosts and Blanchett discuss how default savings rates, target-date defaults, and portfolio transparency affect participant engagement, risk-taking, and the likelihood of abandoning diversified portfolios during volatility. Foreign revenue vs domicile in equity exposure (Priority: 4/5): Blanchett argues that company revenue geography may matter more than legal domicile for understanding global risk and home-country bias. Large domestic-index constituents can be economically global, so domicile alone can overstate local exposure.

Key Arguments: Risk aversion and regret aversion are different: risk concerns losing money, while regret concerns the emotional pain of missing out on gains from assets you did not own. Regret is highly investor-specific; many investors, and even institutions, may rationally hold a small amount of speculative assets to avoid behaviorally costly FOMO. A regret-aware optimization framework can explain why small allocations to volatile assets may improve long-run investor behavior, even if they reduce mean-variance efficiency. Retirement liabilities should not be modeled as a single inflation-linked number because real spending has essential and discretionary components with very different flexibility. Because many retirees have guaranteed income already, the portfolio often funds flexible spending rather than all spending, which changes withdrawal and asset-allocation decisions. Binary success/fail metrics are too crude for retirement planning; utility-based models should reflect degrees of shortfall and how painful those shortfalls are. Dynamic spending rules should be implementable in advisor software and based on funded ratios and changing cash flows, not purely on idealized academic solutions. Higher default savings rates and more transparent portfolio constructions can improve participant behavior and reduce panic trading during market stress. Advisor compensation and credentials likely affect client outcomes, but the effect may reflect advisor selection/quality as well as incentive distortion. Company domicile is an incomplete proxy for economic exposure; revenue location may better capture true global risk and may weaken some arguments for home-country bias.

Data Points: Podcast episode: 254 - Rational Reminder episode featuring returning guest David Blanchett Prior appearance: Episode 137 - Blanchett previously appeared on the podcast a couple of years earlier Published research: Over 100 papers - Blanchett's publication record mentioned in his bio Retirement spending elasticity: 70-ish% - He said about 70% of retiree expenditure is on average elastic/flexible Default savings rates in studied plans: Mostly not over 6% - He noted the data set mostly included plans with default rates at or below 6% Monte Carlo runs: 100,000 - He criticized a tool showing 100,000 simulations with historical averages for not adding real value Planning horizon: 30+ years / 35 years - Used to describe long retirements and the limitations of static models Defaulting age bands: 5-year age bands - Mentioned as an example of target-date fund construction limitations Advisor survey response rate: About 80% - He said in DGMI surveys about 80% of responding advisors consistently use Monte Carlo Sample failure metric example: 90% success rate - Used to illustrate why binary success metrics can be misleading Safety spending default example: 3% vs 6% - He suggested higher default savings rates reduce skepticism and improve acceptance of target-date funds Retirement spending decline: About 1.5% per year vs 3% inflation - He described a 'spending smile' where spending rises slower than inflation for many retirees Public pension example: Claim at 65; spouse retires at 70; deferred income at 85 - Used to illustrate why dynamic cash-flow modeling must handle uneven inflows Employer match example: High match with low default can act like a tax on less sophisticated workers - Because many people simply follow the default contribution rate Retirement age / longevity: 30–35 years of retirement - Used to emphasize the importance of dynamic withdrawal rules

Pivotal Quotes: "People aren't utility maximizing robots across all time periods and all times. We're people, right?" — David Blanchett: Explaining why regret, FOMO, and behavior should be incorporated into portfolio design "The 1990s want their key assumptions back." — David Blanchett: Critiquing retirement-planning tools that still rely on old static assumptions "If you read the paper, I've literally walked through step by step exactly how you can do it in a way that I know for certain can be done very quickly in a Monte Carlo projection." — David Blanchett: Arguing that dynamic spending and utility-based planning can be implemented in practical advisor software

Implications: Advisors should design around human behavior, not idealized models: allow small regret-driven allocations, use dynamic retirement spending rules, communicate risk in intuitive terms, and recognize that revenue geography and compensation models can materially change portfolio and advice quality.