Beyond the Risk Questionnaire
First-generation robo-advisors were essentially digital risk questionnaires paired with automated ETF allocation. Betterment and Wealthfront pioneered the model: answer 6–10 questions, receive a diversified portfolio, and let the algorithms handle rebalancing. It was simple, low-cost, and effective—but limited.
The second generation of robo-advisory platforms is far more sophisticated. They now integrate tax-loss harvesting, direct indexing (owning individual stocks instead of ETFs for more granular tax management), and retirement-specific features like Required Minimum Distribution (RMD) calculations and Social Security optimization.
The Technology of Tax-Loss Harvesting
Tax-loss harvesting (TLH) is the systematic selling of securities at a loss to offset capital gains tax liabilities. What makes it challenging at scale is the need to maintain portfolio exposure while realizing losses:
- Wash sale detection: The system must avoid buying “substantially identical” securities within 30 days
- Replacement security selection: Sold positions must be replaced with correlated but not identical alternatives
- Tax lot optimization: The system must track the cost basis of every tax lot and select the optimal lots to sell
- Client-specific constraints: Different clients have different tax situations, holding periods, and risk tolerances
The Retirement Income Problem
The hardest problem in robo-advisory is not accumulation—it is decumulation. How does a platform manage systematic withdrawals from a portfolio while minimizing sequence-of-returns risk? This requires dynamic asset allocation models that adjust based on market conditions, life expectancy estimates, and spending patterns.
Platforms like Income Lab and LifeYield are pioneering “retirement income optimization” as a distinct category within robo-advisory, using Monte Carlo simulations and dynamic programming to solve the decumulation puzzle.
