Loss Harvesting and the Myth of Tax Alpha (EP.158)

Episode Summary

Executive Summary: Episode 158 covers summer updates and book/news recommendations before a deep dive into tax loss harvesting. Ben argues it is often oversold: its benefits depend heavily on return levels, future tax rates, contributions, realized gains to offset, and trading costs. He concludes that in Canada, average-cost accounting, no step-up at death, and no short/long-term rate differential make the strategy far less compelling than marketing suggests.

Main Topics: Summer update, store promotion, and listener/community notes: Hosts open with light banter about the Dog Days of Summer, promote Rational Reminder merch with a discount code, and respond to listener reviews and feedback about goals and podcast style. Book review: Playing to Win: Ben reviews Roger Martin and A.G. Lafley’s strategy framework, emphasizing strategy as a set of choices: winning aspiration, where to play, how to win, capabilities, and systems. He connects this to business decision-making at PWL. ETF industry and Robinhood market structure news: They discuss Canadian ETF growth, asset managers’ ETF exposure, and Robinhood’s rapid growth, revenue mix, user engagement, and concerns about gamified investing and execution quality. Core topic: tax loss harvesting mechanics: Ben explains how tax loss harvesting works, why it is tax deferral rather than tax elimination, and the role of superficial loss/wash sale rules and substitute securities. When tax loss harvesting works—and when it doesn’t: A model and empirical papers show tax benefits depend on expected returns, realized gains, future tax rates, contributions, transaction costs, and volatility. The strategy is much less attractive under Canadian assumptions. Canada-specific tax and portfolio considerations: Ben highlights Canadian differences: average cost accounting, no step-up in basis at death, no short/long-term capital gains distinction, and fund-level tax dynamics that can make pooled funds more efficient than direct ownership. Bad advice of the week: UK pension fee caps and illiquid assets: A listener submission critiques proposed UK pension reforms that would allow performance fees to fit under fee caps, potentially encouraging illiquid investments and undermining expected returns.

Key Arguments: Tax loss harvesting is fundamentally a tax deferral strategy, not a true tax reduction strategy. The value of deferral rises with higher expected returns and lower future tax rates; if future taxes rise, tax loss harvesting can become unattractive or even negative. Many marketing claims of roughly 1% tax alpha are too aggressive unless assumptions are unusually favorable. The strategy requires realized gains to offset realized losses; if an investor is mostly buy-and-hold and not realizing gains, its usefulness drops sharply. Transaction costs and tracking error materially reduce the net benefit of tax loss harvesting. Canadian rules make the strategy less effective than in the U.S. because Canada uses average cost, lacks step-up in basis at death, and does not differentiate short- and long-term capital gains rates. Pooled funds can have tax advantages over direct ownership because inflows and redemption rules can reduce tax drag across unitholders. Illiquid assets are not automatically attractive just because they generate performance fees or can be smoothed into pension fee caps.

Data Points: Podcast episode: 158 - Episode number for the Rational Reminder discussion Summer discount code: 20% off with code 'rational summer' - Merch promotion running July 15 to July 29 Dog Days of Summer: July 3 to August 11 - Ben explains the seasonal phrase's astronomical origin Canada ETF listings: 1,118 listed ETFs - Canadian ETF market size mentioned in news discussion Canada ETF sales YTD: $29 billion - Year-to-date ETF sales in Canada for the first six months of 2021 Canada ETF sales growth: 35% higher year over year - YTD sales compared with the same period last year Robinhood users: Close to 18 million active users - Discussion of Robinhood S-1 and user engagement Robinhood assets: $80 billion - Assets held on the platform Robinhood crypto assets: $12 billion - Portion of Robinhood assets in crypto Robinhood first brokerage accounts: More than 50% - More than half of users have their first-ever brokerage account there Robinhood new U.S. retail accounts since 2007: 16% - Share of all new retail accounts opened in the U.S. since 2007 Robinhood daily usage: 47% use daily; daily users visit 7 times per day - App engagement metric Robinhood monthly usage: 98% use monthly - Platform engagement metric Robinhood revenue 2019: Under $300 million - Pre-growth comparison year Robinhood revenue 2020: Nearly $1 billion - Annual revenue surged in 2020 Robinhood Q1 2021 revenue: $522 million - Revenue in the first quarter of 2021 Robinhood options share of transaction revenues: 38% - Options generated a large share of transaction revenue despite being a small asset share Robinhood options asset share: 3% - Options represented a small portion of assets but large portion of revenue Base-case tax alpha (simple model): 23 bps annualized - Tax loss harvesting model with 7% total return and constant tax rate Low-return tax alpha (simple model): 5 bps annualized - Model with low capital gains/returns High-return tax alpha (simple model): 49 bps annualized - Model with 10% total return and constant tax rate Canadian low-future-tax case: 46 bps annualized - Model with lower future tax rate and 7% expected return Canadian high-future-tax case: -4 bps annualized - Model with higher future capital gains inclusion and high tax bracket FAJ paper full-period tax alpha: 1.1% - Chartery, Burnham, and Lowe empirical U.S. study from 1926-2018 FAJ paper no monthly contribution case: 0.73% - Tax alpha after removing 1% monthly contributions assumption Estimated Canadian-style tax alpha from paper: ~39 bps - Interpolated estimate using 27% tax rate before transaction costs Transaction costs in paper: 16 bps per year - Estimated drag from turnover in the empirical U.S. study Canadian-style net tax alpha after costs: ~10 bps - Ben’s adjusted estimate after applying Canadian assumptions and transaction costs Great Depression subperiod tax alpha: 2.29% per year - Highest subperiod benefit in the empirical paper, driven by extreme volatility and losses Lowest subperiod tax alpha: 0.57% per year - Lowest subperiod result in the empirical paper Annualized U.S. market return in study: 9.75% - Empirical period used in the tax loss harvesting paper UK DC pension fee cap: 75 bps - Cap discussed in the Bad Advice of the Week segment Average UK pension charge rate: 48 bps - Current average charge rate under the cap Smoothed performance fee period proposed: 5 years - UK proposal to allow performance fees to be smoothed within the cap Canada capital gains inclusion rate discussion: 50% current; possible 75% - Used to illustrate how future tax changes affect tax loss harvesting attractiveness

Pivotal Quotes: "tax loss selling is a tax deferral strategy" — Benjamin Felix: Core clarification early in the tax loss harvesting segment "it should be obvious that it's not obvious" — Benjamin Felix: Ben’s summary of the nuance and uncertainty around whether tax loss harvesting is actually worthwhile "success comes when opportunity meets preparation" — Cameron Passmore: Discussing Amazon’s pandemic-driven growth and broader investing/business preparation themes

Implications: Listeners should treat tax loss harvesting as situational, not universally valuable. In Canada especially, the tax benefit may be modest after costs and can be outweighed by tracking error, fees, and future tax uncertainty.