Which account holds which asset changes your return without changing your risk
The claim
For an unchanged overall allocation, placing tax-inefficient assets inside tax-advantaged accounts raises after-tax returns, and household portfolios show most people do not arrange their accounts this way.
Why we rate it moderate
The direction is analytic and not in dispute. The magnitude depends on yields, turnover and marginal rates, and the low-yield environment of the 2010s shrank it considerably from the estimates in the original literature.
Tax drag scales with how much taxable income an asset throws off each year. A bond fund distributing interest taxed as ordinary income leaks far more annually than a broad equity index fund that pays a modest qualified dividend and defers most of its return into unrealised gains. If some of your money sits in a 401(k) or IRA and some in a taxable brokerage account, which asset goes where is a free variable — and it changes the after-tax result while leaving the risk profile untouched.
Dammon, Spatt and Zhang solved the joint allocation-and-location problem and found the location effect is first-order rather than a rounding error: for realistic parameters, holding taxable bonds inside the shelter and equities outside it produced meaningfully higher lifetime consumption than the reverse. The intuition is that the shelter is scarce, so it should be spent on the asset it protects most.
Bergstresser and Poterba looked at what households actually do, using the Survey of Consumer Finances, and found allocations that depart substantially from that prescription. Location is a decision most people never make explicitly; assets end up wherever the money happened to arrive.
The size of the gain is genuinely uncertain now. The original work was calibrated when bond yields were high and the drag from holding them in taxable was correspondingly large. A decade of near-zero yields made the whole question smaller, and higher yields have made it matter again — which is a good illustration of why an effect size should be quoted with its conditions attached.
The rules of thumb, in rough order: put high-yield bonds, REITs and anything with high turnover inside the tax-advantaged accounts; keep broad equity index funds in taxable, where their gains defer and their qualified dividends are taxed lightly; and hold any Roth space for the highest-expected-return assets, because the tax you avoid there is the tax on growth. All of these are second-order compared with contributing more and paying less in fees.
Where this breaks down
- Location decisions constrain rebalancing. If the bonds all sit in the 401(k), rebalancing has to happen there — usually fine, occasionally awkward.
- The benefit depends on holding both account types with meaningful balances. For someone whose savings are almost entirely in a 401(k), this note is close to irrelevant.
- Tax-loss harvesting and step-up-in-basis at death cut the other way for equities in taxable, and a full analysis has to include them.
- This is a smaller lever than fees, contribution rate, or capturing the employer match. Do not let it delay any of those.
Sources
Follow these rather than taking our word for the summary.
Robert M. Dammon, Chester S. Spatt and Harold H. Zhang (2004). Optimal Asset Location and Allocation with Taxable and Tax-Deferred Investing
The Journal of Finance, 59(3), 999-1037
Finding: Holding taxable bonds in the tax-deferred account and equities in the taxable account raised lifetime welfare substantially for realistic parameter values.
Daniel Bergstresser and James Poterba (2004). Asset allocation and asset location: household evidence from the Survey of Consumer Finances
Journal of Public Economics, 88(9-10), 1893-1915
Finding: Household asset location departs substantially from the tax-efficient arrangement, with many holding tax-inefficient assets outside their sheltered accounts.