Tax-Loss Harvesting with Bonds
Tax-Loss Harvesting with Bonds
Tax-loss harvesting captures unrealized losses in bonds to offset taxable gains elsewhere, increasing after-tax returns without changing the portfolio's economic exposure via substitute positions.
Key takeaways
- When a bond position has declined in value, selling it realizes the loss, which can offset capital gains (stocks, other bonds) or up to 3,000 of ordinary income per year.
- The wash-sale rule prevents repurchasing the same or substantially identical bond for 30 days before or after the sale; substitution with a similar but not identical bond preserves economic exposure while allowing loss realization.
- Common substitute pairs: AGG-BND (both broad IG bond indices), TLT-IEF-SHV (varying Treasury maturities), LQD-VCIT (IG corporate bonds).
- Harvesting is most valuable when gains have been realized elsewhere (stock portfolio surge) and marginal tax rates are high (35%+ for high earners).
- Over-harvesting (repeatedly swapping bonds to realize losses) can trigger IRS scrutiny; the strategy is defensible if positions are held with genuine intent, not traded purely for tax purposes.
The mechanics: realizing losses and offset
A simple scenario: An investor bought BND (Vanguard Total Bond Market ETF) for 100k in early 2021 when yields were low. In 2022, rates rose sharply; BND declined 15% to 85k. The investor has an unrealized loss of 15k.
Separately, the investor's stock portfolio (VTI) has appreciated 80k in 2022, a substantial capital gain. If the investor sells 80k of stock gains and realizes 80k of taxable income, they owe ~15k–28k in taxes (depending on federal, state, and NII tax brackets; assume 35% marginal rate for a high-earning household: 80k × 35% = 28k).
But if the investor harvests the 15k bond loss before year-end 2022, they can offset the stock gain:
Net taxable gain = 80k stock gain — 15k bond loss = 65k. Taxes owed ≈ 65k × 35% = 22.75k. Tax saved: 28k — 22.75k = 5.25k.
The tax saving is substantial (5.25k on a 100k initial bond position is 5.25% annual return, tax-deductible). After realizing the loss, the investor can immediately repurchase a substitute bond position (e.g., AGG) to restore bond exposure, avoiding the 30-day wash-sale window.
Identifying substitute bonds and the wash-sale rule
The IRS prohibits repurchasing the same security within 30 days before or after a loss-harvest sale; this is the wash-sale rule. However, purchasing a "substantially different" security is allowed. The IRS has not provided a rigid definition of "substantially different," but general guidance is:
- Same issuer, different maturity: ✓ (BND and a ladder of individual Treasuries)
- Same maturity, different credit (e.g., Treasury vs. corporate): ✓ (IEF vs. LQD)
- Different ETF, same index methodology: ✓ (BND vs. AGG; both broad IG bond indices)
- Different ETF, different index: ✓ (IEF vs. TLT; different Treasury maturities)
- Substantially different credit: ✓ (IG corporate vs. HY)
Common harvest-and-substitute patterns:
| Sold | Reason | Substitute | Hold Duration | Then Repurchase |
|---|---|---|---|---|
| BND | Down 15% | AGG (both IG, different metrics) | 31 days | BND |
| TLT | Down 20% | IEF (shorter-duration Treasury) | 31 days | TLT |
| LQD | Down 12% | HYG (spreads widened) | 31 days | LQD |
| EWG | Down 10% | Bunds (via BNDX or unhedged) | 31 days | EWG |
The strategy requires discipline: after selling a bond at a loss, you must wait 30 days before repurchasing the exact same security. During this window, you hold a substitute, incurring temporary performance divergence risk (if your original bond rallies while your substitute lags, you lose the relative gain).
Example: On Dec. 15, 2022, you sell 100k of TLT (long-duration Treasury ETF) at a loss after the Fed's aggressive hiking drove TLT down 25%. You buy 100k of IEF (intermediate-duration Treasury ETF) as a substitute, maintaining treasury exposure with a slightly shorter duration. The key assumption: IEF and TLT move similarly (correlation ~0.95), so the substitute is an economic equivalent.
On Jan. 15, 2023 (31 days later), if rates have stabilized or declined, both TLT and IEF have recovered. You sell IEF and repurchase TLT, completing the cycle. If TLT rallied more than IEF during the 31-day hold, you capture extra gain on the repurchase; if IEF outperformed, you forgo that relative gain.
In practice, Treasury futures curves are flat-to-steep, so duration mismatch matters less than credit sensitivity. The substitute swap of TLT-IEF is usually safe; the swap of IG-corporate to HY-corporate might face more tracking error.
Harvest timing and tax-planning integration
Tax-loss harvesting is most valuable when:
-
Gains are realized elsewhere. If a stock portfolio surges 100k, harvesting 15–20k of bond losses offsets the gain efficiently.
-
Marginal tax rates are high. For investors in the 35%+ bracket, saving 1 point of tax is 3,500+ per 100k realized loss. Harvesting is less valuable for lower-income investors in 12% tax brackets (saving 1,200 per 100k).
-
Losses exist and bonds are underwater. In 2022, after rising rates demolished bond funds, virtually all bond positions were underwater; harvesting was possible and valuable. In 2021, when yields were low and bonds had appreciated, losses were rare.
-
The portfolio is large enough. For a 100k portfolio, harvesting 10k of losses nets ~3,500 in tax savings, meaningful. For a 10m portfolio, harvesting 100k of losses nets ~35,000, life-changing.
Planning cycle for a household with significant stock gains:
September–October: Review realized gains in stock portfolio. If gains exceed losses, estimate net taxable gain.
October–November: Audit bond positions. Identify any underwater bonds (common after rising rates). Harvest losses strategically.
November–December: Hold substitutes to avoid wash-sale windows. Reallocate if necessary.
January: Repurchase original positions after 31 days, restoring portfolio to intended allocation. The year-end harvesting locks in losses for the current tax year; year-end timing optimizes the deduction.
Complications: wash-sale rules and ETF swaps
The wash-sale rule is applied strictly by the IRS. Common traps:
1. Inadvertent same-security purchase. If you sell BND on Dec. 15 and buy BND on Jan. 10, the IRS treats the Jan. 10 purchase as violating the wash-sale rule (it is within 30 days before the sale... wait, no, this is within 30 days after). The timeline: no purchases of BND from Nov. 15 (30 days before sale) through Jan. 14 (30 days after sale).
2. Conflating different securities. BND and AGG are both broad bond indices, but they are different ETFs with slightly different methodologies (BND uses an extended index including Treasuries; AGG uses the Bloomberg Aggregate). They are not substantially identical, so swapping BND for AGG avoids wash-sale issues.
3. Spousal accounts. If a spouse purchases the same security (or a substantially identical one) within the 30-day window, the wash-sale rule applies to the selling spouse's loss. Coordinating tax planning between spouses requires caution.
4. Dividend reinvestment. Many bond ETFs distribute monthly coupons. If you harvest a loss in December and the substitute ETF distributes a dividend on Jan. 5, then repurchase the original ETF on Jan. 15, the dividend is not a direct purchase, so it doesn't trigger a wash-sale violation. But the wash-sale rule can still apply to the explicit repurchase on Jan. 15.
Best practice: Consult a tax advisor before implementing large harvest strategies; the IRS has been scrutinizing harvesting for abuse (specifically "round-tripping"—selling and repurchasing with no economic purpose). A legitimate harvest is defensible; a circular trade executed purely to generate tax losses may not be.
Harvesting with bond funds versus individual bonds
Harvesting with index ETFs (BND, AGG, TLT) is straightforward; they are fungible and liquid.
Harvesting with individual bonds requires more care. If you own a ladder of individual Treasuries (a 2-year, 5-year, 10-year, etc.), you can harvest losses on the 10-year, then replace it with a different 10-year issue (e.g., a Treasury bond maturing in 2034 vs. one maturing in 2033). The Treasuries are substantially different (different maturity dates, different coupons, different issuance), so no wash-sale issue arises.
However, individual bond ladders are less tax-efficient in one respect: you cannot easily substitute a 10-year Treasury for a 10-year corporate bond without changing credit exposure. The substitution logic is tighter for individual bonds than for funds.
Most retail investors use ETFs for harvesting because ETFs are liquid, transparent, and easily substitutable.
Harvesting limits and year-end crunch
The IRS limits loss deductions to 3,000 of net losses per year above capital gains. Excess losses carry forward to future years. A household with 50k of realized losses in 2022 can deduct 3,000 against ordinary income in 2022, carrying 47k forward to offset 2023 gains or subsequent-year ordinary income (3,000 per year).
This creates a year-end crunch for large portfolios: if you have significant losses, harvesting early is suboptimal (the deduction is delayed). Harvesting late in December ensures the deduction applies to the current year's income.
Next
Tax-loss harvesting is one component of tax-efficient bond management. The next article examines passive bond strategies—simply buying broad bond indices with minimal active management or harvesting—as the default for most investors, with tax efficiency and cost-effectiveness built in passively.