Are high-yield bonds still worth holding once you count taxes and how closely they move with stocks?
Alpha · deepseek-v4-pro
My answer is: not on the headline yield, and not as an equity diversifier — but the case isn't uniformly dead. The question turns on three moving parts: what defaults take, what tax takes, and what "moving with stocks" actually means.
1. The 8% yield is not the return
The spread of roughly 3.3 percentage points over Treasuries exists to compensate for defaults. The manuscript's long-run numbers — a Moody's credit-loss figure around 2.78% a year after recoveries, or an S&P default rate near 4.1% with roughly a third recovered — point to a similar place: expected losses after recovery consume most of the spread.
So the honest pre-tax comparison is not 8% versus 5%; it is closer to 5.2% versus 4.98% for a Treasury, before tax and before any price move. That is a much thinner edge than the screen suggests.
One weakness in the worked example: comparing the high-yield index effective yield with the 10-year Treasury is not duration-matched. High-yield indexes typically carry much shorter duration than a 10-year Treasury, so the fairer Treasury comparison may be shorter. That can move the small gap in either direction, depending on the curve.
2. Tax changes the ranking…
Beta · gpt-6-sol
Beta: The manuscript’s main warning survives, but its worked example should not be presented as an estimate of what a high-yield investor will earn.
The dated market figures check out: FRED shows a 7.99% high-yield index effective yield and a 3.26-percentage-point option-adjusted spread on October 29, 2026; the 10-year Treasury yield was 4.98%. On October 30, the high-yield figures were 8.14% and 3.44 points. The spread is calculated against a Treasury *curve*, not simply by subtracting the cited 10-year yield from the index yield. (fred.stlouisfed.org)
I would push back harder than Alpha on the conclusion drawn from 3.71% versus 3.78%. The arithmetic works *if* the manuscript’s assumed 2.78-point loss is realized and offsets gains taxed at 15%. But an historical issuer loss average is not necessarily the prospective loss of today’s market-value-weighted index. Nor does an index effective yield, less that loss average, become a reliable one-year return: prices, timing, recoveries and fees remain outside the calculation. FRED also says the index excludes defaulted securities. The comparison to a 1…
Gamma · gemini-3.1-pro
Gamma: Alpha and Beta have thoroughly dismantled the false precision of the manuscript's worked example, and I agree with their assessment: comparing an index effective yield with a historical average issuer-weighted default loss and a non-maturity-matched Treasury is mathematically messy. However, the manuscript's broader qualitative premise holds up well, particularly when looking outside the US.
To add to the verification, I checked the manuscript's claims regarding international tax treatments, as these highlight just how conditionally "worth holding" these bonds are depending on the investor's jurisdiction:
* Korea: The manuscript is correct. Interest income earned from Korean sources is subject to a 15.4% withholding tax (including local income tax). Furthermore, financial income (interest and dividends) exceeding KRW 20 million annually is subject to global/comprehensive taxation at progressive rates. (pwc.com)
* Japan: The manuscript is also correct here. Interest on listed bonds is taxed separately at a flat rate of 20.315% (comprising 15.315% national tax and 5%…
Researcher · gpt-6-luna
Researcher — verification
The strongest qualification raised this round is well founded: the worked example’s arithmetic can be reconstructed, but it should not be treated as a forecast or a reliable head-to-head estimate. It subtracts a historical credit-loss average from an index yield, then compares the result with a 10-year Treasury yield. Those are different measures and maturities; the tiny resulting gap cannot support a firm conclusion.
A few specific checks:
- Market figures: FRED identifies the cited series as the ICE BofA US High Yield Index Effective Yield and Option-Adjusted Spread. Its notes define the yield as an index measure and the spread as a distinct series; the OAS is not simply the index yield minus the cited 10-year Treasury yield. I could verify the series definitions, but not independently confirm the manuscript’s exact October 29–30, 2026 observations from the pages returned in this search.
https://fred.stlouisfed.org/series/BAMLH0A0HYM2EY
https://fred.stlouisfed.org/series/BAMLH0A0HYM2 (fred.stlouisfed.org)
- Historical losses: A Moody’s-hosted report wa…
The Arbiter · gpt-6-sol
Round mediation: The voices agree on the main distinction: an index yield is not an investor’s after-tax return, and high-yield bonds should not be assumed to cushion a stock sell-off. Their real difference is how much the manuscript can infer from its example. Alpha treats the 3.71% versus 3.78% result as a useful illustration under stated assumptions; Beta warns that the seven-basis-point gap is too small, and the inputs too mismatched, to establish a return ranking. Gamma emphasizes that the tax result also changes across countries and accounts.
The source record is firmer on the tax mechanism than on the example’s outcome. IRS guidance supports the capital-loss limit and carryforward, but the tax benefit from a particular bond default depends on how and when a loss is realized. The manuscript cannot simply assume that every default produces an immediately usable deduction at 15%. (irs.gov)
Several publication checks remain open in this round. The Researcher could not independently read the exact October 29–30 FRED observations or S&P’s 4.10% table entry. Moody’s historical loss figures were located in…
The Arbiter's Final Verdict · gpt-6-sol
Verdict
The manuscript’s premise stands, but its numerical conclusion is too precise. High-yield bonds can still have a role, but an 8% quoted yield is not an 8% expected, spendable return, and they should not be counted on as protection when stocks fall. Whether they are worth holding depends on the price paid for credit risk, the account and tax rules, and what risk they are meant to diversify. The record does not establish a general rule to hold or avoid them.
The strongest objection to the manuscript is its comparison of 3.71% for high yield with 3.78% for a Treasury. The arithmetic follows its assumptions, but the seven-basis-point difference cannot bear the weight of a return ranking. It combines an index yield, a historical issuer credit-loss average, an assumed immediately usable tax benefit, and a non-maturity-matched 10-year Treasury yield. Price changes, fees and the investor’s actual tax treatment are left out.
What the publication record supports—and what it does not
- Yield and spread: FRED identifies the relevant ICE BofA high-yield effective-yield and option-adjusted-spread series. Beta reported reading the manuscript’s 29–30 October…
