Even for a rolling investor (whose returns are also driven by mark-to-market spread moves), initial spreads explain nearly half of 5yr forward returns.
Why it is true (even for a rolling/MTM investor) Credit excess returns for a rolling investor (constant-maturity index style, or someone continuously reinvesting as bonds mature/roll) can be decomposed roughly as:
Excess return≈ Carry (spread income) - Duration x ΔSpread+other (defaults, convexity, etc.)
- The carry term is directly tied to the level of spreads.
- The mark-to-market term comes from spread changes. Spreads are mean-reverting over multi-year horizons, so the starting level strongly influences both the path of subsequent spreads and the expected direction/magnitude of MTM moves.
The starting points matter:
- Starting from wide spreads → high ongoing carry + higher probability of subsequent tightening (positive MTM contribution).
- Starting from tight spreads → low carry + higher probability of subsequent widening (negative MTM contribution).
This dual effect is why the initial spread retains substantial explanatory power even after accounting for the MTM channel.
Over 5-year horizons, analyses of US and European IG/HY indices typically show that the starting spread (or yield) explains on the order of 40–60%+ of the variation in subsequent excess (or total) returns — “nearly half” is a fair description of the usual range.
Supporting evidence
- Janus Henderson (and similar index analyses): Correlations between starting yields and subsequent 5-year annualized total returns are high (≈0.77–0.87 for US/Euro IG and HY). That implies R² in the ~0.6–0.75 range for total returns; the excess-return version (driven by spreads) is in a similar ballpark.
- Bucket analyses (e.g., Aptus and others): Forward excess returns rise systematically with starting OAS buckets. Starting OAS <3% has historically produced negative average excess returns; higher starting spreads produce materially higher subsequent excess returns.
- Decomposition work (AQR-style and others): Monthly credit excess returns are approximately (starting spread)/12 − duration × Δspread. Over multi-year rolling periods the integrated spread path (heavily influenced by the starting level via mean reversion) accounts for a large share of cumulative results.
- Broader literature on yield/spread predictability in fixed income (Ilmanen and others) consistently finds that starting levels are among the strongest predictors of multi-year returns precisely because carry dominates and spreads/yields mean-revert.
Caveats
- The relationship is stronger over 3–5+ year horizons than over short periods, where pure MTM volatility can dominate.
- It is not deterministic. Path dependency, regime shifts, and large default clusters can reduce the explanatory power in specific windows.
- “Nearly half” is an approximate description of typical R² figures in the literature; exact numbers vary by sample period, index (IG vs HY, US vs Europe), and whether you look at total or excess returns.
- Current tight spreads (as of late 2025/early 2026 in many series) imply lower expected excess returns on this framework, all else equal.
Bottom line: The statement is consistent with how credit returns actually work for rolling investors. Initial spreads set both the carry and (via mean reversion) a meaningful portion of the expected MTM path, so they continue to explain a large fraction of multi-year forward excess returns.
Carry is not destiny.
Spread mean reversion is not perfectly timed. But the level you start from still explains a surprisingly large share of where you end up.
In High Yield, your 5-yr return has historically landed within ~1 point of the starting yield...whether you bought in a crisis or at the tights.
Starting YTW today is ~8%
Graph source - Alliance Bernstein
