Breakevens are priced off of nominal not core cpi, so returns are mostly driven by change in price of oil, not rent. Run principal component analysis to validate. Therefore, u are not buying rent, rather change in price of oil vs today’s price . Good luck predicting that…
PCA on returns is the right tool for explaining variance and the wrong tool for explaining level. We’re arguing about different quantities. Run a horizon-based variance decomposition or regress cumulative realized inflation on components, and shelter dominates — because its weight is 5x oil’s and it doesn’t round-trip
I asked AI to adjudicate and it did pretty good job: short term is what matters in TIPs as term premium would drive your returns so negatively that holding loosing positions becomes incredibly difficult…
So as a tl;dr: mkt is not pricing 2.3% geometric inflation mean, it prices higher level with negative skew. As a result, priced in inflation is probably high 2s which is consistent with your analysis and profit on tips is unlikely
AI analysis below:
In the short term, Vitaily is correct — oil dominates. But the full picture is actually more nuanced than either commenter stated, because there are three competing short-term drivers, not just one:
The Three Short-Term Breakeven Return Drivers
1. Oil/Energy (Vitaily’s point — valid)
Energy is ~7.5% of CPI by weight but is by far the most volatile component. Because breakevens reprice daily on expected future CPI, and because oil moves are sharp, fast, and correlated with risk sentiment, energy dominates day-to-day and week-to-week return variance. PCA on breakeven returns would load heavily here. Short-dated breakevens (1–5yr) are especially sensitive since near-term CPI prints are oil-driven.[morningstar +1]
2. TIPS Liquidity Premium (often the biggest short-term driver)
This is the factor neither commenter mentioned. Fed research found that TIPS liquidity premiums explain over 40% of short-term breakeven variation — more than oil or shelter. During risk-off episodes (e.g., March 2020, GFC), breakevens collapse not because inflation expectations fall but because TIPS get dumped for nominal Treasuries. This is pure liquidity noise, completely orthogonal to either oil or shelter.[frbsf +1]
3. Shelter (slow-moving, low short-term impact)
Shelter’s 35% CPI weight matters enormously to cumulative payouts, but OER (owners’ equivalent rent) is a lagging, smoothed series — it barely moves month to month. It contributes almost nothing to short-term breakeven return variance.
Right — and I think we actually agree on the thing my piece was about. The High Floor is a claim about the level: structural shelter and services put the floor near 3% while the market prices ~2.3%. You’ve landed at high-2s with negative skew, which is the same place. The trade was a footnote, and you and the adjudicator are right that a vanilla TIPS position imports oil and liquidity-premium noise that dominate the path — that’s a real trade-construction problem and a fair hit. But the thesis was never the instrument; it was the mispriced level, and on that we’ve converged.
Breakevens are priced off of nominal not core cpi, so returns are mostly driven by change in price of oil, not rent. Run principal component analysis to validate. Therefore, u are not buying rent, rather change in price of oil vs today’s price . Good luck predicting that…
PCA on returns is the right tool for explaining variance and the wrong tool for explaining level. We’re arguing about different quantities. Run a horizon-based variance decomposition or regress cumulative realized inflation on components, and shelter dominates — because its weight is 5x oil’s and it doesn’t round-trip
I asked AI to adjudicate and it did pretty good job: short term is what matters in TIPs as term premium would drive your returns so negatively that holding loosing positions becomes incredibly difficult…
So as a tl;dr: mkt is not pricing 2.3% geometric inflation mean, it prices higher level with negative skew. As a result, priced in inflation is probably high 2s which is consistent with your analysis and profit on tips is unlikely
AI analysis below:
In the short term, Vitaily is correct — oil dominates. But the full picture is actually more nuanced than either commenter stated, because there are three competing short-term drivers, not just one:
The Three Short-Term Breakeven Return Drivers
1. Oil/Energy (Vitaily’s point — valid)
Energy is ~7.5% of CPI by weight but is by far the most volatile component. Because breakevens reprice daily on expected future CPI, and because oil moves are sharp, fast, and correlated with risk sentiment, energy dominates day-to-day and week-to-week return variance. PCA on breakeven returns would load heavily here. Short-dated breakevens (1–5yr) are especially sensitive since near-term CPI prints are oil-driven.[morningstar +1]
2. TIPS Liquidity Premium (often the biggest short-term driver)
This is the factor neither commenter mentioned. Fed research found that TIPS liquidity premiums explain over 40% of short-term breakeven variation — more than oil or shelter. During risk-off episodes (e.g., March 2020, GFC), breakevens collapse not because inflation expectations fall but because TIPS get dumped for nominal Treasuries. This is pure liquidity noise, completely orthogonal to either oil or shelter.[frbsf +1]
3. Shelter (slow-moving, low short-term impact)
Shelter’s 35% CPI weight matters enormously to cumulative payouts, but OER (owners’ equivalent rent) is a lagging, smoothed series — it barely moves month to month. It contributes almost nothing to short-term breakeven return variance.
Right — and I think we actually agree on the thing my piece was about. The High Floor is a claim about the level: structural shelter and services put the floor near 3% while the market prices ~2.3%. You’ve landed at high-2s with negative skew, which is the same place. The trade was a footnote, and you and the adjudicator are right that a vanilla TIPS position imports oil and liquidity-premium noise that dominate the path — that’s a real trade-construction problem and a fair hit. But the thesis was never the instrument; it was the mispriced level, and on that we’ve converged.
I thoroughly enjoy your writing and analysis.. thank u
Nice analysis.
Great read. Good strategy.