# Protecting against inflation: what actually works The three previous articles set the scene. They named the enemy (inflation regimes, and the financial repression that comes with them), showed how to measure it ([[tracking-inflation]]) and took apart the way it attacks a plan ([[inflation-and-withdrawal-rates]]). This one assembles the arsenal. The market for "inflation protection" is crowded with products that protect against nothing and advice that mixes up horizons. So the defenses need sorting, and not by popularity: by the **kind of evidence** behind them. Contractual protection rests on indexation written into a contract or into law (linkers, the public pension). Structural protection rests on an economic mechanism that tracks prices over the long run (global equities, real assets whose cash flows are free to reprice, fixed-rate debt you owe). Episodic protection wins **during** episodes and promises nothing in normal times (gold, broad commodities, trend following). Behavioral protection is the settings of the plan itself, and it is usually the cheapest of the lot. Then comes the blacklist, what sells as protection and is not. Then how to put it together phase by phase, because, as we saw, the uncovered phase, the bridge years, is the one to index. This chapter is deliberately a summary armed with links. Each defense has its own article. What stays here is the sorting, the sizing, and the buying mistakes. ::: cle The rule for buying protection Judge an inflation hedge on **two** questions, never one. Does it follow prices, and through what mechanism: a contract, an economy, an episode? And what does it cost you in the regimes where inflation never shows up? The best answers score on both. A delayed public pension is contractual, and subsidized. Linkers are contractual, at the price of the breakeven, with next to no markup. Indexation flexibility is free. Global equities are structural, and you are paid to hold them. That is where to start. Gold and commodities come later, in measured size, because they do **cost** something in normal times. ::: ## Tier 1: the contractual, indexation in writing **Linkers and the indexed ladder** ([[inflation-linked-bonds]]). Linkers are inflation-linked bonds, TIPS on the US side. They are the only protection whose mechanism is a **contract**: inflation plus a guaranteed real yield. At the current breakeven, around 2%, insurance against anything above that is cheap by historical standards. The size runs 25 to 50% of the bond sleeve, plus a ladder covering the floor of the uncovered phase for plans with no give in them. The limit is real-rate risk before maturity, as 2022 showed, which argues for short durations or for holding to maturity. **The public pension, and the way to buy more of it: delaying your claim** ([[annuities-and-safety-first]]). This is the annuity indexed by **law** and backed by the state. It is the most underrated inflation asset a household owns, and the only one you can **buy** at a subsidized price. Every year you wait raises the check for life, and each extra dollar of future pension is a dollar of indexed floor for life. Before any financial product, price what delaying would buy you: that is line one of the indexation program. The limit is political. Raises can be pushed back or set below the index for a stretch. Treat that as a prudence haircut, not as a disqualification. The rules where you live decide the exact price of waiting ([[us-healthcare-and-social-security]]). **The rents you collect** ([[real-estate-in-retirement]]). Almost contractual: they follow prices with a lag, and rent control can cap them in the very years you need the raise. A good indexed flow, never the only one. ## Tier 2: the structural, an economy that follows prices **Global equities, the slow protection** ([[stock-bond-allocation]], [[international-diversification]]). The point that confuses everyone deserves one clear statement. **During** the episode, equities **lose** in real terms, because multiples compress: 1973 to 1974, and again in 2022. They are **not** an episode hedge. But over 5 to 15 years companies reprice, since their revenues and their assets are real. Over that horizon equities are historically the **only** productive asset that catches up with cumulative inflation and then passes it. The engine of the plan is also its deepest protection, which is one more reason never to undersize it out of fear of inflation. That was the mistake of 1975: sell everything for "safe" cash, then get eroded through the decade of catching up. ::: figure actions-rattrapent The S&P 500 with dividends reinvested, in dollars, deflated by the US CPI, base 100 at the end of December 1972. The September 1974 low wipes out more than half the purchasing power, and the index climbs back above the line only in January 1985. It has never fallen below it since, and that is what we mean by catching up for good. ::: **Fixed-rate debt you owe** ([[real-estate-in-retirement]], [[leverage-and-margin]]). This is negative indexation. A loan fixed at 1.5% while prices run at 4% is repaid in melted money. Keeping a cheap fixed-rate loan is, in accounting terms, holding an inflation asset the size of the balance outstanding. All of it under the prudence conditions already set out. **The value factor, half a building block** ([[factors-in-retirement]]). Of all the tilts you can put on the equity sleeve, value has the best report card in inflationary regimes. Free in expectation, give or take the gap to the index you have to sit through. ## Tier 3: the episodic, the crisis winners, in measured size **Gold** ([[gold-in-retirement]]). Episodic in the purest sense: spectacular from 1971 to 1980, respectable in 2022, dead from 1980 to 2000. Hold it at 5 to 10%, sharing the slot. **Broad commodities** ([[defensive-assets]]). A genuine episode hedge, since half the inflation of 2021 and 2022 **was** commodities. Their negative carry is chronic, and for an individual they are dominated by gold plus trend. **Trend following** ([[managed-futures]]). The systematic winner of the **long** regimes, with 2022 as the demonstration: short bonds, long energy. Take 5 to 15% if you can sit through the winter. The logic of the tier is simple. These blocks cost something in normal times, in carry and in a gap to the index that is painful to hold. You buy them in **measured size**, written into the plan, never in reaction to the news. ## Tier 4: the behavioral, the free settings These defenses get forgotten because you cannot buy them. They are the best value in the whole arsenal. The first is the **indexation freeze** on withdrawals after a red year, which disarms half the squeeze, spending climbing with prices while the portfolio stalls, at a cost spread over years ([[fixed-inflation-adjusted-withdrawal]], [[inflation-and-withdrawal-rates]]). The second is the **budgeted drift**, a real spending drift of 0.3 to 0.5 points a year, which pays your true inflation up front in the sizing instead of discovering it at 80 ([[tracking-inflation]]). The third is **real flexibility** in the budget, where a floor at 75% of the comfort level absorbs years of indexation shortfall without drama ([[flexibility-in-practice]]). A plan with those three settings and Tier 1 is already 80% covered. Tiers 2 and 3 do the rest. ::: attention The blacklist: what sells as protection and is not - **Stable value funds and money market funds.** These are the designated victims of financial repression. The rate they pay **follows** inflation by years. Indispensable as liquidity, a category error as protection. - **Dividend stocks**, sold as a way to keep up with the cost of living. They are ordinary equities in a less diversified version ([[defensive-assets]]). The dividend is not indexed. It is discretionary. - **Floating-rate bonds.** They follow **short rates**, not prices. Better than fixed during an episode, but with no contract behind them this is a duration tool, not an indexation tool. - **Structured "inflation" products.** The indexation is capped, averaged, conditioned. Read the formula. What is left of the protection after the clauses rarely covers the fees. - **Crypto.** The digital gold thesis is still a thesis. Through its one lived inflation episode, 2021 and 2022, it behaved like a risk asset with a beta of 2 to 3, falling 70% **while** prices accelerated. Maybe one day. Not protection yet. A withdrawal plan does not buy theses ([[simulator-traps]]: what has no track record has no place in the sizing). ::: ## Putting it together by phase, and the buying mistakes **The uncovered phase** is the heart of the problem ([[inflation-and-withdrawal-rates]]). Its indexation inventory is short, and you fill it in layers. All of Tier 4, which is free. Tier 1 in size: linkers at 25 to 50% of the bond sleeve, the ladder for the floor, a priced decision on delaying the pension. Tier 3 in measured doses: gold 5 to 10%, trend 5 to 15%. **The covered phase** is a different animal. The indexed pension covers the floor, and the exposure left over is only the comfort layer. Tier 2 (the portfolio) and Tier 4 usually do the job, and the Tier 3 doses can come back down at the annual review of the year you claim ([[the-annual-review]]). **The three buying mistakes** are documented after every episode. The first is to **buy after**. Linkers and gold bought in 2023, at the peak of the fear, when breakevens and premiums were at their highest. Protection is bought in normal times, when it costs nothing, on the same contrarian logic as valuations ([[valuations-and-cape]]). The second is to **buy too much**. That is the portfolio 40% in gold and commodities, built by the saver traumatized by 1980 or by 2023, who then gets eroded by twenty years of disinflation. Over-protecting against the last regime is the bet **most** exposed to the next one. The third is to **sell everything** for real assets in the middle of the acceleration. That is the mistake of 1975: abandoning the engine at the moment it is cheapest. ::: exemple Sonia's indexation program, tier by tier Sonia is 54 and leaves in a year. She has $1.7M, a floor of $42,000 and an estimated pension of $19,000 at 66. Her uncovered phase runs 11 years, and her indexation inventory is all nominal. Her program follows the order of the tiers. Tier 4, at zero cost: an indexation freeze written into the rule, a spending drift of 0.4 points a year, a floor reworked down to 76% of comfort. Tier 1: she commits to delaying her pension claim by two years, the cheapest indexed lifetime income she will ever be offered. Then 10% of the portfolio in short linkers, and an approximate indexed ladder covering years 1 to 6 of the floor. Tier 3: 6% gold and 6% trend, taken out of the excess stable value and cash. Tier 2: nothing to do, since her 62% in global equities was already the slow protection. The verdict comes in. The broad sample goes from 12.4% to 8.1%, the 1966 vintage is ridden out, the central case barely moves. The whole program costs about 0.3 points of expected return, half of it handed back by the pension decision. Inflation is still her main risk. But it is now **paid for**, not hoped away. ::: ## The essentials - Sort by the kind of evidence. The **contractual** (linkers, the indexed ladder, the public pension and a delayed claim) is where you start. The **structural** is global equities, the slow protection, plus fixed-rate debt you owe and a value tilt. The **episodic** (gold, commodities, trend) is taken in measured size, because it costs in normal times. The **behavioral** (indexation freeze, budgeted drift, a floor with slack) is free, and comes first. - Here is the point that confuses everyone. Equities lose **during** the episode and win **after**. Neither a crisis hedge nor a long-run victim, the engine stays the engine. - Blacklist: stable value and money market funds (the victims of financial repression), dividend stocks, floating-rate bonds (a duration tool, not an indexation tool), structured products with a formula, crypto (a thesis, not a track record). - The assembly follows the phase. The uncovered phase is the one you index (Tiers 4 and 1 in size, Tier 3 in doses). The covered phase already is, through the pension. - The three mistakes are buying after the episode at the top price, buying in overdose, which is a bet on the last regime, and selling the engine in a panic. Protection is bought in normal times, at written sizes, and checked against the broad sample and the 1966 vintage. --- ## Going further - The other articles of this part, for the whole file: [[tracking-inflation]], [[inflation-and-withdrawal-rates]], [[hyperinflation-and-extremes]]. - The building blocks: [[inflation-linked-bonds]], [[gold-in-retirement]], [[managed-futures]]. - Ilmanen, *Investing Amid Low Expected Returns*, the inflation chapter: the hedges compared, with numbers. - Neville, Draaisma, Funnell, Harvey and Van Hemert, "The Best Strategies for Inflationary Times" (2021). The reference study on what actually worked across eight inflation episodes. Trend and commodities at the top, stocks and bonds at the bottom. It is the academic validation of this chapter's ranking.