Why Allocators Get Fired: The Dual Mandate
Allocators get fired for two reasons. They lose more in a drawdown than the client was prepared for, or they trail the benchmark long enough that confidence runs out. Whether it’s an institutional fund or a family’s savings, most firings trace back to one of those two stories, and the ending tends to be the same. Faith breaks, the plan is abandoned, the compounding stops.
Think of them as two budgets that together form the dual mandate. The absolute budget is for losses: how far the account can fall before it breaks the mandate the client agreed to. The relative budget is for patience: how long it can trail the benchmark before confidence runs out. Every portfolio spends from both at once, whether its manager realizes it or not.
Building a portfolio that stays inside both budgets is the whole purpose of the endowment method we began in Part One. This piece explains why the problem is hard; the rest of the series shows how the method resolves it.
Start with the absolute budget, the expected loss limit that the investor and advisor thought was possible but still digestible. Somewhere in writing, everyone agreed how much risk this money would take: the board’s policy, the advisor’s risk tolerance bands, the “moderate growth” portfolio the client believed they chose. Some go all the way and hold nothing but stocks, aggressive and counting on a long horizon. But a 50% or 60% decline arrives eventually, and it runs deeper than almost anyone truly bargained for. That is what stocks do, and even the investors who set out to be aggressive, long horizon and all, struggle to stick with the plan once it is actually happening, which is why most stop short of all-equity. The sensible default is a mix of stocks and bonds, the traditional global 60/40.
But dialing down to a moderate mix doesn’t buy the calm it promises. The loss a client can truly stomach is smaller than anyone admits, and even a balanced portfolio serves up stretches that are brutally hard to sit through. A 30% or 40% decline asks more than most investors can give. Near the bottom, when the headlines are terrible and every brokerage statement is a fresh insult, many capitulate and sell. That is how bottoms form; the market stops falling when the last rattled holder has sold. Charlie Munger’s first rule of compounding is never to interrupt it needlessly, and a loss beyond what the client can bear is the most common way that rule breaks.
The safe answer isn’t safe enough. Bonds soften the ride but don’t solve the problem; when stocks fall hard, they drag the whole 60/40 down with them. Sometimes stocks and bonds fall together, which tends to happen during bouts of inflation. Exhibit A traces the full record: over the past half-century a global 60/40 has dropped more than 20% again and again, and in the worst episodes far more: roughly 37% in the 2008 crisis, ~26% in the 1973–74 stagflation bear, ~25% as the dot-com bubble unwound, and ~21% as recently as 2022.
So the “balanced” portfolio has fallen more than 20% four times in fifty years, and as much as 37% in 2008. An investor can understand that in advance and still find it more than they are prepared to sit through, which is the gap that matters. If even the traditional 60/40 blows past what the client turns out to be able to carry, splitting money between stocks and bonds isn’t enough. The mandate demands something built to fall less.
The one guaranteed benefit of the plain 60/40 is that you can never look wrong against it; you can’t fall behind the benchmark when you are the benchmark. But that safety against losing to the benchmark locks in every decline in Exhibit A above. Toggle it to stocks alone and the troughs cut deeper still, or stretch the window back to 1926 and the same breaches repeat across a full century. Stocks and bonds recover, which matters only for an investor who survives the fall to collect the rebound, and not every investor does. To make it fall less you have to build something different, and anything different from the benchmark will, sooner or later, trail it. Trailing too long, or by too much, is the second way an allocator or advisor gets fired.
The obvious response to the absolute risk problem is to build something better and more robust than the plain 60/40, adding genuine diversifiers (i.e., assets that tend to zig when stocks and bonds zag) so the portfolio has a better chance at avoiding a 40% drawdown. It has worked quite well, and it is why they call diversification the only free lunch in investing. But the bill actually lands on the other budget. A portfolio that looks different from the benchmark performs differently, and in the long stretches when the benchmark is roaring, different typically means lagging. Practitioners call it tracking error. In plain terms, it means stretches, sometimes years long, when a good portfolio trails the popular index with nothing wrong except that it was built to do something else.
Take a plain global 60/40 and set beside it a mix that spreads the same money across four sources of return instead of two: 40% global stocks, 30% bonds, 15% gold, and 15% trend-following.
| Full Period · 1973–2025 | Diversified 40/30/15/15 | Global 60/40 | Difference |
|---|---|---|---|
| Growth of $1 | $104 | $66 | +$38 |
| Annualized return (CAGR) | 9.2% | 8.3% | +0.9% |
| Volatility (annualized) | 7.8% | 9.9% | -2.1% |
| Maximum peak-to-valley decline | -21.5% | -37.3% | +15.8% |
| Sustained Underperformance vs. 60/40, cumulative total return | |||
| 1988–2000 · the long bull (12.7 yrs) | +153% | +232% | -79% |
| 2009–2018 · post-crisis bull (9.6 yrs) | +85% | +122% | -37% |
Over the past half-century the broader construction did what the absolute budget demands. It cut the worst loss by more than a third, from ~37% to ~21.5%, lowered volatility by two points, and still earned more than the 60/40, turning a dollar into $104 rather than $66. Its edge showed most in the crises. In the 1973–74 stagflation bear the 60/40 fell ~26% while the diversified mix fell just ~12%, because gold and trend-following have historically held up in the same inflation that sinks stocks and bonds together. The diversified mix held up again in the 2008 deflationary crash, drawing down only ~21.5% versus the 60/40’s ~37%. Over this period it came out ahead on return, volatility and worst loss. Yet look at the shaded years. Through the 1990s bull, and again for most of the decade after 2008, the 60/40 steadily closed the gap while the diversifiers held the portfolio back. An allocator graded on a three-year scorecard would have thrown the diversifiers overboard right before they paid off again.
The growth chart understates how those lean years felt, because the wealth line is the analyst’s hindsight view. The client’s view is the relative score in real-time, which is what Exhibit C plots. For every dollar the plain 60/40 grew to, how many did the diversified mix grow to? That running ratio is the client’s real experience; the +56% where it stands at the end of 2025 means the mix is worth about 1.6 times the 60/40. A rising line means the mix is growing faster, while a falling line means the 60/40 is. A falling line does mean ground lost to the benchmark, but it does not necessarily mean the portfolio is losing money, because the line tracks only the gap between the two, not the dollars either one earned. The account can be generating positive returns the whole time the line slides lower; it’s just generating lower returns than the stated benchmark. Looking at the historical record, the mix built a commanding lead in the 1970s, then bled it back twice, through the 1990s bull and again for the decade after the 2008-2009 crash, each time surrendering much of a hard-won advantage for years on end.
The index takes its losses in crises; the diversified mix loses relative ground in booms and builds its lead in the crises, when the protection is being paid for. A client holds this portfolio through 2008 and feels like a genius, then holds it through 2014 and feels like a fool. Nothing about the portfolio changed, only the budget being spent. And look at the right edge. The mix ends 56% ahead, yet a client who joined near the 2009 peak has spent 16 years watching the lead sit below where it was when they arrived. Nobody anchors to where the line started; they anchor to the best it ever looked. That is how a portfolio can be winning the war and still feel like it is losing.
The better portfolio agreed to lose to the benchmark, patiently, in the good times, in exchange for losing far less in the bad ones. In this period, protection came at no cost to long-run returns; the mix finished ahead. What it cost was a decade of looking wrong, paid from the relative budget.
For an advisor or an investment committee, this is the slow firing, the “so what are we paying you for” conversation that lands after two or three years of trailing the index everyone is watching. Impatience isn’t just a small-investor failing. In a landmark Journal of Finance study, Goyal and Wahal tracked roughly 3,400 large institutions and found the same reflex at scale; they hired managers after hot streaks and fired them after cold ones, and the swap bought nothing. The managers they fired went on to match or beat the ones hired to replace them. Even professionals run their patience to zero and fire at the worst moment.
And it reaches even the greatest. Picture Warren Buffett and Berkshire Hathaway at the height of the dot-com mania. In 1999, Buffett refused to chase technology stocks, and it looked like he had lost his touch; Berkshire fell more than 20% while the S&P 500 rose about 20%. The most celebrated investor alive had actually lost money while the market soared, and the press wrote him off, with Barron’s running “What’s Wrong, Warren?” on its cover. His edge was intact. But the client lives in the comparison, and years like that draw down the relative budget, the finite patience that decides how long a portfolio may look different before it is declared broken.
Nor was 1999 a fluke. Since 1980, a dollar in Berkshire grew to nearly $3,000 while a dollar in the S&P 500 grew to about $224.∗ Even so, Berkshire’s lead over the index peaked in October 2008, and nearly eighteen years later it still sits about 40% below that high. That is not the same as losing money, or even trailing the index in every one of those years; it means the lead already built was smaller than it had once been, which is the number a client anchors to. The 2008 stretch is the extreme but not the exception: across the full record, Berkshire spent roughly nine months in ten below its best relative high, and about three in four more than 10% below it. Any portfolio built to look different from the benchmark pays a price in stretches of looking wrong. The four pillars are designed to manage that price, limiting how far and how long the portfolio trails in the good years while still cushioning the drawdowns in the bad ones.
Overdraw either budget and the plan is over. Whether the client bolts because the portfolio fell too far or lagged too long, the compounding stops the moment faith is gone, no matter how sound the plan looked on paper.
This is the problem we believe can be solved. The endowment method we laid out in Part One is a deliberate attempt to defend both budgets at once, limiting the depth of the bad years to protect the mandate, and keeping the portfolio earning in the good years to conserve the client’s patience. Nothing here is watered down to track an index. The differences are engineered to pay for themselves often enough to keep faith intact. The four pillars split the work, two guarding each budget.
True Diversification
Spread the load across genuinely different sources of return (stocks, bonds, true diversifiers) so no single crisis dictates the depth of the worst year.
Dynamic Positioning
Adjust exposures as conditions shift, so the portfolio can lean away from risk when the environments that produce deep drawdowns take hold.
Factor Tilts
Layer deliberate, evidence-based return sources on top of the market’s, so the portfolio keeps competing while its defenses sit quiet.
Capital Efficiency
Let a single dollar do more than one job, so adding defense never requires abandoning the growth engine the benchmark runs on.
The allocator’s job is building something the client, the committee, and the board can hold long enough to compound, with losses inside the risk they agreed to carry and lags inside the patience they can actually supply. As we publish through the pillars, each will be graded against the two questions this piece posed. Does it soften the loss, and does it hold up on the relative scoreboard?
Every portfolio spends from two risk budgets at once. The absolute budget is written into the mandate: how far the account may fall. The relative budget is written into the client’s patience: how long it may look wrong. Indexing spends all of one to keep the other at zero; heavy diversification does the reverse. Even Warren Buffett, up nearly 3,000-fold, has seen his lead over the index sit about 40% below its 2008 high for nearly eighteen years, a relative gap rather than a loss of capital. The craft of the endowment portfolio is refusing to overdraw either.
Data & methodology. Performance figures are monthly total returns compiled by RQA from the following sources: global equity: MSCI ACWI Index from January 1988 forward, extended with the MSCI World Index (developed markets) for prior periods; U.S. aggregate bonds: Bloomberg U.S. Aggregate Bond Index from 1976 forward, extended with representative proxy data from Norgate Data for earlier periods; gold: Norgate Data; trend-following: an AQR managed-futures simulation over the early period, the SG Trend Index (a benchmark of trend-following CTAs) through April 2019, and the iMGP DBi Managed Futures Strategy ETF (DBMF) thereafter. DBMF is an actively managed fund that seeks to replicate the performance of the managed-futures (CTA) trend index; it is used for the most recent period to make the sleeve investable. Exhibit A’s extended history draws on Norgate Data for long-run equity and bond series. Crisis drawdowns are measured over the window shown; the 1973–74 figures run from the March 1973 start (about two months after the January 1973 market peak) and so understate the full peak-to-trough decline.
∗ Berkshire figures. Berkshire Hathaway Class A share-price total return (the stock pays no dividend) versus the S&P 500 total return, monthly, April 1980 – August 2026 (the span of available Berkshire history), from Norgate Data. One dollar grew to roughly $2,925 in Berkshire against about $224 in the S&P. “Below a prior high-water mark” means the Berkshire-to-S&P wealth ratio sitting below its own running maximum: the case in about 90% of months, and more than 10% below its peak in roughly 73%.
The recent exhibits span March 1973 through December 2025, the earliest date from which all four components of the diversified allocation, and the associated strategy backtests, are continuously available. Exhibit A’s extended view begins in January 1926 and reflects equities and bonds only, as investable gold and trend-following histories begin later. Where a component’s primary index does not span the full horizon, the series is extended with representative proxy data selected to preserve the return and risk characteristics of the asset class.
Component returns are total returns and are net of the underlying product’s fees and expenses where an investable product is used. The equity and bond sleeves are represented by indices throughout and therefore bear no product-level fees. The trend-following sleeve uses an investable product where one exists, and an index or simulation before that. All figures are gross of RQA’s advisory management fee and gross of transaction costs, custody fees, and taxes, any of which would reduce the returns shown.
All allocations are fixed-weight and rebalanced monthly. The allocations shown are hypothetical and illustrative, intended to demonstrate a portfolio-construction principle. They do not represent an RQA product, an account managed by RQA, or a recommendation. Indices are unmanaged and cannot be invested in directly. Hypothetical and model results carry inherent limitations: they are constructed with the benefit of hindsight, do not reflect the effect of material economic and market factors on decision-making, and do not represent the results of actual trading. Past performance is not indicative of future results.
Disclosures. The portfolios and exhibits shown are for illustrative purposes only. They are not real strategies or strategies offered by RQA, and any historical performance presented is hypothetical and simulated from historical data believed to be accurate. These materials have been prepared solely for informational and educational purposes and do not constitute investment advice or a recommendation to make or dispose of any investment or to engage in any particular investment strategy. They include general information and have not been tailored for any specific recipient. Information and data shown were obtained from sources believed to be reliable, but accuracy is not guaranteed. All investments involve risk, including the potential loss of principal. Past performance is not indicative of future results. References to individual securities, including Berkshire Hathaway, are for illustrative and educational purposes only, do not represent holdings of any RQA strategy, and are not a recommendation to buy or sell any security. Comparisons to indices or benchmarks are provided for illustrative purposes only and may differ in composition, risk profile, and rebalancing assumptions. Third-party research referenced (Goyal & Wahal, “The Selection and Termination of Investment Management Firms by Plan Sponsors,” Journal of Finance, 2008) is cited for educational purposes; the summary is RQA’s own. Richmond Quantitative Advisors, LLC is an SEC-registered investment adviser; registration does not imply a certain level of skill or training.