Quantise research note · risk-weighted portfolio construction
Encyclopedia of Return Streams
22 Return Streams, Organized by Economic Driver
How to Read Each Entry
Notation
Each entry answers three different questions:
How does the stream behave across different growth and inflation environments?
What happens during a major deleveraging or funding shock?
Does the stream usually earn steady gains with occasional large losses, or does it tend to lose gradually and then gain sharply during a crisis?
The Dial: Growth and Inflation
Each return stream includes a four-cell dial showing how it tends to perform under different combinations of growth and inflation.
Growth runs vertically, while inflation runs horizontally. The colour of each cell represents performance on a six-step scale, from great to bad. There are two important points to remember.
First, performance is measured relative to the stream's own long-term average, not relative to zero. A green cell therefore means that the stream tends to perform better than usual in that environment. It does not mean that returns are guaranteed to be positive.
Second, the bottom-right cell represents a severe deflationary bust involving defaults, financial stress and forced selling. It is not simply a mild slowdown in which inflation gradually declines. Some return streams behave very differently in these two situations.
When all four cells are pale, the stream has no strong growth or inflation bias. This is not missing information. In many cases, that lack of sensitivity is precisely the diversification benefit the stream is meant to provide.
Deleveraging: The Funding Shock
The deleveraging reading asks what happens when market behaviour is driven less by the economy and more by funding pressure. During a severe shock, investors may be forced to sell positions simply to raise cash. At that point, the growth and inflation dial becomes less useful. Assets that appeared unrelated can suddenly fall together because they are owned by the same leveraged investors. Each stream receives one of three readings:
Helps means the stream tends to gain while other investors are being forced to sell.
Holds means the stream does not usually benefit from the crisis, but it also does not participate meaningfully in the loss. This often applies to risks that are not primarily financial.
Hurts means the stream tends to fall alongside other assets during the funding shock. Most of the streams in this encyclopedia fall into this category.
Payoff Profile: Which Tail Is Larger?
The payoff profile describes how gains and losses are distributed over time. It is shown in the index rather than repeated in every entry. A short-volatility stream tends to produce small, regular gains followed by an occasional large loss. A long-volatility stream behaves in the opposite way. It may lose small amounts for an extended period and then make a large gain when markets break down. Only a genuinely balanced stream has a symmetrical payoff profile.
Under this definition, equities are classified as short volatility. Equity declines are generally faster and deeper than equity rallies. Realised volatility rises when markets fall, and leveraged equity positions are often liquidated in the same conditions that damage strategies selling options.
Cell colour: performance in that box, against the stream's own average
GreatGoodOKNeutralPoorBad
Worked exampleGold, B-12. Great in stagflation, good in a boom and in a bust, and poor in the one box where real rates rise with monetary confidence intact.
Read the other two columns with it. Gold is one of the four streams that gain in a funding shock, and its payoff profile is long vol, because it does nothing for years and then reprices quickly. Three readings, three different questions.
Deleveraging: behaviour in a funding shockhelpsGains while others are forced to sell.
4 streams: sovereign duration, gold, trend, cash.holdsNo gain, but not the same loss either.
4 streams, including inflation-linked bonds and the insurance cluster, where the risk is real but not financial.hurtsLoses alongside everything else.
14 streams, from equity beta through credit and carry to volatility selling. Most of the list sits here.
Payoff profile, in the index: which tail is fatterlong volFat right tail. A slow bleed most of the time, then a large gain when something breaks.
3 streams: commodities, gold and managed futures.neutralNeither tail is reliably fatter. Losses and gains scale with the move.
4 streams: sovereign duration, inflation-linked bonds, crypto, cash.short volFat left tail. Small regular gains, then a loss that arrives all at once.
15 streams. Equity beta belongs here alongside credit, carry, relative value, volatility supply and insurance, because its drawdowns are faster and deeper than its rallies.
Index
22 streams · 13 clusters
Plate
Dial
Return stream
Cluster
Deleveraging
Payoff profile
Cost band
BasicB-01 to B-14 · no manager, no prime broker, no institutional minimum
Cost in short form. bp = basis points per annum on a listed vehicle. 1.5–2 + 20 = 1.5–2% management plus 20% performance. A slash separates the cheap implementation from the full one. Complete bands, including hedging, borrow and financing, sit in each entry.
Environment coverage
Where the set is thin
The dial is more useful for spotting an empty box in the set you own than for ranking streams against each other. Below is every stream that rates great or good in each environment, plus the short list that genuinely gains when other people are being forced to sell. Look at how crowded the disinflationary growth column is next to the deflationary bust column. Read that imbalance before you read any individual rating, and note that the three streams covering the empty corner are all cheap and listed.
Boom
growth ↑ · inflation ↑ 10 of 22 rate great or good
Everything else either loses with the rest of the book or, at best, sits the episode out.
Long volatility
convex payoff · 3 of 22 15 are structurally short volatility
B-09Broad commodity futures
B-12Gold and monetary metals
B-13Managed futures / time-series trend
15 streams carry the fat left tail, spread across 8 clusters, from equity beta through credit and carry to volatility supply and insurance. They make small regular gains and then lose a great deal at once. Count them before you decide the book is diversified, and keep the question separate from deleveraging, since insurance premia are short volatility on risks that have nothing to do with markets.
Basic
B-01 – B-14
Every stream in this tier is available directly, with no manager to select, no prime broker, no subscription documents and no institutional minimum. Twelve of the fourteen are one or two exchange-listed instruments held in an ordinary brokerage account with daily liquidity. The other two, a building and a collection, anybody can buy directly, and both price on a horizon of years rather than seconds. That tier carries the large majority of the diversification available to any investor, so do not skip past it looking for something more sophisticated. Going from two bets to eight is worth several times what going from eight to thirty is worth, and it costs a small fraction as much.
The return comes from the policy rate and the yield available at the very front of the government curve. Cash earns the prevailing short-term rate, with almost no duration or credit exposure, and provides the benchmark against which every other return stream should be judged.
Realistic vehicle
A short-dated government bill ETF in your accounting currency, or bills bought directly at auction. Money market funds are fine provided you read what sits inside them, because some of them reach for yield in ways that defeat the purpose of holding cash.ExamplesUCITSIB01 for USD bills · XEON or CSH2 for EUR overnightUS-listedSGOV at 0 to 3 months · BIL at 1 to 3 monthsDirectA three-month bill ladder rolled quarterly, which removes the fee entirely at the cost of doing the adminWatchMatch the currency to your liabilities rather than to the highest yield on the screen. An unhedged foreign bill is a currency bet with a coupon attached.
Works well when
Cash works well when short-term interest rates are high or still rising, and when stress is pushing down the prices of longer-duration assets. Because bills mature quickly, their yield resets toward the prevailing policy rate while their price remains relatively stable. During a funding shock, liquidity becomes valuable in its own right: it allows the investor to meet obligations, avoid forced sales and buy assets from others who need cash immediately.
Fails when, and how
Cash performs poorly when its yield is below inflation, because the balance appears stable while its purchasing power declines. It also loses income quickly after central banks cut rates, since maturing bills must be reinvested at lower yields. Three-month bills bought in January 1981 at about 15% returned less than 8% a year over the following two decades as rates declined. Cash is therefore useful as liquidity and optionality, but a large permanent allocation can create a substantial long-run opportunity cost.
Cost band
3–12 bp p.a. bill ETF · a few bp at auction if bought directly
The return comes from the term premium and from changes in nominal government-bond yields. Longer-duration bonds gain when growth and inflation fall below what the yield curve had priced, because expected policy rates and usually the term premium decline.
Realistic vehicle
A currency-hedged government bond UCITS ETF, or futures (ZN, ZB, RX, G) for cleaner exposure with no manager and no tracking drift.ExamplesUCITSDTLA or IDTL for 20+ year Treasuries at roughly sixteen years of duration, with EUR- and GBP-hedged share classes of the same fundUS-listedTLT at ~16 years · IEF at ~7 · EDV at ~24FuturesZN, ZB, UB in Chicago · FGBL in Frankfurt · long gilt in LondonEquivalence10% in TLT carries about the same rate risk as 23% in IEF, so the two are not interchangeable at equal weights.
Works well when
Nominal sovereign bonds work well in a disinflationary slowdown or recession, particularly when central banks cut rates more aggressively than markets expected. Lower expected growth and inflation push government-bond yields down, and the longer duration converts that fall in yields into a capital gain. Safe-haven demand can reinforce the move, allowing the position to offset equity losses at relatively low cost.
Fails when, and how
They perform poorly when inflation surprises to the upside or investors demand a higher term premium for holding long-dated government debt. Yields then rise and bond prices fall; if equities are also being repriced because discount rates are rising, the usual diversification between stocks and bonds can disappear, as it did in 2022. Fiscal concerns can produce the same result when heavy issuance, inflation risk or weaker sovereign credibility pushes yields higher at the same time as risk assets decline.
Cost band
5–20 bp ETF; futures ≈ roll + collateral spread, a few bp
B-03
Inflation-linked sovereign bonds
Sovereign bondsDeleveraging: holds
Performs best
Stagflation· GREAT growth ↓ · inflation ↑
Underperforms
No losing box no environment costs it money
Primary driver
The return combines a real yield with compensation for realised inflation. Unlike a nominal bond, the principal and coupons are linked to an inflation index, but the market price still moves with real yields and with changes in breakeven inflation.
Realistic vehicle
A currency-hedged global inflation-linked UCITS ETF. A short-duration linker fund gives you the inflation accrual with much less real-rate duration if that is what you are after.ExamplesUCITSIGIL for global linkers hedged to USD · ITPS for US TIPSUS-listedTIP at ~6.5 years · SCHP at 3 bp · VTIP or STIP at ~2.5 years · LTPZ for the long endDurationThe choice of duration decides the outcome. Through 2022 short TIPS lost low single digits while long TIPS lost more than 30%, on identical inflation prints.
Works well when
Inflation-linked bonds work well when realised inflation is higher than the market had priced and real yields remain stable or fall. The inflation adjustment increases the bond's principal value and future coupons, while a stable real discount rate prevents that benefit from being offset by a fall in the bond's market price.
Fails when, and how
They perform poorly when real yields rise faster than inflation increases the bond's principal value. This is what happened in 2022: inflation-linked bonds received large inflation adjustments, but those gains were more than offset by the sharp rise in real yields, which pushed bond prices down. Longer-maturity linkers suffered the most because their prices are more sensitive to changes in real yields. Liquidity can also weaken during market stress, especially outside the large US and UK markets.
Cost band
10–25 bp p.a. hedged
B-04
Investment-grade credit spread, rate-hedged
CreditDeleveraging: hurts
Performs best
Goldilocks· GREAT growth ↑ · inflation ↓
Underperforms
Deflationary bust· BAD growth ↓ · inflation ↓
Primary driver
The return comes from the corporate credit spread after government-rate exposure has been removed. Investors collect compensation for expected defaults and downgrades, for uncertainty around those outcomes and for accepting weaker secondary-market liquidity.
Realistic vehicle
Long an investment-grade corporate ETF, short government futures of the same duration. The long leg gives you the spread and the rates; the short leg takes the rates back out, and what remains is the spread on its own. Index CDS (CDX IG, iTraxx Main) does the same thing more cleanly and more cheaply if you have the account, because it strips the rate leg by construction.ExamplesStructureLong corporate bonds, short government bonds of the same duration. What is left over is the spread, which is the thing you actually wanted.USD · longLQD in the US, or LQDE in UCITS form. Roughly 8.4 years of duration.USD · shortZN Treasury futures at about 6.5 years, so you need roughly 1.3 units of futures notional per unit of bond exposure to flatten the rate legEUR · longIEAC, roughly 4.5 years of durationEUR · shortFGBM Euro-Bobl futures at about 4.6 years, so close to 1.0 unit of notional, and the ratio differs from the USD one because the two credit markets are not the same lengthPre-hedgedIGHG runs the USD version pre-hedged for about 30 bpWatchHeld without the short leg, most of the variance in a corporate bond ETF is rates rather than spread.
Works well when
Rate-hedged investment-grade credit works well when economic growth is healthy, defaults and downgrades remain low, and companies can refinance without difficulty. In that environment investors require less compensation for corporate risk, so credit spreads tighten while the position continues to earn spread carry. Because the sovereign-rate exposure has been hedged, the return is driven mainly by the corporate credit cycle rather than by changes in government-bond yields.
Fails when, and how
It performs poorly when recession risk rises and investors expect more downgrades, defaults or difficulty refinancing. Credit spreads widen, reducing the value of the corporate bonds, and the loss often occurs alongside an equity drawdown because both markets are responding to weaker earnings and tighter financing conditions. The payoff is asymmetric: spread income and tightening provide limited upside, while a severe deflationary downturn can produce much larger losses through defaults and illiquidity. Without the rate hedge, the position also retains substantial sovereign-duration exposure and may look more diversified than it really is.
Cost band
10–20 bp long leg + futures roll; index CDS a few bp per turn
The return comes from corporate earnings growth exceeding what is already priced and from changes in the equity risk premium. Prices rise when expected cash flows improve, discount rates fall or investors accept less compensation for bearing equity risk.
Realistic vehicle
A global or developed-market UCITS ETF, either in a share class hedged to your accounting currency or hedged separately with rolling forwards. Index futures do the same job while tying up less capital, which helps when you want the cash working as collateral elsewhere.ExamplesUCITSIWDA or SWDA · SWRD at 12 bp · VEVE or VHVG for the FTSE version, with hedged share classes where the currency is not meant to be part of the betUS-listedVT at 6 bp for all-world · VTI plus VXUSFuturesES and MES when you want the cash working as collateralWatchMSCI World is around 70% United States, with roughly a quarter of it in the ten largest names.
Works well when
Developed-market equities work well when economic activity and corporate earnings grow faster than investors had expected. Stable or falling real rates reduce the discount applied to future profits, while greater confidence can lower the equity risk premium and raise valuation multiples. The strongest environment therefore combines improving cash-flow expectations with financing conditions that do not become materially tighter.
Fails when, and how
They perform poorly when expected earnings fall, recession risk increases or discount rates rise enough to compress valuations. Prices can then be hit from both directions: investors reduce their estimates of future cash flows and apply a lower multiple to those cash flows. At the portfolio level, the more important failure is concentration. At conventional capital weights, equities can contribute 80 to 90% of total portfolio variance, leaving the other holdings with little ability to change the overall outcome.
The return comes from faster economic and earnings growth than in developed markets, a higher equity risk premium, commodity terms of trade and access to dollar funding. The balance between these forces varies by country, but global liquidity is often the common constraint.
Realistic vehicle
A broad EM UCITS ETF. Use an ex-China version if you want to isolate the China policy factor, or an EM minimum-volatility version if the concentration described below bothers you.ExamplesUCITSEIMI at 18 bp, small caps includedUS-listedIEMG or VWO · EMXC ex-China · EEMV for minimum volatilityWatchTaiwan and Korea are together about a third of the index and TSMC alone is close to a tenth, so removing China raises the semiconductor weight rather than reducing risk.
Works well when
Emerging-market equities work well when global trade and industrial activity are accelerating, the US dollar is stable or weakening, and local inflation is falling enough to give EM central banks room to ease. Easier dollar funding supports capital flows and currencies, while lower domestic rates improve financing conditions and equity valuations. Commodity-exporting markets receive an additional benefit when stronger export prices improve national income and corporate earnings.
Fails when, and how
They perform poorly when the dollar strengthens sharply, global funding tightens or investors reduce risk across countries at the same time. Capital outflows weaken local currencies and tighten domestic financial conditions, while falling global demand reduces earnings; these forces can reinforce one another. Correlations with developed-market equities usually rise during the stress, so the diversification expected from the allocation becomes weakest when it is most needed. Broad EM indices can also hide a large concentration in Taiwan and Korean semiconductor companies, and mechanically removing China increases that concentration rather than removing it.
Cost band
14–35 bp p.a. + 30–60 bp unrecoverable dividend withholding
B-07
Emerging-market local-currency debt
Emerging marketsDeleveraging: hurts
Performs best
Goldilocks· GREAT growth ↑ · inflation ↓
Underperforms
Stagflation· BAD growth ↓ · inflation ↑
Primary driver
The return has three components: the local real yield, changes in domestic bond yields and movements in the issuer's currency. Investors are paid for holding markets with different policy cycles and for the possibility that undervalued currencies converge toward long-run fair value.
Realistic vehicle
An EM local-debt UCITS ETF, unhedged, because the currency is the stream. Add a partial hedge overlay only if the volatility budget cannot carry the full currency exposure.ExamplesUCITSIEML or SEMLUS-listedEMLC at 30 bpHedgingNone. The currency is the stream, and hedging it back leaves you with local rates and little else.Not thisEMB and IEMB are USD sovereign spread, which is a credit exposure belonging to the credit cluster
Works well when
Emerging-market local-currency debt works well when credible central banks have already raised rates, local inflation is falling and the US dollar is stable or weakening. Bondholders can then earn a high starting yield while falling local rates support bond prices and a firmer local currency adds to the return for foreign investors. Improving commodity terms of trade can strengthen both public finances and currencies in the large exporter countries represented in the index.
Fails when, and how
It performs poorly during a dollar-liquidity shock. Foreign investors sell local bonds, EM currencies weaken together and local yields may rise, so the bond-price loss and the currency loss occur at the same time. The foreign-exchange move can easily overwhelm several years of coupon income. This creates a negatively skewed payoff: long periods of steady carry can be reversed within weeks, often during the same periods in which global equities are also falling.
Cost band
25–55 bp p.a.; wide underlying bid-offer, real tracking error in stress
The dominant driver is global liquidity, expressed through risk appetite, real rates and the availability of leveraged dollar funding. A separate adoption cycle can add an independent source of demand, but over shorter horizons crypto assets often behave like extremely long-duration risk assets.
Realistic vehicle
A physically backed spot ETP from a regulated issuer, with segregated cold storage and a published custody policy. Regulated futures if you want the exposure to sit lightly on capital. Self-custody only if key management is a discipline you already run properly.ExamplesEuropePhysically backed ETPs from 21Shares, WisdomTree, CoinShares and VanEck, with segregated cold storageUS-listedIBIT or FBTC for bitcoin · ETHA or FETH for etherFuturesCME bitcoin and micro bitcoinEquivalenceAt roughly 70% annualised volatility, a 2% position contributes about as much risk as a 9% position in developed equity.
Works well when
Crypto assets work well when dollar liquidity is abundant, real rates are falling and investors are willing to take more risk. Easier funding supports leverage and speculative demand, while lower real yields reduce the opportunity cost of holding an asset with no contractual cash flow. Adoption can provide an additional source of return when broader access, usage or institutional participation expands during an already supportive market environment.
Fails when, and how
They perform poorly when funding tightens and leveraged investors are forced to reduce positions. Because the market is highly sensitive to marginal liquidity, the resulting drawdowns can be much larger and faster than in equities, and correlation with the Nasdaq often rises during the sell-off itself. Custody failures, exchange failures and regulatory changes create additional losses that are separate from the market cycle and are not guaranteed to earn a compensating premium. Crypto can therefore be used as a satellite growth exposure, but it should not be expected to protect the portfolio during a funding crisis.
Cost band
35–95 bp p.a. spot ETP, plus tracking and spread on entry and exit
The return comes from changes in physical commodity prices and from the futures curve. Supply and inflation shocks drive the spot component, while roll yield and producer hedging pressure determine how much of that move reaches an investor holding futures.
Realistic vehicle
A broad commodity UCITS ETF with an optimised roll. Roll methodology matters far more than the fee, because front-month indices bleed structurally when curves sit in contango.ExamplesUCITSCOMF for a longer-dated roll · ICOM or CMOD at about 19 bpUS-listedPDBC for an actively optimised roll · BCI at 25 bp with no K-1 · COMT for dynamic rollNot thisFront-month single-commodity notes, for broad exposureWatchAcross a cycle the gap between an optimised roll and a front-month one has run to several hundred basis points a year in contango, which dwarfs every fee on this page.
Works well when
Broad commodity futures work well when physical supply is constrained while demand remains resilient. Production outages, transport disruption or geopolitical restrictions can lift spot prices and move futures curves toward backwardation, allowing the investor to benefit from both rising prices and a more favourable roll. Because the same supply shock can raise inflation while hurting equities and nominal bonds, commodities can diversify both sides of a conventional stock-and-bond portfolio.
Fails when, and how
They perform poorly during prolonged contango, because each expiring contract must be replaced with a more expensive later-dated contract and the repeated roll creates a persistent drag. A demand-led recession can also pull commodity prices down as consumption and industrial activity weaken. During an acute funding shock, investors may sell futures for cash regardless of the inflation outlook, so the position can fall with other risk assets. A broad index further combines many unrelated physical markets, which means strength in one commodity can be diluted by weakness elsewhere.
Cost band
19–45 bp p.a.; roll structure worth several hundred bp over a cycle
B-10
Direct real estate, income-producing
Real assetsDeleveraging: hurts
Performs best
Goldilocks· GREAT growth ↑ · inflation ↓
Underperforms
Deflationary bust· BAD growth ↓ · inflation ↓
Primary driver
The return comes from rental income, changes in property values and the spread between the asset yield and its financing cost. Inflation-linked leases can raise nominal income over time, while leverage amplifies both the income spread and changes in valuation.
Realistic vehicle
An apartment, a small building, or a share in one, bought with a mortgage and let. A core open-ended fund does the same job at size, provided its redemption queue has been tested in a real drawdown. Listed REITs own the same kind of asset and give you the same rent economics far more cheaply, so the question is not whether they are property. It is that three things genuinely differ: what sits in the index, how the leverage is financed, and who sets the price at the margin.ExamplesDirectAn apartment or small building at 50 to 60% loan to value, financed long and fixed, with rent covering debt service at a rate two points above the one you signedAt sizeA core open-ended fund whose redemption queue has been tested in a real drawdownListedVNQ, IPRP and IWDP hold hundreds of properties for a few basis points, against 3 to 7% of round-trip cost for a single building. Correlation to equities runs 0.7 to 0.8, and they fell roughly 37% in 2008 and 40% between February and March 2020.Same assetOver long holding periods listed and private property deliver close to the same thing. Once private returns are de-smoothed, the two correlate somewhere around 0.7 to 0.9 across multi-year windows. Anyone claiming a REIT is not real estate is arguing with the rent roll.Different clockAppraisals update on stale comparables, so private indices report roughly half their true volatility and lag by several quarters. Through 2022 listed REITs fell about 25% while the main US private index still printed a positive year, then caught down through 2023. Most of the reported diversification is that lag rather than a different exposure.What differsIndex composition, financing and the marginal price-setter. Listed indices carry heavy weights in towers, data centres, logistics and self-storage, which is not what you own when you buy a flat. Ten-year fixed debt at 55% loan to value cannot be margin-called, while a REIT reprices its leverage daily through equity investors who are sometimes selling for reasons that have nothing to do with the rent.
Works well when
Income-producing real estate works well when rents are rising, local supply is constrained and debt has been fixed at a cost below the property's income yield. Rental income can then grow while financing costs remain stable, increasing the cash flow available to the owner. Stable or falling real rates can also reduce capitalisation rates and raise the value placed on those rents, while indexed leases help income keep pace during an inflationary expansion.
Fails when, and how
It performs poorly when real rates and capitalisation rates rise, because investors apply a higher discount rate to future rents and the property's value falls. Leverage magnifies that decline. A deflationary bust is more dangerous because rents and occupancy can weaken while the debt balance and interest payments remain fixed, creating a path to negative equity. Appraisal-based valuations often recognise the loss only gradually and therefore understate the true volatility, while high entry and exit costs require a long holding period before the investment earns them back.
Cost band
60–120 bp core open-ended fund · 1.5% + 20% value-add · 3–7% round-trip transaction cost, which usually dominates the fee
B-11
Collectibles: art, watches, wine, classic cars
Real assetsDeleveraging: hurts
Performs best
Goldilocks· GREAT growth ↑ · inflation ↓
Underperforms
Deflationary bust· BAD growth ↓ · inflation ↓
Primary driver
The return comes from scarcity interacting with the wealth, liquidity and preferences of a small buyer base. Supply of a specific artist, vintage or provenance cannot expand, so prices depend heavily on how much capital affluent collectors have and what they currently regard as desirable. Some objects also provide a portable store of value without a financial counterparty.
Realistic vehicle
Direct purchase at auction or through a dealer, held in insured storage or a free port where that makes sense. Fractional platforms and art funds exist, but they stack a fee layer onto an asset that already carries the widest spread on this page. Buy things you would be content to keep, because the exit is never guaranteed.ExamplesDirectPurchase at auction or through a dealer, held in insured storage, in a category you can read yourselfReferenceLiv-ex for wine · HAGI for classic cars · Art Market Research and Knight Frank for art and luxuryNot thisFractional platforms and art funds, which add a management layer to an asset already losing 10 to 25% each way at the auction house
Works well when
Collectibles work well when the wealth and confidence of their natural buyers are rising and demand is concentrated on genuinely scarce objects. Because supply cannot respond to a higher price, additional demand is reflected mainly in the sale price rather than in greater production. Concern about currency debasement can add support for portable real assets, provided authenticity, condition and provenance are trusted.
Fails when, and how
They perform poorly when buyers withdraw, because an illiquid object may have no executable bid near its last recorded sale. The asset produces no income to offset storage, insurance, conservation and authentication costs, while auction and dealer spreads can reach 10 to 25% in each direction. Published indices tend to include the objects that successfully traded and omit those that could not find a buyer, which makes measured volatility and correlation look artificially low. A permanent change in taste can impair an entire category, while failed provenance or authenticity can turn an investment loss into a total loss.
Cost band
10–25% round-trip auction and dealer spread · 0.5–1.5% p.a. storage, insurance and conservation
B-12
Gold and monetary metals
Precious metalsDeleveraging: helps
Performs best
Stagflation· GREAT growth ↓ · inflation ↑
Underperforms
Goldilocks· POOR growth ↑ · inflation ↓
Primary driver
The return is driven mainly by real-rate expectations, confidence in currencies and sovereign credit, and official-sector demand for reserve diversification. Gold is a monetary asset rather than an industrial commodity, so its economic behaviour should be assessed separately from broad commodity exposure.
Realistic vehicle
A physically backed ETC with allocated bars, a published bar list, and vaulting in a jurisdiction you are content to hold metal in. No manager, no roll, no counterparty games.ExamplesUCITS ETCsSGLN or SGLD at 12 bp, physically backed and allocatedDeliverableXetra-Gold, and the Swiss-vaulted ZKB and UBS products, where the option to take metal mattersUS-listedGLDM at 10 bp · IAU at 25 · SGOL for Swiss vaults · GLD for the deepest options marketFuturesGC and micro MGCMarginal benefitAdding gold to a conventional stock and bond portfolio in 4% steps improves return per unit of risk sharply at first, and the improvement flattens out beyond roughly a 16% weight.
Works well when
Gold works well when real yields are falling, confidence in money or sovereign credit is weakening, or official institutions are diversifying away from traditional reserve assets. Lower real yields reduce the opportunity cost of holding an asset that pays no income, while monetary or geopolitical uncertainty increases demand for an asset without a direct counterparty. The repricing often arrives quickly after a long quiet period, which gives gold its long-volatility character.
Fails when, and how
It performs poorly when real yields rise and confidence in the monetary system remains intact. Higher real yields make interest-bearing assets more attractive, while gold has no cash flow to cushion the decline or reward the investor for waiting. Gold moved largely sideways through much of the 1980s and 1990s and lost roughly a third of its value in 2013. The practical risk is therefore not only a sharp drawdown, but also a decade of weak returns that exceeds the patience of the investor or the governance structure holding it.
Cost band
12–25 bp p.a. all-in for allocated, vaulted physical
The return comes from persistent price trends. Information is incorporated gradually, investors initially under-react and institutional capital adjusts slowly; once a move is established, extrapolation and continuing flows can extend it.
Realistic vehicle
A replication ETF or UCITS for cheap index-like exposure, a flagship CTA for the real thing, or your own book across 20 to 40 liquid futures with a volatility target. Running it yourself is decisively the right answer if you already have futures infrastructure, clean tick data and someone who owns the roll calendar at three in the morning, because the fee saving compounds into a very large number over a decade.ExamplesUS-listedDBMF at about 85 bp, which replicates the average large CTA · KMLM and CTA run their own systemsEuropeDaily-dealing alternative UCITS from a trend house such as Man AHL, Aspect, Winton or Transtrend. The ETF market here is thin.Own book20 to 40 liquid futures across equities, bonds, currencies and commodities, one or two moving-average crossovers, positions scaled inversely to volatility, the book targeted at 10 to 15% annualised
Works well when
Time-series trend works well when prices continue moving in the same direction for long enough for the strategy to identify the move, build a position and remain invested. The direction itself does not matter: the strategy can be long a rising market or short a falling one across equities, bonds, currencies and commodities. Slow-building inflationary or deflationary regimes are especially favourable because they create persistent moves rather than one-day shocks.
Fails when, and how
It performs poorly in choppy, mean-reverting markets, where prices repeatedly begin a move and then reverse before a trend develops. The strategy is forced to enter and exit several times, producing a sequence of small losses known as whipsaw. A sudden reversal after a long trend can be more damaging because the position is largest just before the direction changes. Drawdowns can also last for years, creating a governance risk that investors abandon the strategy shortly before the next sustained trend begins.
Cost band
60–95 bp replication · 1–2% + 20% flagship CTA · infrastructure only if built
The return comes from systematic differences in expected returns across stocks after broad market beta has been removed. Value, quality and profitability, momentum and low volatility are linked to behavioural biases, institutional constraints and the compensation required for holding unfashionable or difficult positions.
Realistic vehicle
The crude two-line version is long a factor ETF and short the market ETF or an index future to strip out the beta. The real version is a sector- and country-neutral market-neutral mandate, because the crude version leaves residual bets that swamp the factor you wanted.ExamplesUCITSIWVL for value · IWMO for momentum · IWQU for qualityUS-listedVLUE · MTUM · QUAL · USMVShort legIndex futures sized at the factor's measured beta rather than at oneHedge ratioMTUM and VLUE run close to 1.0 · USMV nearer 0.7. Hedging USMV at 1.0 leaves you structurally short the market rather than long low volatility.WatchThe honest version is a sector- and country-neutral mandate. The two-line proxy leaves sector bets that routinely swamp the factor.
Works well when
The component equity-style premia work in different environments. Value benefits when unusually cheap companies or markets reconnect with their fundamentals; momentum benefits when existing winners and losers continue in the same direction; and low volatility benefits when leverage-constrained investors overpay for high-beta, lottery-like exposure. The portfolio is most robust when several factors are combined, because no single style works consistently through every cycle.
Fails when, and how
The strategy performs poorly when a factor becomes crowded, reverses sharply or remains out of favour for many years. Momentum crashes in 2009 and 2020, and value's weakness from 2007 to 2020, show that the holding period required can be much longer than most investors expect. Borrow, financing and beta-hedging costs reduce the premium that reaches the investor. Simple factor ETFs can also retain large sector or country exposures, turning what appears to be a market-neutral factor allocation into an unintended directional bet.
Cost band
20–50 bp long leg + 30–60 bp borrow/financing · 1–1.5% + perf for a real mandate
Specialized
S-01 – S-08
Every stream in this tier needs a manager, a prime broker, leverage, physical infrastructure or private documentation. Each entry states the access requirement honestly. Most of this tier is out of reach below roughly USD 100 m of dedicated risk capital, and several entries are only worth holding if the access is genuinely good. A mediocre allocation to a specialized stream is worse than no allocation at all, because you end up paying alternative fees for beta you already own.
S-01
Credit arbitrage and event-driven
CreditDeleveraging: hurts
Performs best
Goldilocks· GREAT growth ↑ · inflation ↓
Underperforms
Deflationary bust· BAD growth ↓ · inflation ↓
Primary driver
The common driver is compensation for bearing corporate credit and event risk inside a hedged structure. Merger arbitrage earns the deal-completion premium; capital-structure arbitrage trades inconsistent pricing across a company's securities; convertible arbitrage combines an issuer concession with a relatively cheap embedded option; and synthetic risk transfer earns a premium for absorbing junior losses and releasing bank regulatory capital. The techniques differ, but the underlying risk remains corporate credit.
Realistic vehicle
Sub-types: merger arbitrage and event-driven, capital-structure arbitrage, convertible arbitrage, synthetic risk transfer. A UCITS merger-arb fund for the liquid version, a credit-focused hedge fund or a multi-strategy platform for the arbitrage books, and negotiated private transactions with multi-year lockups for risk transfer. All of it needs CDS documentation, reliable borrow and genuine modelling capability. On the convertible side the quality of the prime broker is the actual investment decision, because financing and borrow terms determine whether the strategy works at all.
Works well when
Credit-arbitrage and event-driven strategies work well when growth is positive, defaults remain low and financing markets are orderly. Merger arbitrage earns its spread when announced transactions close; capital-structure trades profit when inconsistent prices across a company's securities converge; convertible arbitrage benefits when the embedded option is bought cheaply and financing remains available; and synthetic risk transfer earns its premium when realised loan losses stay below the level priced into the transaction. An active corporate calendar creates more opportunities, but the common requirement is that credit conditions remain stable enough for the hedge or catalyst to work.
Fails when, and how
They perform poorly when recession, default risk or a funding shock overwhelms the hedge. Deals can break together, a jump to default can destroy the assumed relationship between equity and credit, and losses on transferred loan books arrive in the same downturn that is hurting the rest of the portfolio. Convertible arbitrage is especially vulnerable when financing disappears: in 2008 the strategy lost roughly a third within weeks even though the underlying bonds had not fundamentally changed. Fees, limited capacity and political review can further reduce the realised premium. The entire group should share one risk budget with corporate and private credit rather than receive four separate allocations.
Cost band
0.9–1.5% UCITS merger arb · 1.5–2% + 20% for capital-structure, convertible and risk-transfer mandates, plus the financing spread that decides the outcome
S-02
Commodity curve carry and calendar spreads
CommoditiesDeleveraging: holds
Performs best
Boom· GREAT growth ↑ · inflation ↑
Underperforms
Deflationary bust· POOR growth ↓ · inflation ↓
Primary driver
The return comes from the shape of the commodity futures curve, which reflects physical inventories, storage costs, financing and convenience yield. Calendar spreads isolate changes in that curve shape rather than taking a simple long position in the commodity price.
Realistic vehicle
Sub-types: enhanced and optimised roll indices, dedicated calendar spreads, storage and inventory trades where the physical leg is reachable. An enhanced-roll commodity index for the beta-plus version, or a dedicated calendar-spread book for the pure version. Spread margins are low, so the capital efficiency is excellent and the risk is easy to underestimate.
Works well when
Commodity curve carry and calendar spreads work well when inventories are tight, near-term supply is scarce and futures curves are backwardated. Producers or commercial users may also create persistent hedging pressure that leaves one part of the curve expensive relative to another. The strategy earns a return when that pricing difference persists or converges in the expected direction, without requiring a large move in the outright commodity price.
Fails when, and how
They perform poorly when the physical mechanics of the market change faster than the model can adapt. Storage limits, delivery constraints, a sudden inventory shift or a crowded position can move one futures contract sharply relative to another, producing a large loss even if the long-run supply thesis remains reasonable. The negative WTI price in April 2020 showed that settlement and storage conditions can dominate normal financial relationships without warning. Because the trade is often leveraged, a temporary spread move can force liquidation before convergence occurs.
Cost band
30–50 bp enhanced index · 1–2% + perf managed book
S-03
Levered basis and funding relative value
Relative valueDeleveraging: hurts
Performs best
Goldilocks· OK growth ↑ · inflation ↓
Underperforms
Deflationary bust· BAD growth ↓ · inflation ↓
Primary driver
The return comes from supplying scarce balance sheet to close small pricing gaps across rates, currencies and mortgage markets. These include slope and curvature dislocations, swap spreads, futures basis, cross-currency basis and compensation for carrying mortgage prepayment risk and negative convexity. The common economic function is to take the other side of investors whose balance-sheet or regulatory constraints prevent them from arbitraging the spread themselves.
Realistic vehicle
Sub-types: rates and macro relative value across slope, curvature, swap spreads and futures basis; cross-currency basis; agency MBS prepayment and convexity. Realistically only inside a multi-strategy platform, a dedicated macro relative-value manager or a bank-adjacent mandate. It needs repo lines, a prime broker and ten to twenty times leverage on very tight spreads. An agency MBS ETF gives you the unlevered version of one leg, which is mostly duration plus a modest option-adjusted spread. For the basis trades there is no standalone product, and there should not be.
Works well when
Levered basis and funding-relative-value strategies work well when a pricing dislocation is visible but financing remains reliable. Policy divergence, reporting-date balance-sheet pressure or predictable flows from liability-driven and mortgage investors can create small gaps between closely related instruments. Stable rate volatility and prepayments that behave near the model allow those gaps to converge repeatedly, while leverage turns a small spread into a meaningful return. The high normal-period Sharpe ratio therefore reflects stable funding as much as it reflects forecasting skill.
Fails when, and how
They perform poorly when funding is withdrawn at the same time that the pricing basis widens. Margin calls and tighter financing terms then force investors to sell the cheap asset and close the hedge at a loss, turning a convergence trade into a liquidation trade, as in September 2019 and March 2020. A volatility spike can also damage short-convexity positions and extend duration precisely when the portfolio is least able to carry it. Fair value provides no protection if the position cannot be financed until convergence, which is why the strategy requires institutional scale and the ability to survive infrequent but severe funding losses.
Cost band
5–20 bp for the unlevered MBS leg · 1.5–2% + 20% for a levered basis mandate, where haircut and financing terms matter more than the fee
S-04
Short-horizon flow and statistical arbitrage
Relative valueDeleveraging: hurts
Performs best
Goldilocks· GOOD growth ↑ · inflation ↓
Underperforms
Deflationary bust· BAD growth ↓ · inflation ↓
Primary driver
The return comes from providing immediacy and absorbing temporary supply-demand imbalances at very high turnover. Index rebalances, forced trades, ETF-versus-basket deviations, futures rolls and dealer inventory create short-lived price pressure; statistical arbitrage applies the same principle to temporary mispricing between related securities.
Realistic vehicle
Sub-types: liquidity provision and short-term reversal, index-rebalance and ETF-versus-basket flow, statistical arbitrage proper. Only through a manager, and the good books are closed. It needs execution infrastructure and latency that a family office does not have and should not try to build. Capacity is small and jealously defended, so if it is open to you, ask why.
Works well when
Short-horizon flow and statistical arbitrage work well when volatility and cross-sectional dispersion are elevated but market liquidity is still functioning. Predictable institutional flows and temporary order imbalances create small price deviations, while wider bid-offer spreads provide enough gross return to cover execution costs. The strategy makes money by absorbing those temporary flows and closing the position when prices normalise.
Fails when, and how
It performs poorly when transaction costs, slippage and signal decay consume the small gross edge. Losses become much larger when the order flow is informed or forced rather than temporary, because the price continues moving instead of reverting while market makers withdraw liquidity. Crowded managers may also hold similar signals built from the same data, so simultaneous deleveraging turns apparently independent books into the same trade. August 2007 showed how quickly that common unwind can destroy diversification.
Cost band
2% + 20–30%; capacity-constrained, frequently closed, and worth paying only if the access is real
The return comes from applying value, carry, momentum and defensive styles across countries, currencies, bond markets and commodities. Value seeks convergence toward fundamental anchors; carry earns the yield differential for holding an unpopular exposure; momentum follows persistent flows; and defensive strategies exploit investors' preference for high-beta, lottery-like assets. The diversification comes from combining the styles, not from carry alone.
Realistic vehicle
Sub-types: FX style premia across carry, purchasing-power value and currency momentum; cross-asset value across countries, bond markets and commodities; multi-asset defensive, also sold as betting against beta. A multi-asset style-premia UCITS or a managed account, which is one product rather than the three or four separate line items the industry likes to sell. Parts of it are crudely expressible with country ETF pairs, bond futures and an FX forward book if you accept the tracking noise and the tax friction. The cross-sectional equity version of the same idea sits in the Basic tier at B-14 and costs a fraction as much.
Works well when
Multi-asset style premia work well when markets offer persistent differences in valuation, yield, trend or defensiveness and funding remains stable enough to hold the positions. Value benefits when dislocated prices move back toward fundamental anchors; carry benefits when yield differences persist without a funding shock; momentum benefits when trends continue; and defensive strategies benefit when leverage constraints leave safer assets under-owned. Combining the styles is important because each earns its return in a different part of the cycle.
Fails when, and how
The portfolio performs poorly when several styles lose at the same time or when one style dominates the risk budget during its own long drawdown. Carry can unwind violently during a funding shock, value can remain cheap for years without converging, and momentum can be hit by a sudden reversal. Financing and trading costs reduce the relatively modest gross premia, while heavy demand for quality or minimum-volatility assets can make the defensive leg expensive. The styles are therefore more defensible as a balanced portfolio; evaluated separately, they are often abandoned near the worst point in their cycle.
Cost band
0.7–1.5% in a multi-style UCITS · 1% + 10–20% dedicated, plus the financing spread on the levered legs
S-06
Closed-end fund and holding-company discounts
Relative valueDeleveraging: hurts
Performs best
Goldilocks· GOOD growth ↑ · inflation ↓
Underperforms
Deflationary bust· BAD growth ↓ · inflation ↓
Primary driver
The return comes from the discount between the market price of a listed wrapper and the value of its underlying assets. The discount is realised only if a catalyst such as activism, a tender offer, a buyback, corporate simplification or a wind-up causes the market price to converge toward net asset value.
Realistic vehicle
Sub-types: closed-end fund discounts, listed holding-company and conglomerate discounts, activist situations with a discount-control mechanism. Direct if you can hedge the underlying, which means long the holding company and short the listed stakes. Otherwise a specialist activist or value manager with the standing to force the catalyst.
Works well when
Closed-end-fund and holding-company discount trades work well when the starting discount is wide and there is a credible, time-bounded mechanism for narrowing it. A tender offer, asset sale, liquidation, improved capital allocation, active discount control or successful shareholder pressure can convert the apparent valuation gap into cash or a higher market price. Without such a catalyst, the discount is only a number on a screen rather than a realisable return.
Fails when, and how
They perform poorly when the discount has no enforceable catalyst and can remain wide indefinitely. During market stress, the discount may widen further just as the cost of hedging rises and borrow on the underlying assets becomes difficult or is recalled. Entrenched family, foundation or management control can make convergence structurally impossible. In that case the discount is not temporary mispricing; it is compensation for a permanent governance limitation.
Cost band
1–1.5% + 15–20%; borrow cost on the hedge leg often decides it
S-07
Volatility risk premium and dispersion
VolatilityDeleveraging: hurts
Performs best
Goldilocks· GREAT growth ↑ · inflation ↓
Underperforms
Deflationary bust· BAD growth ↓ · inflation ↓
Primary driver
The return is an insurance premium. Implied volatility usually exceeds the volatility subsequently realised because investors pay for protection and certainty. Dispersion earns a related premium from the gap between expensive index options and options on the index constituents, which reflects one-sided demand for index hedges and a difference between implied and realised correlation.
Realistic vehicle
Sub-types: index and cross-asset variance selling, put and strangle writing, correlation and dispersion, skew and term-structure trades. A risk-controlled variance programme spread across equities, rates, FX and commodities, with hard tail limits and explicitly purchased crisis protection, or an options desk running dispersion. Dispersion only counts as separate from generic volatility selling once you have neutralised net variance, direction, skew, sector concentration and earnings-event exposure, all five of them. Buy-write and put-write ETFs are the single-ticker proxy and are strictly inferior, being long equity beta with the upside removed.
Works well when
The volatility risk premium works well when option prices imply more volatility than the market subsequently realises. The seller keeps the difference as long as price moves remain contained and no large jump occurs. Dispersion adds a second source of return when index options imply very high correlation between constituents but individual stocks continue to move more independently, provided the single-stock options are liquid enough to trade efficiently.
Fails when, and how
It performs poorly during a systemic shock because the strategy is short several risks at once: volatility, jumps, convexity, liquidity and, in dispersion, correlation. As markets fall, realised volatility rises, correlations move toward one and option hedges become more expensive, so losses accelerate rather than remain linear, as in February 2018 and March 2020. Trading frictions absorb part of the theoretical premium even in normal markets. This is not a crisis diversifier: the steady premium is earned precisely because the strategy absorbs losses when a severe event occurs.
Cost band
1–2% + 20% for a programme · 35–60 bp for the ETF proxy you should not use as a diversifier
S-08
Insurance and other non-economic risk
InsuranceDeleveraging: holds
Environmental bias
None return does not depend on the growth or inflation box
Primary driver
The return comes from underwriting risks that are largely independent of the macroeconomy: natural catastrophe, mortality and longevity, weather, cyber loss, litigation outcomes and physical power or grid-delivery risk. Pricing is set by the probability and severity of claims and by available underwriting capacity rather than by expected growth or inflation. Litigation funding is structurally closer to buying an option, but the group as a whole is best understood as insurance risk.
Realistic vehicle
Sub-types: catastrophe bonds and reinsurance sidecars, mortality and longevity including life settlements, parametric weather, cyber, litigation finance, and power, gas and grid congestion. Catastrophe bond UCITS funds with weekly to monthly liquidity, reinsurance sidecars for the illiquid version, specialist ILS and life-settlement managers, parametric structures placed through reinsurance intermediaries, closed-ended litigation funds with lives of four to eight years, and a specialist commodity manager with a physical desk for power and gas. Zurich and Bermuda are where the underwriting expertise actually sits. Every route requires underwriting diligence that you either perform yourself or buy.
Works well when
Insurance and other non-economic risk strategies work well when premiums are set above expected claims and underwriting capacity is scarce. Hard markets after large losses, higher attachment points and diversification across perils, jurisdictions and counterparties can improve the expected return. The portfolio benefit comes from the source of the claims: hurricanes, legal outcomes or power disruptions are generated by mechanisms that are largely different from the business and interest-rate cycle.
Fails when, and how
They perform poorly when claims are larger or more correlated than the pricing model assumed. Short histories and changing conditions make frequencies difficult to estimate; contract wording can determine whether a loss is covered; and cyber, cloud or infrastructure exposures can create many claims from one event. Litigation outcomes are binary and slow to resolve, while power strategies add counterparty and regulatory risk, as Texas demonstrated in February 2021. Climate change can also make historical event frequencies unreliable. The underlying risk may be independent of financial markets, but the securities can still be sold for cash during a broad funding shock, so their weekly market prices may fall with other assets.
Cost band
100–150 bp cat bond UCITS · 1.5–2% + 15–30% for sidecars, private ILS, life settlements, litigation and power, plus diligence cost you should book as real
Excluded, and why
Partial credit only
The candidates below can be excellent investments. None of them is an independent return stream. Giving each one a separate risk budget is how a portfolio ends up with forty holdings and two and a half bets.
Candidate
What it actually reduces to
Private equity
Levered small-cap value plus illiquidity, with the volatility suppressed by appraisal-based valuation. Smoothing is not diversification.
Venture capital
Long-duration growth equity plus illiquidity, plus a power-law selection problem that ten commitments cannot diversify away.
Private credit
Credit spread, leverage and illiquidity. It is exposed to the same default cycle as everything else in the credit cluster.
High yield, leveraged loans, CLO equity
Credit beta with a short position in default correlation. It is equity risk with a coupon attached.
Listed real estate and infrastructure equity
Over a quarter it trades with the equity market; over a decade it tracks the buildings. Either way it is the same economic exposure as B-10 rather than a second one, so it earns a place in the portfolio but not a separate risk budget. In regulated infrastructure the inflation linkage is often negotiated away as well.
Distressed debt
Credit beta plus legal complexity plus liquidity risk. Real manager skill exists here, but the underlying risk factor does not change.
Alternative and consumer lending
Consumer or SME credit plus servicing risk plus illiquidity. The first loss arrives with unemployment.
Royalties, music and pharma
Long-duration contractual cash flows with sector-specific uncertainty. Macro beta is low, but what you are being paid for is duration rather than a risk premium.
Put writing and covered calls
Equity beta with the upside removed. It belongs inside the equity sleeve, sized in delta terms, rather than counted as an alternative.
Multi-asset liquid-alt wrappers
Once you measure it, correlation to equity is typically 0.7 to 0.85, and you are paying alternative fees for it. Test any candidate against a naive blend of equity and cash first, because most do not survive that test.
Important information
This note is published for information and discussion only. It is not investment advice, an offer, or a solicitation to buy or sell any instrument, and it takes no account of the objectives, circumstances or constraints of any particular investor. Cost bands are indicative ranges observed in the market and will differ by domicile, size and counterparty. Instruments named as examples illustrate how a stream can be expressed and are not recommendations; availability, domicile, tax treatment, fees and index methodology differ by investor and change over time, so verify current terms before acting. Any allocation decision should be taken with reference to a specific mandate and risk budget.