Property, commodities, gold and cash.
These are the assets that do not fit neatly into shares or bonds. Some of them produce income, some produce nothing at all, and the difference matters more than anything else when working out what to expect from them.
The idea behind all of them
Every forecast here starts with the same question: what does this asset actually pay you, and what happens to that payment over time?
A property fund collects rent. An infrastructure fund collects tolls and utility bills. Cash collects interest. Those three have something to compound and their expected returns follow from the income plus whatever that income does over time.
Gold has no contractual income. Commodity futures combine collateral income and uncertain futures returns. These need different models from shares or property.
Current alternative models
These ten-year annual estimates use the same calculations as the default current Assumptions, Forecaster and optimizer transfer. Sources publish at different frequencies. The observation dates below are retained; a quarterly report is never labelled as a daily market quote. No external expert endorsement is claimed.
| Asset / model | Observed | Central return | ||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| US REITs cash flow Calculation, source dates and limitationsNareit All Equity REIT income yield 3.68%; real dividend-per-share trend 1.449% over 2011Q2-2026Q2, four-quarter sums deflated with quarter-end CPI. Our 50% shrinkage towards zero gives 0.725% real growth. FFO-per-share trend 1.792% is a cross-check, not added income. Share issuance, repurchases and distributions enter per-share growth once. Changing industry membership and aggregate per-share measurement limit this proxy. Nominal cash-flow IRR, unchanged exit yield centrally; growth ±2pp and terminal value ±25% over ten years. No automatic yield reversion or extra capex deduction from dividends; gross of wrapper fees. Unhedged, unchanged nominal exchange rates in other investor currencies; FX uncertainty remains. Sensitivities are not probability bounds.
| 2026-08-31 | 7.12% | ||||||||||||||||||
| South African Listed Property cash flow Calculation, source dates and limitationsSatrix Property ETF (STXPRO), the same fund as the risk history. Four consecutive quarterly income distributions through 2026-06-30 total 0.8361 ZAR per unit, using the issuer's rounded cents-per-unit figures. Dividing by issuer NAV 14.0027 ZAR on 2026-09-22 gives 5.971% net distribution yield. NAV is a fund valuation, not an exchange closing trade; trading premiums and discounts can change an investor's yield. Income is after the fund's costs: no second TER deduction. The fund changed from S&P to FTSE/JSE All Property J803 on 2026-06-19; much of the trailing income was earned by the preceding basket. The current fund includes property holding/development companies and foreign exposures, so the narrower SA REIT industry index's yield and distribution growth are not substituted. Constant real per-unit income is a structural baseline, not measured forward rental or FFO growth. Shared South African CPI 4.012% enters once; unchanged exit income yield centrally. NAV must be no more than seven days old and the issuer document must have been retrieved within ten days. Quarterly distributions and their disclosure have a 135-day maximum age (one quarter plus publication allowance); a monthly factsheet date is not treated as a daily yield observation. No extra buybacks, leverage, capex or property appreciation is added to distributions. Non-recurring distributions and changing membership remain limitations. Growth ±2pp and terminal value ±25% over ten years are sensitivity cases, not probability intervals. Other investor currencies assume unchanged nominal FX, retaining FX risk. Net of fund costs, before investor taxes. One conditional cash-flow IRR is not probability-calibrated expected compound growth. Unhedged, unchanged nominal exchange rates in other investor currencies; FX uncertainty remains. Sensitivities are not probability bounds.
| 2026-09-22 | 10.22% | ||||||||||||||||||
| Canadian Listed Property cash flow Calculation, source dates and limitationsXRE, the same Canadian REIT fund as the risk history. Investment income earned in its latest 6-month accounts to 2026-06-30, after all reported fund expenses, is 0.278182 CAD per weighted-average unit. Annualising that period gives 0.556364 CAD; divided by the 2026-09-25 market close 15.53 CAD, the net investment-income yield is 3.583%. This is a reported-period run rate, not a trailing-twelve-month distribution yield. It includes underlying trust distributions recognised as income, interest and securities lending, while excluding the fund's realised trading gains and unrealised revaluations. Actual cash payouts can exceed earned income; the excess is not added as new return. Tax return-of-capital classifications alone do not measure economic income. Underlying trust distributions can still contain non-recurring items; the available data do not establish sustainable FFO or rental growth. Constant real per-unit income is the central structural baseline, with horizon-matched Canadian inflation 2.041% applied once and unchanged exit income yield. Current income may be seasonal; the half-year annualisation is a forecast assumption. No extra buybacks, capex, leverage premium or second fund-fee deduction. Growth ±2pp and terminal value ±25% over ten years are sensitivities, not probability intervals. Net of this wrapper's reported expenses, before investor taxes. One conditional cash-flow IRR is not probability-calibrated expected compound growth. Unhedged, unchanged nominal exchange rates in other investor currencies; FX uncertainty remains. Sensitivities are not probability bounds.
| 2026-09-25 | 5.70% | ||||||||||||||||||
| Global REITs cash flow Calculation, source dates and limitationsNasdaq Developed Markets Real Estate Index: trailing gross ordinary-distribution yield 3.877%, recovered from matched daily price and gross total-return index levels. Same Nasdaq broad real-estate sector proxy used in this row’s risk history. Includes property companies beyond REITs; not an exact FTSE EPRA or fund replication. USD distributions contain historical currency effects. No separate FX trend is added. Real USD index-unit payout trend -2.246% over 2006 to 2025, 20 complete years. Annual payouts are deflated with December US CPI; the missing October 2025 CPI is not filled. Half the log trend gives -1.130% real growth, then horizon-matched US inflation 2.573% gives nominal payout growth. This shrinkage is a planning choice supported by limited chronological diagnostics, not a statistically established best forecast. Historical USD payouts include currency and membership changes. Nominal cash-flow IRR with unchanged exit yield centrally; no extra buyback, FFO, inflation, capex or leverage premium. Dividend growth already reflects issuer reinvestment, dilution and financing. Growth ±2pp and terminal value ±25% over ten years are sensitivities, not probability intervals. Gross of wrapper fees and investor taxes. One conditional path is not probability-calibrated expected compound growth. Unhedged, unchanged nominal exchange rates in other investor currencies; FX uncertainty remains. Sensitivities are not probability bounds.
| 2026-09-25 | 5.35% | ||||||||||||||||||
| UK Listed Property cash flow Calculation, source dates and limitationsNasdaq United Kingdom Real Estate Index: trailing gross ordinary-distribution yield 5.289%, recovered from matched daily price and gross total-return index levels. Same Nasdaq broad real-estate sector proxy used in this row’s risk history. Includes property companies beyond REITs; not an exact FTSE EPRA or fund replication. USD distributions contain historical currency effects. No separate FX trend is added. Real USD index-unit payout trend -4.476% over 2006 to 2025, 19 complete years; incomplete 2013 excluded. Annual payouts are deflated with December US CPI; the missing October 2025 CPI is not filled. Half the log trend gives -2.264% real growth, then horizon-matched US inflation 2.573% gives nominal payout growth. This shrinkage is a planning choice supported by limited chronological diagnostics, not a statistically established best forecast. Historical USD payouts include currency and membership changes. Nominal cash-flow IRR with unchanged exit yield centrally; no extra buyback, FFO, inflation, capex or leverage premium. Dividend growth already reflects issuer reinvestment, dilution and financing. Growth ±2pp and terminal value ±25% over ten years are sensitivities, not probability intervals. Gross of wrapper fees and investor taxes. One conditional path is not probability-calibrated expected compound growth. Unhedged, unchanged nominal exchange rates in other investor currencies; FX uncertainty remains. Sensitivities are not probability bounds.
| 2026-09-25 | 5.55% | ||||||||||||||||||
| Europe Listed Property cash flow Calculation, source dates and limitationsNASDAQ Developed Markets: Europe Real Estate Index: trailing gross ordinary-distribution yield 4.938%, recovered from matched daily price and gross total-return index levels. Same Nasdaq broad real-estate sector proxy used in this row’s risk history. Includes property companies beyond REITs; not an exact FTSE EPRA or fund replication. USD distributions contain historical currency effects. No separate FX trend is added. Real USD index-unit payout trend -4.409% over 2006 to 2025, 20 complete years. Annual payouts are deflated with December US CPI; the missing October 2025 CPI is not filled. Half the log trend gives -2.230% real growth, then horizon-matched US inflation 2.573% gives nominal payout growth. This shrinkage is a planning choice supported by limited chronological diagnostics, not a statistically established best forecast. Historical USD payouts include currency and membership changes. Nominal cash-flow IRR with unchanged exit yield centrally; no extra buyback, FFO, inflation, capex or leverage premium. Dividend growth already reflects issuer reinvestment, dilution and financing. Growth ±2pp and terminal value ±25% over ten years are sensitivities, not probability intervals. Gross of wrapper fees and investor taxes. One conditional path is not probability-calibrated expected compound growth. Unhedged, unchanged nominal exchange rates in other investor currencies; FX uncertainty remains. Sensitivities are not probability bounds.
| 2026-09-25 | 5.24% | ||||||||||||||||||
| Listed Infrastructure cash flow Calculation, source dates and limitationsMSCI World Infrastructure current dividend yield 3.54%, P/E 17.34 (valuation context). Our real per-share growth prior 1.750% is half the broad-world structural growth assumption, shrunk towards zero for mature capital-intensive businesses; not measured sector earnings growth. No separate buyback premium. Nominal cash-flow IRR, unchanged multiple centrally; growth ±2pp and exit value ±25% over ten years. Developed-market infrastructure benchmark proxy, not an exact replication of every listed infrastructure fund. Gross of fees. Unhedged, unchanged nominal exchange rates in other investor currencies; FX uncertainty remains. Sensitivities are not probability bounds.
| 2026-08-31 | 8.06% | ||||||||||||||||||
| Commodities (Broad) structural premium Calculation, source dates and limitationsFully collateralized broad commodity futures: horizon-matched ACM expected cash × a 1.5% annual relative-wealth futures premium. Our premium shrinks the published roughly 3% long-history equal-weighted result halfway towards zero because broad commercial indexes differ in weighting and rolling. The premium includes spot changes, rolling and diversification; inflation and roll yield are not added again. No claim that today's futures curve persists for ten years. −3%/+4.5% premium sensitivities, gross of implementation costs; index design materially changes results. Unhedged, unchanged nominal exchange rates in other investor currencies; FX uncertainty remains. Sensitivities are not probability bounds.
| 2026-09-22 | 6.01% | ||||||||||||||||||
| Gold monetary scenario Calculation, source dates and limitationsPhysical bullion monetary scenario: observed spot 2026-09-25, US M2 2026-08, no projection of missing global money data. Matched monthly bullion/M2 ratios over twenty years; latest spot divided by latest observed M2 is 1.80 times the historical median. Twenty-year observed M2 growth 6.263% is the future money-growth assumption. 0.3333333333333333 log-ratio reversion centrally; no/full reversion and money growth ±2pp in sensitivities. Growth, valuation and 0.4% annual holding cost compound multiplicatively. Monthly averages and latest daily spot are different observation frequencies. US M2 is a monetary proxy, not all global gold demand or intrinsic value; this heuristic is not the World Gold Council GLTER model or a validated causal forecast. Unhedged, unchanged nominal exchange rates in other investor currencies; FX uncertainty remains. Sensitivities are not probability bounds.
| 2026-09-25 | 3.79% | ||||||||||||||||||
| Listed Private Equity factor proxy Calculation, source dates and limitationsListed private equity: joint ac_world equity beta 1.395 and us_treasuries beta -0.075, solved together from the app's covariance assumptions. Current arithmetic factor premia over horizon-matched cash, converted with 25.59% economic volatility. Zero net alpha; no assumed manager selection skill. Global listed managers and investment companies differ from unlisted buyout funds. No extra illiquidity premium or automatic discount-to-NAV closure is added. Scenarios cross both factor forecasts with net alpha ±2pp; sensitivities are deliberately broad and are not calibrated confidence intervals. Unhedged, unchanged nominal exchange rates in other investor currencies; FX uncertainty remains. Sensitivities are not probability bounds.
| 2026-09-22 | 8.03% | ||||||||||||||||||
| Listed Private Debt factor proxy Calculation, source dates and limitationsListed private debt / BDC shares: joint us_large_cap equity beta 0.826 and high_yield beta 0.415, solved together from the app's covariance assumptions. Current arithmetic factor premia over horizon-matched cash, converted with 20.24% economic volatility. Zero net alpha; no assumed manager selection skill. The credit factor uses current portfolio yields after default losses. BDC distributions are not added to total equity returns. This is a factor proxy, not a loan-by-loan or price-to-NAV valuation. Scenarios cross both factor forecasts with net alpha ±2pp; sensitivities are deliberately broad and are not calibrated confidence intervals. Unhedged, unchanged nominal exchange rates in other investor currencies; FX uncertainty remains. Sensitivities are not probability bounds.
| 2026-09-22 | 6.51% | ||||||||||||||||||
| Private Equity private planning Calculation, source dates and limitationsBuyout cash-flow model: 12x entry and 12x exit EBITDA, 37% debt/value, 5.6% observed EBITDA growth fading to 4.62%, 60% cash conversion after tax/capex/working capital. Borrowing at cash + 5pp; cash repays debt first. Management 1.75% stepping down to 1.25% after year 5, 20% carry after 8% preferred return with catch-up, 1% entry cost on invested equity. Central 8.16% assumes replacement investments earn the same net IRR; cash reinvestment gives 8.16%. No additional volatility subtraction from this cash-flow rate. No blanket private premium. The default holds the exit multiple equal to the entry multiple; an edited exit multiple changes that assumption. Cash conversion and later fund terms remain judgement inputs. Net of modelled fund fees, before investor tax. This is a diversified maintained-allocation planning forecast, not the return of a chosen manager. Cash-flow and valuation estimates set geometric planning growth; arithmetic means are inferred once using the same economic volatility as portfolio risk. A lognormal moment approximation is used consistently across the app. It is a modelling assumption, not empirical probability calibration. Sensitivities are not confidence intervals; capital calls, liquidity and manager selection can change realised outcomes. Unhedged, unchanged nominal exchange rates in other investor currencies; FX uncertainty remains. Sensitivities are not probability bounds.
| 2026-09-22 | 8.16% | ||||||||||||||||||
| Venture Capital private planning Calculation, source dates and limitationsVenture capital: median of three planning models. (1) Current buyout net return with JPM's structural net VC/buyout spread -1.7pp: 6.46%. (2) Current small-cap return plus PGIM's 2.2pp gross private premium, less 2% management and 20% carry above 8%: 7.69%. (3) Current small-growth factor return plus shrunk net fund residual 1.28pp in arithmetic space: 6.52%. Institutional spreads are dated structural priors, not current house forecasts. Residual estimate 4.32pp, standard error 3.08, prior SD 2pp; 2009-2025 sample misses dotcom/GFC and has benchmark/backfill uncertainty. We do not use pooled IRR as a return series or claim company-level failure modelling. Net of modelled fund fees, before investor tax. This is a diversified maintained-allocation planning forecast, not the return of a chosen manager. Cash-flow and valuation estimates set geometric planning growth; arithmetic means are inferred once using the same economic volatility as portfolio risk. A lognormal moment approximation is used consistently across the app. It is a modelling assumption, not empirical probability calibration. Sensitivities are not confidence intervals; capital calls, liquidity and manager selection can change realised outcomes. Unhedged, unchanged nominal exchange rates in other investor currencies; FX uncertainty remains. Sensitivities are not probability bounds.
| 2026-09-22 | 6.52% | ||||||||||||||||||
| Global Infrastructure private planning Calculation, source dates and limitationsPrivate infrastructure: SIPA Infra300 Q2 valuation-implied equity discount rate 10.44% anchors gross planning growth. Treating a required discount rate as long-run geometric growth is an explicit equilibrium assumption, not a published expected CAGR. Project borrowing is already reflected in equity value: no second leverage multiplier. Fund layer: 1.25% annual wealth fee and 10% carry after 6% preferred return. ±3pp gross-rate sensitivity spans materially different cash-flow and valuation views. Benchmark includes broader infrastructure and mixed local currencies; treating its central rate as a USD planning anchor is approximate, with FX risk retained. Net of modelled fund fees, before investor tax. This is a diversified maintained-allocation planning forecast, not the return of a chosen manager. Cash-flow and valuation estimates set geometric planning growth; arithmetic means are inferred once using the same economic volatility as portfolio risk. A lognormal moment approximation is used consistently across the app. It is a modelling assumption, not empirical probability calibration. Sensitivities are not confidence intervals; capital calls, liquidity and manager selection can change realised outcomes. Unhedged, unchanged nominal exchange rates in other investor currencies; FX uncertainty remains. Sensitivities are not probability bounds.
| 2026-09-22 | 8.41% | ||||||||||||||||||
| US Core Real Estate private planning Calculation, source dates and limitationsCore property: actual ODCE income 4.08% already includes interest. Recover NOI, grow it with current inflation plus 1% real, deduct fixed interest once and 1.2% asset-value maintenance capex. 27.1% debt/value, 5.5% financing, unchanged exit yield. Gross/net observed fund wealth gives 0.823% annual fee allowance. Central reinvestment maintains property exposure; cash reinvestment gives 6.21%. Current inflation does not revalue the starting income. Net of modelled fund fees, before investor tax. This is a diversified maintained-allocation planning forecast, not the return of a chosen manager. Cash-flow and valuation estimates set geometric planning growth; arithmetic means are inferred once using the same economic volatility as portfolio risk. A lognormal moment approximation is used consistently across the app. It is a modelling assumption, not empirical probability calibration. Sensitivities are not confidence intervals; capital calls, liquidity and manager selection can change realised outcomes. Unhedged, unchanged nominal exchange rates in other investor currencies; FX uncertainty remains. Sensitivities are not probability bounds.
| 2026-09-22 | 6.40% | ||||||||||||||||||
| Diversified Hedge Funds private planning Calculation, source dates and limitationsDiversified hedge funds: covariance-consistent equity and credit factor exposures, plus 1.15pp net strategy residual in the arithmetic mean. This rounded middle of recent matched-fund residual evidence (~0.55pp) and Cliffwater's strategy mixture (1.77pp) includes omitted arbitrage/trend premia and skill after fees. It is not guaranteed alpha. Whole-sample historical alpha is not extrapolated; net residual 0-2.3pp is a sensitivity. Economic risk is anchored to the diversified JPM fund category, with source residual dependence estimated from the broad net-fund history. The former full ARMA risk estimate is retained as a comparison, not treated as validated economic volatility. Strategy composition and the transfer of the net premium remain explicit uncertainties. Net of modelled fund fees, before investor tax. This is a diversified maintained-allocation planning forecast, not the return of a chosen manager. Cash-flow and valuation estimates set geometric planning growth; arithmetic means are inferred once using the same economic volatility as portfolio risk. A lognormal moment approximation is used consistently across the app. It is a modelling assumption, not empirical probability calibration. Sensitivities are not confidence intervals; capital calls, liquidity and manager selection can change realised outcomes. Unhedged, unchanged nominal exchange rates in other investor currencies; FX uncertainty remains. Sensitivities are not probability bounds.
| 2026-09-22 | 6.42% | ||||||||||||||||||
| Private Debt (Direct Lending) private planning Calculation, source dates and limitationsDirect lending: cash 4.44% + 1.75× private loan spread 5.25% + 1.75× annual origination income 0.40% − 1.75× annual credit loss 0.825% − 0.75× funding spread 2%. Management 1.375% and administration 0.35% on investor equity; 12.5% incentive on annual gains above 6%. Expected incentive fees integrated over pre-fee outcomes, not applied just to average income. Post-fee mean 8.79% and 11.50% economic risk are matched to the app's lognormal return convention once. Fee-transformed outcomes are approximated by their moments; annual reset is not a perpetual high-water mark. Private spreads already include illiquidity compensation. No second premium, dividend or discount-to-NAV gain is added. Net of modelled fund fees, before investor tax. This is a diversified maintained-allocation planning forecast, not the return of a chosen manager. Cash-flow and valuation estimates set geometric planning growth; arithmetic means are inferred once using the same economic volatility as portfolio risk. A lognormal moment approximation is used consistently across the app. It is a modelling assumption, not empirical probability calibration. Sensitivities are not confidence intervals; capital calls, liquidity and manager selection can change realised outcomes. Unhedged, unchanged nominal exchange rates in other investor currencies; FX uncertainty remains. Sensitivities are not probability bounds.
| 2026-09-22 | 8.19% |
General alternative review: 2026-09-07; private-fund review: 2026-09-08; next required review by 2027-03-01. Unhedged returns assume unchanged nominal exchange rates in other currencies. The risk model still reflects currency exposure. Sensitivities are not probability bounds. Private-asset and hedge-fund models remain proxies; gold remains a monetary scenario.
Property and REITs
US REITs use the published Nareit All Equity income yield and fifteen-year real dividend-per-share growth, shrunk halfway towards zero. Four-quarter sums reduce payout seasonality. FFO per share is a cross-check, not a second source of return. Share issuance and repurchases enter through per-share results. The central exit yield stays unchanged; growth and terminal-value sensitivities make valuation uncertainty explicit.
US core property uses current ODCE fund income, observed leverage, maintenance spending, growth and fund fees. Its central risk matches the diversified JPM core-property volatility and US REIT correlation as one pair, while other dependence is estimated from model paths. The resulting USD volatility is 11.39%. The full appraisal reconstruction remains a comparison case. A statistical reconstruction alone cannot establish the true size of economic price movements.
Infrastructure
Listed infrastructure uses current MSCI World Infrastructure dividend yield, conservative real per-share growth and a terminal sale value. Growth is half the broad-world structural prior, not measured current sector earnings growth. The central valuation multiple is unchanged. The benchmark covers developed markets and is a proxy for the broad asset class.
Private infrastructure starts from the current SIPA valuation-implied equity rate, then subtracts modelled fund fees. Project borrowing is already reflected in the equity rate. Treating that required rate as long-run growth is an explicit equilibrium assumption; it is not a published expected CAGR. The source is broader than a pure core infrastructure fund. The shared calculation below gives the inputs and sensitivities.
Commodities
The central futures premium is 1.5%: our conservative shrinkage of a roughly 3% long-history equal-weighted result discussed by AQR. Commercial indexes differ in commodity weights and roll schedules. The premium already includes spot changes, rolling and diversification. We do not add inflation or a separate roll yield. Negative and higher premium cases show uncertainty; these are not fitted confidence intervals.
Gold
We use observed bullion prices, the latest published US M2 and twenty years of matched monthly gold/M2 observations. Future money growth is assumed to match its twenty-year pace. One-third of the log valuation gap closes over the forecast horizon centrally; no reversion, full reversion and different money growth form sensitivities. Growth, valuation and a 0.4% annual holding cost compound multiplicatively.
US M2 is a monetary yardstick, not global money, gold demand or intrinsic value. Its release lags the daily bullion quote and we show both dates. This model does not invent missing money observations, and it is not the World Gold Council GLTER model. Its central result is an uncertain scenario; an investor can reasonably choose a different assumption.
Cash
Cash pays whatever short-term interest rates happen to be, which is knowable today and unknowable in three years. The question is what they will average over your horizon.
We use the New York Fed Adrian, Crump and Moench model at the selected horizon. Its fitted Treasury yield less its estimated term premium gives a risk-neutral yield. This estimates future short rates with the model’s convexity convention; it is not an observed future deposit rate.
The published yields use continuous compounding. We convert them before the app compounds returns annually. Daily observations are refreshed automatically; observations older than ten calendar days are rejected. Horizons outside the published one-to-ten-year range are unavailable. The main USD assumptions use ten years; the Forecaster uses its selected horizon.
Bear and bull are sensitivities half a percentage point either side, not estimated confidence bounds. The model can be revised, and realised cash returns can differ substantially. Source: New York Fed ACM term premia.
Canadian-dollar cash
We use the Bank of Canada’s latest Market Participants Survey, including its dated policy targets and its own long-run neutral-rate estimate. Starting from the observed policy rate, we interpolate between the calendar targets, join the neutral rate at five years, and model daily reinvestment on ACT/365. Targets for a month or quarter are endpoints, not an average rate for a whole year.
This is a conditional gross cash forecast before fees. Survey quartile paths show sensitivity, not the probability of future cash returns. Policy observations expire after ten days; a survey more than 120 days past publication is unavailable. The same calculation feeds CAD cash in the workstation and the cash forecaster.
Source information is available free of charge from the Bank of Canada. Portfolio Lab interpolates and compounds the source figures; the Bank does not endorse this derived forecast. USD and CAD cash are different currency instruments.
Euro overnight cash
Euro cash uses the ECB Survey of Monetary Analysts’ euro short-term rate (€STR) expectations, anchored to the latest observed €STR. Governing Council periods use official reserve-maintenance start dates; quarterly targets use quarter ends. We interpolate these targets and join the survey’s own long-run rate at five years, compounding daily on ACT/360.
This is a conditional gross cash forecast, not a realized ECB benchmark index. Low and high paths use survey quartiles as sensitivities. The same calculation feeds euro cash in the workstation and forecaster. Rate observations expire after ten days; survey publications expire after ninety days.
Source: ECB survey and releases. The source statistics are available free. Portfolio Lab interpolates and compounds them; this is not an ECB forecast or endorsement.
Sterling and rand cash
Sterling starts with current SONIA and follows the latest Bank of England participant survey, including its dated policy targets and neutral-rate quartiles. The observed SONIA-minus-Bank-Rate spread is held constant by assumption. Rand starts with ZARONIA and transitions over five years to Portfolio Lab’s 6% nominal overnight-rate assumption, with 5% and 7% sensitivities. That terminal rate is not a numerical SARB forecast. Both use daily ACT/365 compounding before fees; neither treats a government-bond term premium as cash income. These same forecasts feed the workstation and Forecaster.
Private credit
Direct lending starts with the shared cash path and private loan spreads. It adds origination income, subtracts expected credit losses, and applies fund borrowing and its financing cost. Management and operating fees are charged on investor equity. Incentive fees are calculated across possible outcomes, because a fee on the average return understates average fees.
The central fund owns £1.75 of loans for each £1 of investor equity. Its debt is £0.75. The 11.5% economic risk estimate reflects this leverage and fee model, rather than smooth reported loan values. Private spreads already include compensation for illiquidity; no extra bonus is added. Loan data and fund mandates remain imperfect matches. Listed BDCs remain a separate market-priced investment.
Private equity, venture capital and hedge funds
Buyouts use entry valuation, business cash generation, debt repayment, financing costs, management fees and carried interest. Central exit multiples stay unchanged. Management fees step down after year five. Reinvesting distributions into replacement funds maintains the allocation; reinvesting only in cash is a separate sensitivity.
Venture capital takes the median of three models: a relationship to the current buyout forecast, a private premium over current small-cap returns after fees, and a small-growth factor model with strongly shrunk historical net residual returns. The recent fund sample misses earlier crises. This is a provisional planning ensemble, not a company-level prediction of failures and exits.
Hedge funds use equity and credit exposures consistent with the risk matrix, plus an explicit net strategy premium. That premium includes effects such as arbitrage and trend as well as manager skill; it is not guaranteed alpha. We do not carry the large full-history residual forward unchanged.
Private risk uses matched JPM 2026 USD fund categories as the central reference for volatility and public-benchmark correlation together. Direct lending retains its stated borrowing and fee risk budget. Empirical residual dependence and actual currency conversions connect these pairs to the whole portfolio; the resulting matrix is Portfolio Lab’s model, not JPM’s published matrix. The former 34% buyout estimate and full property and hedge-fund reconstructions remain research comparison cases. They are not discarded evidence, but neither are they proven universal defaults. No fresh price observation is inferred from a structural risk assumption.
These estimates describe diversified fund equity, including its borrowing. Capital calls, withdrawal restrictions and forced-sale discounts also matter and require separate investor-specific stress assumptions. A valid covariance matrix checks mathematical consistency; it does not prove forecast accuracy or eliminate uncertainty about private valuations.
All six private forecasts use the same net, geometric planning-return convention. The app derives arithmetic means once for simulation and portfolio calculations. A central cash-flow path does not prove a probability-calibrated expected return; this remaining uncertainty is stated in every model.
Central risk and research comparison
| Private asset | Central USD volatility | Previous research-model volatility |
|---|---|---|
| private equity | 19.78% | 34.00% |
| venture capital | 22.20% | 23.00% |
| infrastructure | 10.25% | 13.30% |
| us core real estate | 11.39% | 15.06% |
| hedge funds | 5.75% | 9.51% |
| private debt | 11.50% | 11.50% |
The last column is another model estimate, not a confidence bound or a daily observation. Its benchmarks and historical periods can differ. Central means the planning assumption currently used throughout the app; it does not mean proven correct. Direct lending retains its explicit borrowing and fee risk budget.
Could you meet a capital call?
Illustrative assumptions, not predictions. Enter all amounts in the same currency. This checks one funding scenario separately from return volatility. A locked fund starts with no sale capacity; increase it only if a sale or redemption is actually possible.
Capital call: 25,000. Distributions received: 5,000.
Cash shortfall before sales: 10,000.
Fund value sold: 0. Loss from the sale discount: 0.
Still to fund: 10,000. The available cash and permitted sales do not cover this call.
A capital call is an investment payment, not automatically a loss. Sale discounts are applied to the fund value entered; use its stressed value if modelling a market fall as well. Taxes, financing, subsequent calls and sale execution are outside this example. These inputs do not change your portfolio forecasts.
Listed and private exposures
Listed and private vehicles have different liquidity, leverage, fees and valuation practices. The app therefore keeps separate rows and risk assumptions. The current infrastructure return models are described above; older published-reference choices remain explicitly separate in Assumptions.
| Private exposure | Volatility assumption | Listed exposure | Volatility assumption |
|---|---|---|---|
| Private Equity | 19.78% | Listed Private Equity | 25.59% |
| Global Infrastructure | 10.25% | Listed Infrastructure | 16.88% |
| Private Debt (Direct Lending) | 11.50% | Listed Private Debt | 20.24% |
These are risk-model assumptions, not daily observations or guaranteed future volatility. Private appraisals can smooth measured risk. No return premium is added simply because a holding is illiquid; actual fund cash flows and fee terms are needed to assess one.
Bitcoin
Bitcoin is the one asset on this page we do not forecast. You enter your own number, and the tool shows reference points beside the box to help you pick it.
Every other forecast here starts from income. A property fund pays rent, a bond pays a coupon, a company pays a dividend out of earnings, and the yield you can see today does most of the work in the answer. Bitcoin pays nothing to its holders, so that arithmetic has nowhere to start. This is a property of the asset rather than a mark against it, and it is the same reason gold gets the thinnest treatment of anything above. The difference is that gold has several thousand years of price history to lean on and Bitcoin has about sixteen.
What the published models are worth
Three models get quoted often enough to be worth answering directly. Stock to flow compares the existing supply to yearly new issuance. It fits the past closely, but its explanatory power largely disappears once you control for the passage of time, because scheduled halvings and the calendar move almost in lockstep. Metcalfe style models tie value to the square of the number of users, and share the same problem of describing history better than they call the future. The power law, which fits price against time on a log scale, has the best fit of the three, and it is a fitted trend line rather than a tested forecasting model.
Independent tests at horizons of one to six months have not found a model that reliably beats the assumption that tomorrow’s price is today’s. That finding is the reason this tab asks you for a number instead of printing one.
The four reference points
The past rates are compound annual growth over one, three, five and seven years, taken from a weekly price series that we store and refresh every Monday. They are the record, not a projection, and the spread between them is the useful part: seven year and one year figures on the same screen usually disagree violently.
Those two constants are fitted here over the whole stored series, back to September 2014, rather than copied from a chart. The fit has an R squared of 0.92 on log scaled data, which sounds better than it is, because almost anything that rises over time will fit a rising line on a log scale.
The adoption line asks a simpler question: if roughly three times as many people hold Bitcoin in ten years and the price scales with that, what annual return does it imply. It is written as a multiple rather than as a move between two ownership percentages because estimates of how many people hold Bitcoin today disagree with each other by a wide margin, and only the ratio changes the answer.
The digital gold line asks what return is implied if Bitcoin reaches half the value of all the gold ever mined. The World Gold Council reports 222,600 tonnes above ground as of 2026-06-30; the app combines that with a source-dated bullion quote. Stock is refreshed quarterly and price is refreshed daily. This scenario holds gold’s market value and Bitcoin’s supply constant; it is not a gold or Bitcoin price forecast. A target below today’s Bitcoin value produces a negative implied return.
None of the four is a forecast, and they disagree with each other on purpose. If they all pointed at the same number it would mean the assumptions behind them were doing no work.
The inflation number, and why there is only one
Real forecasts are converted to nominal dollars by compounding with one shared US inflation expectation matched to the selected horizon from the Cleveland Fed monthly model. Its inputs include Treasury yields, inflation swaps, CPI and surveys. The source month travels with the input; a retrieval date does not make an old estimate new. The collector rejects missing, future, inconsistent or expired releases.
Source: Inflation Expectations, Federal Reserve Bank of Cleveland. Portfolio Lab extracts the monthly term structure and converts decimals to percentage points. The adapted data are shared under CC BY-SA 4.0. No endorsement is implied. The three-, five- and ten-year views use their matching US estimates; non-US inflation references remain separately sourced.
Where the inputs come from
The current-model table above links directly to each source and uses the same model text and observation dates as the app. Nareit supplies REIT income and per-share history; NCREIF supplies quarterly private-property income; MSCI supplies listed-infrastructure valuations. The New York Fed supplies the cash term model. The Federal Reserve supplies observed US M2 and CPI; the World Bank supplies monthly bullion history and State Street supplies the latest bullion quote. AQR research informs explicitly identified structural assumptions without implying an endorsement or copying its complete forecast model.
The daily refresh job discovers current quarterly and monthly workbooks, validates identities, units, dates and historical continuity, and tests proposed changes before publication. Daily market inputs expire after ten days, monthly market observations after forty days, quarterly property and REIT growth inputs after 130 days, and monthly M2 after 65 days. A failed download never refreshes an observation date. Structural assumptions have a separate review deadline.
What this cannot tell you
The lower and upper estimates change stated inputs, including growth, terminal value, factor returns and monetary assumptions. They are modelled sensitivities, not statistical confidence bounds or exhaustive worst cases.
Most rows are gross of wrapper charges and taxes. Gold includes its stated holding cost; the six private models include their stated fund-fee assumptions. Actual fees and fund exposures may require different assumptions.
Property income provides a cash-flow anchor, while gold and commodity returns have less direct valuation support. Joint factor consistency does not establish fair value for a private fund. These methods still need ongoing empirical evaluation, especially for gold, private assets and hedge funds; no claim of independent expert validation is made.
Frequently asked questions
How do you forecast REIT returns?
Current Nareit income yield and a conservative real dividend-per-share growth assumption feed an annual cash-flow model. The central case keeps the exit yield unchanged. Capital distributions are counted once; growth and terminal-value changes are separate sensitivities.
Does gold have a reliable expected return?
Gold has no contractual income. Our monetary scenario uses observed bullion prices and US M2, with a twenty-year ratio history and explicitly assumed partial reversion. It is a heuristic, not intrinsic value or a reliably validated prediction.
How do commodities earn a return?
A fully collateralized futures portfolio earns cash interest and a futures excess return. Our conservative structural premium already includes price movements, rolling and diversification, so inflation and roll yield are not added again.
How is cash forecast?
The New York Fed ACM horizon-matched risk-neutral yield is converted from continuous to annual compounding. It is a model estimate, not a guaranteed deposit rate.