Separate Accounts in Real Estate: Dedicated Allocations, IMAs, and the Institutional Single-LP Vehicle

Key Takeaways

The Institutional vs Retail SMA

Two unrelated structures share the abbreviation SMA. The retail version — the one that dominates Google's first page for "separately managed account" — is a tax-managed wrapper of public equities or municipal bonds sold by Schwab, Fidelity, Morningstar, and the RIA channel to high-net-worth individuals. It is a $5–$10M minimum, a quarterly statement, and a tax-loss-harvesting overlay. It has almost nothing in common with the institutional real estate SMA except the name.

The institutional real estate separate account is a $100M–$3B+ single-LP private vehicle governed by a customized Investment Management Agreement, with a dedicated investment team, bespoke mandate scope, custom fee and promote economics, and allocator-specific reporting. The LP is a pension fund, sovereign wealth fund, large insurance general account, or the very largest family offices. The vocabulary is institutional — dedicated mandate, fund-of-one, single-LP vehicle, captive fund, discretionary mandate — not the retail "SMA" branding.

This article addresses the institutional structure. If you are searching for the retail product, the Schwab and Fidelity SMA product pages are the right destination; nothing below applies to that market.

What an SMA Actually Is

A separate account in private real estate is a single-LP investment vehicle. The investor — typically a public pension plan, sovereign wealth fund, or large insurance company — commits capital that the manager (the GP) invests under a customized mandate. The LP either owns the underlying assets directly through a holding structure (a series of single-purpose LLCs or partnerships) or sits in a single-LP fund vehicle (a "fund-of-one") that the manager operates the same way it would operate a commingled fund. The defining feature is the LP base: one investor, not many. Everything else — mandate scope, fees, reporting, governance — flows from that.

The wedge against commingled funds is customization. A commingled fund is a pooled vehicle with a fixed strategy disclosed in the LPA, standard fees, a standard waterfall, and standard reporting. Every LP gets the same exposure pro rata. A separate account inverts the relationship: the LP specifies the mandate, the manager builds a portfolio inside it, and the LP receives reporting, fees, and economic terms tailored to its allocation strategy. The manager is no longer a product seller; it is a fiduciary executing a discretionary mandate.

The IMA: Investment Management Agreement

The governing document for a separate account is the Investment Management Agreement — the IMA. It is to a separate account what the LPA is to a commingled fund: the document that defines the mandate, the manager's authority, the fee economics, the reporting obligations, the consent rights, the termination triggers, and the indemnification framework. The LPA and the IMA serve the same role; the IMA is the bilateral-contract version because there is only one LP to sign.

A well-drafted institutional IMA covers nine sections that determine the operating economics of the relationship:

• Mandate scope. The investment strategy (core, core-plus, value-add, opportunistic), the geographic concentration limits, the property-type concentration limits, the leverage ceiling, the hold-period target, the development-risk tolerance, and the ESG screen. Each of these is a single negotiated parameter, not a paragraph of marketing text. A typical CalPERS mandate might read: "U.S. core and core-plus, max 30% single market, max 40% single property type, max 50% leverage on portfolio basis, target 7–10 year hold."
• Discretion. Whether the manager has full discretion to invest within the mandate or whether individual deals require LP consent. Full-discretion IMAs are the institutional norm above $250M because deal pace requires it; below that, "advisory" arrangements with deal-by-deal consent are more common.
• Fee structure. The base management fee (as a percentage of NAV, gross asset value, or committed capital), any incentive fee or promote above a hurdle, and the fee-offset mechanics for transaction and acquisition fees the manager might otherwise earn at the property level.
• Reporting cadence and content. Quarterly, monthly, or both; portfolio-level NAV, property-level NAV, debt schedules, hold-sell analysis, capital-call forecasts. The reporting in an institutional SMA is typically more granular and more frequent than the ILPA template a commingled fund delivers.
• LP consent rights. Affirmative LP consent on specified matters: new strategy areas outside the mandate, related-party transactions, leverage ceiling increases, side-letter arrangements with third parties (rare in an SMA but possible if the SMA invests through a commingled co-invest sleeve).
• Replacement and termination rights. The LP's ability to terminate the IMA "for cause" (key- person triggers, fraud, material breach) and, in many institutional SMAs, "no-fault" termination with notice and a wind-down period. No-fault termination rights are one of the biggest economic differences from a commingled-fund LPA, where LP-side termination is far more constrained.
• Standard of care. The fiduciary standard the manager owes the LP. Most institutional IMAs explicitly invoke the prudent-investor or prudent-fiduciary standard under the applicable state law (Delaware for the manager entity, the LP's home-state law for the LP's compliance overlay).
• Indemnification. The scope of LP indemnification of the manager, and the carve-outs (gross negligence, fraud, willful misconduct). Institutional IMAs typically narrow the indemnification scope versus a standard commingled-fund LPA.
• Audit rights. The LP's right to audit the manager's books, valuation methodology, and portfolio-level expense allocations on demand. Institutional SMAs typically have far broader audit rights than a commingled-fund LP.

Most institutional separate accounts are documented in 60–100 pages of IMA text. The LPA equivalent for a commingled fund runs 200–300 pages because it has to govern many LPs with potentially conflicting interests. The IMA is shorter and more bilateral because there is only one investor on the other side.

Typical Size and Fee Structure

Institutional separate accounts have a floor and a practical ceiling. The floor is roughly $100M of committed capital, though most established managers will not open an SMA below $250M because the setup, dedicated-team, and reporting costs are uneconomic below that. A new SMA requires a dedicated portfolio manager (or at least dedicated PM hours), a separate fund-admin and audit relationship, separate legal documentation, separate valuation workpapers, separate compliance reporting, and a custom reporting build. None of that scales linearly with size; a $100M SMA absorbs roughly the same overhead as a $500M SMA, which is why the fee discount only funds itself at scale.

The practical ceiling is north of $3B for the largest sovereign wealth and public pension relationships. ADIA, GIC, CPP Investments, CalPERS, and Norges Bank each operate multiple billion-dollar real-estate SMAs across managers. These are essentially captive operating relationships: the manager dedicates a team to the LP, deal flow gets routed to the SMA first, and the LP becomes the manager's largest single client by a wide margin.

Fee economics scale with size. A typical commingled value-add fund charges 1.5% on committed capital during the investment period stepping down to roughly 1.25% on net invested capital in harvest, plus 20% carried interest above an 8% preferred return. An equivalent-strategy SMA at $500M might price 100–125 bps on NAV with 15–17.5% carry above a 9% pref — roughly a 25–40 bps cut to the base fee and a 250–500 bps cut to the carry rate, with the hurdle moved up 100 bps. At $1B+ commitment levels SMAs price below 100 bps on NAV with carry compressed further or eliminated entirely for core strategies. The fee discount is real money: a 50 bps cut on a $500M mandate is $2.5M per year of fee revenue the manager is foregoing to win the relationship.

The discount funds itself for the manager in three ways. First, dedicated capital removes fundraising overhead — the manager is not raising a new commingled fund every three to four years for the same capital. Second, deal-flow allocation routes deals to the SMA first, which improves win rate. Third, the relationship creates a referral channel into the LP's broader allocation (co-invest, follow-on SMAs, other strategy mandates). The headline fee discount is real but the manager's effective P&L on an SMA is comparable to a commingled fund of similar size once these adjustments are baked in.

Why Pension Funds and Sovereign Wealth Use Them

The five institutional reasons an allocator opens a separate account instead of committing to a commingled fund are reasonably stable across the largest LPs:

1. Control over investment guidelines. A pension fund with a defined real-estate strategy — say, "U.S. industrial and multifamily, no office, max 50% leverage, ESG-compliant operating partners only" — cannot get that mandate inside a commingled fund whose LPA defines a broader strategy. An SMA lets the LP write its own mandate. CalPERS, CalSTRS, and Norges Bank have all used SMA mandates to enforce strategy shifts (e.g., underweighting office, overweighting industrial and life sciences) that a commingled LPA could not accommodate.
2. Customized fee structure. The 50–100 bps of fee compression versus a commingled-fund equivalent is significant relative to the LP's target real-estate return. On a $500M mandate, 50 bps of fee compression is $2.5M per year of incremental net return — effectively 50 bps added to the LP's reported IRR.
3. Bespoke reporting. Institutional LPs increasingly demand monthly or even biweekly portfolio reporting, asset-level transparency, custom risk metrics (geographic concentration, tenant credit, lease maturity ladders), and the ability to integrate the manager's reporting into the LP's internal portfolio management system. A commingled fund delivers a standardized quarterly ILPA-format report; an SMA delivers whatever the IMA specifies.
4. Direct ownership of assets. In a true separate account (not a fund-of-one wrapper), the LP holds direct legal title to the underlying real estate through SPV ownership structures it controls. This matters for tax structuring (UBTI considerations for ERISA plans, treaty-based withholding for sovereign LPs), collateral pledging, and the LP's ability to dispose of individual assets if portfolio strategy shifts. Commingled-fund LPs own a partnership interest, not the real estate.
5. Larger relative voice with the manager. A $500M SMA LP is, by definition, the manager's single most important client on that mandate. The LP gets first call on new deal flow, priority access to co-invest opportunities outside the SMA, dedicated portfolio-management attention, and a credible threat to terminate the relationship if performance or service degrades. A $50M commingled-fund commitment delivers none of these.

Named institutional examples bear this out. CalPERS operates multiple SMAs across CBRE Investment Management, GI Partners, and other partners covering domestic core, value-add, and global opportunistic mandates. CPP Investments operates a programmatic SMA structure across the largest North American developers and operators. GIC uses an SMA framework for a substantial share of its $50B+ global real-estate allocation. ADIA structures its global real-estate exposure primarily through SMAs and direct investments rather than commingled fund commitments. Norges Bank Investment Management operates SMA partnerships with TIAA, MetLife, and others for its $50B+ direct-real-estate program. GPIF in Japan and the major Canadian pension plans (OTPP, OMERS) operate similar structures. Specific commitment sizes and counterparties are reported in Pensions & Investments and IREI as awards are announced, though terms are not public.

SMA vs Commingled Fund Tradeoffs

The decision between SMA and commingled fund is not "SMA is always better" — it is a tradeoff between customization and scale. A large LP picks the structure that fits the specific allocation. Below a certain size, commingled is the only viable option; above a certain size, SMA almost always wins on net economics; in the middle range ($100M–$300M), the right answer depends on the LP's strategy specificity and operational capacity.

On the four axes where the SMA wins, the wins are sharp. Control: the SMA LP writes the mandate and can terminate the manager with notice. Fees: 10–25% cheaper than a commingled equivalent, with the carry on better LP-side terms. Customization: every parameter that matters to portfolio strategy is negotiable, not fixed. Reporting: monthly cadence and asset-level detail rather than quarterly portfolio-level ILPA.

On the three axes where the commingled fund wins, the wins are real but different in kind. Diversification: a commingled fund spreads the LP across more assets than a single SMA can practically hold. Operational simplicity: commingled-fund LPs have lower internal staffing requirements — they monitor a quarterly report; SMA LPs effectively run a small portfolio team. Scale of entry: a commingled fund commitment can be $25M–$50M; an SMA below $100M rarely works.

The institutional answer is usually both, not one or the other. Large public pension funds typically run their real-estate allocation as a mix: SMAs for the strategy areas where they want control and customization (say, U.S. core industrial, U.S. value-add multifamily) and commingled funds for the strategy areas where they want diversified exposure to a manager's wider pipeline (say, global opportunistic, European value-add, emerging-markets debt). The mix evolves as the allocation grows.

Fund-of-One: The Hybrid Structure

A fund-of-one is the most common practical form an institutional SMA actually takes. The LP commits capital not to a direct-ownership SMA but to a single-LP fund vehicle — a Delaware LP or LLC with one LP and a GP — that the manager operates exactly the way it would operate a commingled fund. The fund holds the assets, the manager runs the fund, the LP receives partnership interests instead of direct title to real estate.

Why use the fund wrapper instead of direct ownership? Three institutional reasons:

• The manager's operating model. Established managers operate commingled funds. The fund-of-one wrapper lets the manager run the SMA on the same operating chassis — same fund admin, same audit, same portfolio accounting, same valuation cadence — rather than building a parallel direct-ownership stack. Operational economics for the manager are materially better with the wrapper.
• Liability segregation. The LP's exposure is limited to its commitment amount inside the fund. Direct ownership exposes the LP to property-level liability (subject to the LLC subsidiary structures it would ordinarily use), which most institutional LPs prefer to avoid even when those structures are well drafted.
• Tax structuring flexibility. The fund vehicle is the standard tax-structuring shell. Sovereign LPs use offshore blocker structures; pension LPs use UBTI-blocked structures. Both are easier to implement inside a fund wrapper than around a direct-ownership SMA.

The fund-of-one is effectively an LPA-governed vehicle with the economic terms of an SMA. The LPA is shorter and simpler than a commingled-fund LPA because there is only one LP, but it covers the same structural territory — capital calls, distributions, GP authority, waterfall (if any), key-person provisions, indemnification. Most "SMAs" reported in Pensions & Investments and IREI are technically fund-of-one structures under the hood; the "separate account" label refers to the dedicated mandate, not the legal form.

A pure direct-ownership SMA — where the LP holds direct title to the real estate through SPVs it controls — is more common for sovereign wealth funds and the largest public pensions, where tax, treaty, and sovereign-immunity considerations make direct ownership cleaner than a fund wrapper. Norges Bank, GIC, and ADIA all operate direct-ownership programs in addition to fund-of-one SMAs with various institutional managers.

2026 Institutional Context

The 2024–2026 institutional allocator trend is a measured shift from commingled funds toward separate accounts, co-invest sidecars, and direct investments. The drivers are well documented in the Preqin Investor Outlook reports, the PERE LP Investor Survey, and the Pensions & Investments mandate coverage.

Three forces drive the shift:

• Cost. Large LPs have systematically more leverage on fees in an SMA than in a commingled fund. The 2024–2026 fee-compression cycle (Preqin median management fee drifting from 2.0% in 2018 toward 1.7% on new vintages) is even more pronounced inside SMAs.
• Customization. Allocator strategy is increasingly specific — underweight office, overweight industrial and life sciences, ESG-screen the operating partners, manage geographic concentration against the LP's broader portfolio. Commingled-fund LPAs cannot accommodate this level of specificity; SMA mandates can.
• Transparency. The 2024 ILPA Reporting Template 2.0 raised the bar on disclosure for commingled funds, but SMAs continue to deliver materially more granular reporting than even the upgraded ILPA standard.

The counter-force is the emerging-manager dynamic: managers without an established commingled franchise are increasingly offering fund-of-one structures as the entry point for institutional capital. The pitch is that an emerging manager can offer SMA-style terms (lower fees, better carry, custom mandate) in exchange for the LP becoming an anchor in what eventually becomes the manager's flagship commingled fund. For the LP this is a way to access emerging talent on bespoke terms; for the manager it is the path from boutique to institutional scale. The 2025–2026 vintage of emerging-manager real estate funds has been dominated by this fund-of-one anchor structure rather than traditional commingled raises.

A second 2026 trend is the rise of the club SMA — a separate account with two to five LPs sharing a single mandate. The structure is technically a small commingled fund (multiple LPs, single LPA) but operates economically as an SMA: each LP gets near-bespoke terms via side letters, the mandate is narrowly defined and LP-driven, and the LP base is small enough that governance can be consensus-based rather than majority-vote. Club SMAs reconcile the LP-side desire for customization with the LP-side desire to spread setup costs across more than one balance sheet.

How Apers Models SMA Structures

Modeling a separate account in Excel from scratch is similar work to modeling a commingled fund, with extra overhead in three places. First, the fee structure is bespoke — not a standard 1.5/20 schedule but a custom blend of NAV-based or committed-based management fees, with a custom carried-interest rate above a custom hurdle. The model has to parameterize all of these without hard-coding the standard defaults. Second, the reporting is bespoke — the LP wants per-asset NAV, per-asset debt, per-asset hold-sell analysis, and a portfolio-level dashboard, on monthly cadence. The model has to produce reporting outputs that match the LP's specification, not the standard ILPA template. Third, the mandate constraints are live — if the LP's IMA caps single-asset concentration at 15% of NAV and the model is showing a deal that would push portfolio concentration to 17%, the model has to flag the violation before the IC discussion, not after.

Apers builds SMA models from the IMA itself. The dedicated SMA model is upcoming; in the meantime, Apers constructs the structure on a per-mandate basis: the management-fee schedule from the IMA fee section, the carried-interest waterfall from the IMA economics section, the reporting template from the IMA reporting schedule, and the mandate-constraint checks from the IMA mandate-scope clauses. The output is a working SMA-specific model that ties to the IMA terms, not a commingled-fund template with a few parameters overridden.

For the underlying waterfall math — preferred return accrual, catch-up gross-ups, promote tier computation — CS-001 Multi-Class Equity Waterfall is the pre-built model that handles the institutional mechanics. SMA economics typically run a simpler waterfall (no LP catch-up because there is only one LP class, often a single promote tier rather than a ladder, sometimes no promote at all for core mandates), but CS-001's framework adapts.

Common Mistakes and Misreads

The misreads below show up in allocator-side memos, manager-side pitch materials, and consulting recommendations when the analyst hasn't structured SMAs at scale before.

• Confusing the institutional SMA with the retail SMA. The retail "separately managed account" sold by Schwab, Fidelity, and Morningstar is an entirely different product — a tax-managed equity or municipal-bond wrapper for high-net-worth individuals at a $5M–$10M minimum. The institutional real estate SMA shares only the name. Junior staff sometimes pull retail-SMA marketing into an institutional memo; it is not relevant.
• Assuming SMAs are always cheaper than commingled. The 50–100 bps fee discount on an SMA is real at the headline rate, but the LP absorbs internal staffing, governance, and operational costs that are diversified across many LPs in a commingled fund. Below roughly $200M–$300M of commitment, the LP-side internal costs can eat much of the fee discount. The structure decision is not "SMA = cheaper" but "SMA = cheaper at sufficient scale."
• Treating "fund-of-one" as a different structure from an SMA. A fund-of-one is the most common legal form an institutional SMA takes. Both are dedicated-mandate single-LP vehicles. The distinction is legal-form only (LP/LLC fund wrapper vs direct ownership), not economic.
• Reading the IMA as a "simpler LPA." The IMA is bilateral and shorter, but the institutional negotiation is harder, not easier. Every clause is bespoke; there is no market-standard template to fall back on. No-fault termination rights, audit rights, and mandate-scope language take more negotiation in an IMA than in a commingled-fund LPA precisely because there is no other LP to defer to.
• Underestimating the manager's deal-flow tension. A manager operating both a commingled fund and a large SMA in the same strategy faces an allocation problem on every deal: does this deal go to the fund, the SMA, or both pro rata? Mature managers solve this with a written allocation policy disclosed to both LP bases; less mature managers can allocate informally, which creates fiduciary risk and LP-side disputes. An institutional LP opening an SMA with a manager that runs a competing commingled vehicle should diligence the allocation policy carefully.
• Forgetting that SMAs reduce manager attention compared to a flagship fund. The flagship commingled fund is the manager's franchise — the deal-flow priority, the marquee deals, the senior partner attention. An SMA can get less of all three if the manager treats it as a side mandate. Institutional LPs typically negotiate this directly into the IMA: dedicated portfolio-management resources, allocation priorities, key-person commitments.
• Quoting "no carry" for SMAs as if it were universal. Some core SMAs run with zero carried interest (the manager earns only the base fee); most value-add and opportunistic SMAs include carried interest, just at lower rates and higher hurdles than the equivalent commingled fund. The carried-interest profile depends on strategy, not on the SMA structure per se.

Related Articles

The articles below cover the surrounding cluster — the LP/GP framework that an SMA inherits from, the commingled-fund vehicles that an SMA is structured against, and the waterfall mechanics that determine the dollar math inside an SMA's economic terms.

Capital Structure — Equity (sibling articles)

• LP/GP Structures, Promote, Catch-Up, and Clawback — The institutional LP/GP framework that an SMA inherits from. The IMA replaces the LPA, but the four-axis structure (capital, control, liability, economics) is the same.
• Fund Structures: Open-End vs Closed-End — The vehicle decision that frames the SMA-versus-commingled tradeoff. Most fund-of-one SMAs are structured as closed-end vehicles; some core SMAs are open-end.
• Management Fees, Carried Interest, and GP Economics — The fee economics that get compressed in an SMA. The 50–100 bps discount this article references is computed against the commingled-fund baseline that article describes in detail.
• Co-Investment Sidecar Economics and Fee Breaks — The other large-LP fee-break mechanism: co-invest sidecars that sit alongside a commingled fund with reduced or zero fees and promote. Often run in parallel with an SMA program.
• JV Equity Structures: Major / Minor Partner — The deal-level analogue: programmatic JVs between an anchor LP and a sponsor GP. Many SMAs execute through programmatic JV structures at the deal level.
• Capital Calls, Distributions, and the J-Curve — Capital-call mechanics inside an SMA mirror commingled-fund mechanics, with bespoke timing rules negotiated in the IMA.
• Syndication Structures: Reg D and Investor Minimums — The retail-accessible cousin of the SMA structure: many smaller LPs in a syndicated vehicle. The structural opposite of the single-LP SMA.

Application

• Fund Reporting (use case) — The bespoke reporting cadence and format that defines the institutional SMA experience. Apers builds reporting against the IMA reporting schedule.
• For Private Equity (audience landing) — Apers for fund managers operating commingled funds and SMA programs alongside each other.
• For Family Offices (audience landing) — Apers for the largest family offices, which increasingly operate SMA-style mandates with their preferred managers rather than committing to commingled funds.

Sources

Institutional data and reference sources cited in this article. Most institutional benchmarking on SMA terms and fee data is paywalled; the references below mix freely available authority documents with paid data products cited by name.

• ILPA Principles 4.0 — the Institutional Limited Partners Association framework for alignment, governance, and transparency in private fund LPAs. The ILPA framework applies to commingled fund LPAs and informs IMA negotiation by analogy; most recent revision late 2024.
• ILPA Reporting Template 2.0 (2024) — the institutional reporting standard for commingled funds. SMA reporting typically extends well beyond the ILPA template scope.
• Preqin Investor Outlook and Term Intelligence — institutional benchmarking on LP allocation trends, separate-account commitment data, and fee benchmarks. Paid data product; cited by name. preqin.com.
• PERE LP Investor Survey — annual survey of institutional LP allocation intent, including SMA-versus-commingled preference data; cited by name.
• Pensions & Investments — trade press coverage of named SMA mandates and awards across U.S. public pensions, sovereign wealth funds, and Canadian pension plans; pionline.com.
• Institutional Real Estate, Inc. (IREI) — trade press coverage of real-estate SMA mandates, manager selection, and allocator trends; cited by name.
• PREA — the Pension Real Estate Association's institutional research on real-estate allocation trends and SMA usage; cited by name.
• NCREIF — ncreif.org for institutional real-estate index data; the NPI and ODCE indices are reference benchmarks for SMA performance reporting.
• Cliffwater LLC — institutional investment consulting and research; SMA structure and fee benchmarking; cited by name.
• Townsend Group and StepStone Real Estate — institutional real-estate consultants whose SMA advisory work covers manager selection, IMA negotiation, and benchmark fee data; cited by name.