Industry

What a Korean real estate trust company actually sells

A Korean 부동산신탁회사 (real estate trust company) sells a legal container. Under the Trust Act, a settlor transfers legal title in a property to a trustee, which then manages, develops, or disposes of that property for the benefit of a named beneficiary and takes a fee for doing so [1]. Nothing about that is unique to Korea. What makes it an industry is that the trustee must hold a licence: real estate trust is a category of financial investment business under the Financial Investment Services and Capital Markets Act, licensed and supervised by the Financial Services Commission under a single licence unit, 4-121-1, covering movables, real property, and real-property-related rights [1].

The economic organising principle follows from the transfer of title. Once the land sits in the trust, the Trust Act separates it from the bankruptcy estate of both the settlor and the trustee, and Article 22 bars compulsory execution against it. That gives the structure four functions the filings name explicitly: custody of development cash so pre-sale proceeds cannot be diverted, bankruptcy remoteness for the land and building, dispute mediation among the parties to a development, and conversion of the property into divisible, tranched beneficiary certificates [2].

So the customer buying a trust is rarely buying property management. A developer buys credibility with a lender; a lender buys a controlled disbursement account and an enforceable security route; an apartment buyer paying instalments years before completion buys protection against the developer failing. The trust company is paid by the developer or landowner, usually out of pre-sale receipts, but the party whose comfort actually closes the deal is the project-finance lender.

Price is a function of how much risk the trustee absorbs

The industry publishes its price list. Korea Asset In Trust's base fee schedule shows the whole logic of the business in one table: the same building, the same trustee, and a fee that moves by seven times depending on who is on the hook for the construction money.

Source: FY2025 Annual Report, II. 사업의 내용 — 기본보수 산정 기준 및 요율 [3]. Rates are the published base schedule; actual contracts are negotiated.

The filings describe the two ends of that ladder in the same terms. A loan-type land trust is the arrangement in which the trust company directly raises the construction cost, carrying funding risk in exchange for a larger fee; a management-type trust leaves funding with the settlor or contractor, so the trustee has no advance at risk and earns correspondingly less [4]. The competing listed trust, KOREIT, words it the same way and adds the consequence: in loan-type work the trustee bears loss risk on money advanced in its own name [5].

The collateral trust sits at the other end and earns 0.2%, because it competes against an alternative that costs nothing to the trust industry — a conventional mortgage. The filings set out why lenders pay anyway: a collateral trust needs no pre-attachment to exercise subrogation, is not caught by the debtor's rehabilitation proceedings, is realised by the trust company's own public tender rather than court auction, and blocks the priority that wage claims would otherwise take [4].

Where the assets sit and where the money is made

Two different scales are quoted for this industry and they measure different things. Assets under trust (수탁고) is trust principal accepted; operating revenue is fees plus interest. The filings warn that the correlation between the two is low, because each trust type prices off a different base [6].

Scope: Korean real estate trust business only, as at 31 December 2025, sourced by both listed trusts from the Korea Financial Investment Association. Assets under trust ₩560tn in total, of which ₩457tn at the dedicated trust companies [6] [7]; operating revenue, licensed-company count and headcount [8] [9].

Banks, securities firms, and insurers may also act as real estate trustees, but only for the passive products. Of the ₩560tn of real estate trust assets outstanding at the end of 2025, banks held ₩89.9tn, securities firms ₩7.9tn, and insurers ₩4.6tn — and none of them held a single won of land trust or pre-sale management trust, both of which sit entirely with the 14 dedicated companies [6]. Development work is the licensed specialists' preserve; the banks compete only where the trust is a security wrapper.

Within the specialists, the split between where assets sit and where fees are earned is stark.

Source: derived from the FY2025 Annual Report tables of assets under trust by type [6] and industry trust fees by type [8]. Fee yield is the ratio of the two; the filings caution that trust principal is not the fee base for every product.

Land trust is 23% of the assets and 80% of the fee pool. Collateral trust is 74% of the assets and 19% of the fees. That ratio is the industry's central fact: the passive, growing, low-risk book generates roughly 3 basis points of fee on assets, while the development book generates roughly 44. Every strategic argument inside these companies is about how far up that gradient to climb.

The profit pool along the development chain

Source: assembled from the product and fee descriptions in the FY2025 Annual Report [3] [4] [10], and the HUG-guaranteed redevelopment disclosure in the contingent-liability note [11].

Two features of that chain matter more than the rest. First, the trust company's fee is charged on pre-sale proceeds, so revenue is recognised over the life of a project — the filings put the conversion of a signed mandate into revenue at three to four years [9]. Orders won in a boom therefore land as revenue into a downturn, and orders lost in a downturn hollow out revenue years later. Second, in redevelopment and completion-guarantee work the trustee stands behind other parties' performance, which converts a fee business into a contingent credit business. That is the industry's principal structural conflict, and the last four years are a live test of it.

The arena: fourteen licensed companies, and what they earn

Fourteen companies hold the licence. Four date from the 1990s, seven from the 2000s, and three were licensed in 2019 — 대신자산신탁 in July and 신영부동산신탁 and 한국투자부동산신탁 in October [8]. None of them is a REIT; all are supervised financial institutions.

Sources: operating revenue, fee income, staff and fee per head for FY2025, Korea Asset In Trust FY2025 Annual Report [9] [12]; equity and net income on a separate-financial-statement basis at 31 December 2025, KOREIT FY2025 Annual Report [13]. Share is derived from operating revenue.

Five of the fourteen lost money in 2025, and the losses are not small relative to the capital behind them: Woori Asset Trust lost ₩220.6bn against ₩239.4bn of equity, Kyobo ₩149.6bn against ₩358.9bn, and Mugunghwa ₩93.3bn against ₩25.1bn of remaining equity [13]. Capital, not revenue, is what separates the companies now — and the two largest capital bases, KOREIT at ₩892.9bn and Korea Asset In Trust at ₩867.2bn, belong to firms with mid-table fee income.

Fee productivity per employee ranges from ₩577m at KB to ₩148m at Shinyoung, against an industry average of ₩352m [9]. This is the metric the filings themselves propose for comparing competitiveness, on the reasoning that fee income is a return on people rather than on assets, and that interest income should instead be read against each firm's equity [9]. One caution: the equivalent table in the FY2021 filing ran on operating revenue rather than fee income, and its column totals ₩1,605.0bn rather than ₩909.5bn, so the two vintages of "fee per head" are not comparable [14].

The 2019 licences reshaped the share table

Source: derived from the industry market-share tables for 2017-2021 [14] and 2021-2025 [12]. Top-4 is the four largest companies by operating revenue in each year; entrants are Daishin, Shinyoung, and Korea Investment, all licensed in 2019.

In 2017 the four largest companies took 60.5% of industry revenue and the incumbent leader, Korea Asset In Trust, took 20% on its own [14]. By 2025 the top four took 43.5% and the three companies licensed in 2019 took 10.8% between them [12]. Two effects are tangled here and the filings do not separate them: new licences added supply into a market that stopped growing, and the ranking also churned for reasons of ownership — Asia Trust, sixth by revenue in 2021, appears in the 2025 table as Shinhan Asset Trust and leads it. The record of who took share from whom belongs to Competition; what the arena shows is that a licensing decision made in 2019 is still redistributing revenue in 2025.

The industry's revenue changed character, not size

Sources: 2017-2021 from the FY2021 Annual Report industry table [15]; 2021-2025 from the FY2025 Annual Report industry table [8]. Both are compiled from Korea Financial Investment Association filings on a separate-financial-statement basis.

Industry operating revenue in 2025, at ₩1,664.1bn, was 3.7% higher than in 2021 and 4.0% below the 2022 peak [8]. On the top line, nothing much happened. Underneath, the composition inverted. Trust fees fell 43% from ₩1,039.5bn in 2022 to ₩589.6bn in 2025, and land-trust fees specifically fell from ₩894.4bn to ₩468.6bn. Over the same three years, interest on loans to trust accounts rose from ₩103.4bn to ₩388.9bn — 3.8 times [8].

That second line is not a new product. Interest on trust-account loans accrues when the trust company advances its own money into a project that cannot fund itself — either because the structure always required it, as in loan-type land trust, or because something went wrong. Revenue of that kind grows fastest when projects stall. Industry-wide, roughly ₩1 in ₩4 of 2025 operating revenue was interest on the industry's own advances into the schemes it manages, against ₩1 in ₩16 in 2022.

A caveat on these tables. Both listed trusts publish the same Korea Financial Investment Association series, and their versions disagree in places: 2024 trust fees are ₩784.8bn in the Korea Asset In Trust filings and ₩764.8bn in KOREIT's, and the itemised fee lines do not always sum to the fee-income total shown above them [8] [16]. The series is self-reported and revised; differences of one or two percent between filings should not be read as signal.

The structural conflict: a fee business that guarantees other people's performance

Three of the industry's products put the trustee behind a third party. In loan-type land trust the trust company funds the build. In completion-guarantee management trust, it promises to finish the building if the contractor fails and to compensate the lenders if it cannot. In HUG-guaranteed redevelopment trust, the standard agreement makes the trust company, as project implementer, the party carrying the completion obligation [11].

Korea Asset In Trust named the risk in its FY2022 filing, writing that as the risk of project-finance distress widened, the risk factors in completion-guarantee products came into focus, because a contractor falling behind schedule or a long unsold period increases the trust company's exposure [17]. The size of that exposure, and its subsequent unwind, is disclosed year by year.

Sources: FY2022 and FY2023 columns from the FY2023 Annual Report contingent-liability note [18]; FY2024 from the FY2024 note [19]; FY2025 from the FY2025 note [11]. Completion-guarantee sites are the 책임준공토지신탁 line; the total adds HUG-guaranteed and other redevelopment trust obligations.

At the end of 2022 one company stood behind ₩1,126.5bn of drawn lender debt across 25 projects, of which 18 were completion-guarantee sites carrying ₩600.5bn [18]. By the end of 2025 the same disclosure covered five projects and ₩126.3bn [11]. The intervening year shows what the obligation means in practice: at the end of 2024 one of the seven remaining sites had passed its contractor's completion deadline, with ₩80.3bn of loans outstanding against it, and the trust company had applied for temporary occupancy approval on part of the building to meet its own deadline [19].

The obligation is disclosed but not provisioned. The filings state each year that the loss cannot be reliably measured and so no amount is recognised, with the exposure monitored project by project instead [11]. Where a completion promise is actually performed, the money appears elsewhere — as advances (대지급금) inside the loan book.

Where the exposure lands on the balance sheet

The visible cost of the cycle shows up in loans to trust accounts. At the end of 2025 Korea Asset In Trust carried ₩802.6bn of gross trust-account loans against ₩161.2bn of allowance — 20.1% — up from ₩819.1bn against ₩107.8bn, or 13.2%, a year earlier [20]. These claims are not unsecured lending in the ordinary sense: the filings note that trust-account loans and advances are recoverable from the trust estate under the Trust Act's expense-reimbursement right [21]. Recovery therefore depends on the project selling, which is the same variable that determines the fee.

Source: FY2025 Annual Report, 재무건전성 — 자산건전성현황 and 고정이하자산비율 [22]. Classification is of consolidated assets subject to soundness classification, and the filings note that trust-account loans are conservatively classified using project balance, sales rate, and construction progress.

Two constraints bound how far this can run. Regulation 3-26 of the Financial Investment Business Regulation requires a type-3 financial investment business to hold a net operating capital ratio of at least 150%, and Regulation 3-41 requires a real estate trust company to hold a won liquidity ratio of at least 100% [23] [24]. Korea Asset In Trust reported 363% at end-2025, having dipped to 284% at end-2024 [23] [24]; KOREIT reported 230.1%, down from 306.3% two years earlier [25]. The prudential ratio is what converts a bad property cycle into a binding limit on how much new development work a trust company may write.

Where the industry sits in its cycle

The filings' own framing is that real estate trust follows the property cycle, but not simply. Recessions reduce the volume of development while raising the demand for the safety features the trust structure provides — so the industry describes itself as exposed to both quantitative and qualitative swings in property activity [2]. That claim has held up in an uncomfortable way: assets under trust at the dedicated companies rose every year from ₩342.4tn in 2021 to ₩457.5tn in 2025 while trust fees fell 43% from the 2022 peak [6] [8]. Volume grew; price and mix collapsed.

The outlook language in the filings moved in three steps. In FY2021 the development market was described in terms of domestic policy to cool an overheating property market and externally rising protectionism [26], and continued demand for loan-type land trust was expected regardless of property-market volatility [15]. In FY2024 it was a global rate-tightening cycle contracting the market [27]. In FY2025 it is project-finance distress and macro uncertainty together, with loan-type land trust described as affected by weakened investment sentiment [2].

The transmission mechanism is set out plainly in the same filing's divisional review: through 2025 the trust industry faced slowing property activity, a contracting project-finance market, and rising construction costs; development starts fell and pre-sales at some sites were delayed, which cut trust and management fees while the deteriorating soundness of trust-account assets pushed provisions up [36]. The forward risk named there is specific — the workout and recovery of provincial loan-type land trust projects may run slow, with government-led court and public auctions gradually restructuring distressed sites while non-capital-region and non-residential uncertainty persists [36].

This is not the industry's first shakeout. The two trust companies that ran Korea's land-trust market in the 1990s — Daehan Real Estate Trust, later renamed Koret Trust, and Korea Real Estate Trust — were both wound up, and Korea Asset In Trust was built in 2001 out of the assets, staff, and systems it acquired from them [28]. The mechanism that killed them was the same one visible in the loss column today: a development trust that funds construction is a lender to a property market, and in a bad enough market a lender to a property market can lose its capital. What differs this time is that the losses are concentrated in the smaller and newer licensees rather than the largest.

Three currents to carry into the other tabs

The fee pool is migrating from greenfield development to redevelopment

Both listed trusts are writing more of their new work in 정비사업 — reconstruction and redevelopment of existing estates, opened to trust companies by the 2016 amendment to the urban-maintenance law, under which the trust company acts as the scheme's operator or agent in place of the owners' association [29].

At Korea Asset In Trust the shift is a four-year trend: redevelopment land-trust fees contracted rose from ₩4bn in 2022 to ₩43bn in 2025, while conventional loan-type mandates fell from ₩38bn to ₩16bn.

Source: 1Q 2026 investor presentation, Trust business — contracted trust and REITs fees [30]. Contracted fees are amounts agreed in the year, recognised as revenue over the following three to four years.

KOREIT's disclosed order mix moves in the same direction but far more erratically, which is itself a feature of the industry: individual mandates are large enough that a single year's mix says little.

Source: KOREIT FY2025 Annual Report, 사업군별 수주비중 추이 (5개년) [31]. Other covers management-type land trust, non-land trust, consulting, and agency work.

The economics of the switch are set out in the fee card: redevelopment trust prices at 2.5% of pre-sale proceeds against 0.5% for plain management-type work, and the trustee carries operator obligations rather than funding risk [3]. Redevelopment schemes also run on long approval clocks, so the fees contracted in 2024 and 2025 sit several years from recognition.

Revenue is becoming more interest-like and less fee-like

The industry earned ₩388.9bn of interest on its own advances into trust accounts in 2025, against ₩589.6bn of trust fees [8]. At Korea Asset In Trust specifically, trust-account interest was ₩51.5bn in 2025 against ₩41.4bn of land-trust fees — the interest line is now the larger of the two [32]. Two readings of the same number are available and the evidence does not yet separate them: interest that accrues on a performing loan-type project is contracted return on capital deployed, while interest accruing on a stalled project is an accounting entry whose cash realisation depends on eventual sales. The 20.1% allowance now carried against the trust-account loan book is the company's own estimate of how much of that balance will not come back [20].

The property recovery is regional, and the trust book is not

Management's own account of 2025 draws the line explicitly: the capital region recovered gradually while the non-capital regions where its main project sites are concentrated kept a high share of unsold housing, deepening a regional split that delayed sales and impaired project viability [33]. For 2026 the same filing expects a full market recovery to stay limited, citing rate volatility, possible changes in government housing policy, and rising construction costs, while stating that provisioning taken across 2024 and 2025 should ease further credit costs [33]. The stated strategic response — diversify regions and products, tighten underwriting on loan-type land trust and other high-risk categories, and push into redevelopment and non-land trust — reads as the same response the whole industry is attempting [34].

What the evidence here cannot settle

Three limitations are worth carrying forward rather than glossing.

There is no independent market-size series. Every figure above traces to the Korea Financial Investment Association's member filings, republished by the trust companies themselves, and the two listed trusts publish versions that differ in detail [8] [16]. There is no third-party market study in the record.

Only one listed pure-play comparable exists. Of the six companies in this run's automatically assembled peer set, KOREIT is the only other listed Korean real estate trust; LF Corp and Hana Financial Group are listed holders of Koramco Asset Trust and Hana Asset Trust rather than trust companies themselves, SK D and D is a developer, and Hulic and TOC are Japanese property companies retained from the 2016 IPO valuation screen [35]. Peer economics in this tab therefore come from the two Korean trusts' own filings, which disclose all fourteen licensees, rather than from the listed peer set.

Product-level profitability is not disclosed anywhere. Korea Asset In Trust's management discussion splits operating profit between its trust division and its capital-finance subsidiary [36], and KOREIT declares three divisions but reports only one because trust exceeds 90% of revenue [16]. Neither breaks profit out by trust product, so the margin on land trust against collateral trust against REIT management cannot be observed — only revenue by product. The fee-yield table above is the closest available substitute, and it is a revenue measure, not a margin.

Neither can these records fix the boundary between contracted return and deferred loss inside the interest line, nor date the trough of the fee cycle. Those are questions for the chapters, which have the company's quarterly record and the analyst estimates to work with; this tab records the arena they take place in, and History records how this particular company arrived in it.