The Side Room. From the bridge to the fund: Türkiye's third round of asset turnover, the model we proposed in 2004, and where the hunting lodge sits
Presidential Decision No. 11750 of 5 September 2026 placed the two Bosphorus bridges and eight motorways into the privatisation programme for 30 years, with no transfer of ownership and with "revenue partnership" among the permitted methods. That is the structure we presented to the Privatisation Administration in 2004 on behalf of Farallon, Perry Capital and Hellman & Friedman — an SPV, transfer of operating rights, revenue sharing, title retained by the state — reappearing 22 years later in the state's own decree. In the same letter, the first asset we listed was toll roads and bridges; everything else on the list has since been sold. This was the one left.
This note turns that spark into a working model. Our thesis: Türkiye runs a recurring cycle in which public assets are privatised and leveraged, fall into distress and return to the state, and change hands again. Value is created — and lost — not in ownership but at the moment of turnover. Oaktree and Brookfield are investors of exactly that moment; VMG's job is to be at the table when it happens. We first sat at that table in 2004 — with a proposal to the Privatisation Administration, and with a signature at the SDIF. The documents are attached.
| Lane | Asset | Instrument and size | Counterparty |
|---|---|---|---|
| Lane CBrookfield | Bridges and motorways — 2 bridges, 8 motorways, 2 ring motorways; 30 years; title with the state | Transfer of operating rights and/or revenue partnership; ~$600 m/yr gross revenue base; 2013 reference $5.72 bn (25 yrs); see Section 02 for the range | ÖİB, KGM, Treasury and Finance |
| Lane BOaktree | TVF portfolio — TRY 12.7 tn of assets, TRY 2 tn of equity, 32 companies, 188,631 employees: Türk Telekom, Turkish Airlines, the state banks, Borsa İstanbul, Eti Maden | AMC-type value-management partnership: minority stake + management rights + incentive fee above a baseline valuation; $100–500 m per asset. The skeleton is the one signed with the SDIF in December 2004 (80/20, tiered profit share, public representation on the board) | TVF Board; the addressee of the VMA report series |
| Lane AOaktree | Corporate bridge/rescue credit — large, solvent, liquidity-squeezed groups (Zorlu simulation) | Share-pledged senior/structured credit; $50–300 m; exit through asset monetisation | Group owners; state banks (Stage-2 credits) |
| Lane A′Oaktree | Exhausted PE/VC portfolios — 2007–2014 vintage funds, no exits, lives extended; USD NAV / lira cash | GP-led secondary, continuation vehicle, preferred equity or purchase of portfolio debt; at a discount to NAV; $50–200 m. Filter: "no market" is separated from "bad asset" | GPs; DFI LPs (EBRD, IFC, EIF) |
Out of scope (per the distinction in Desk Note No. 1): companies under court-appointed trustees and assets seized by the SDIF (TMSF) on terrorism grounds — excluded from every lane for contested title and LP-mandate risk. Clean capital's absence from those auctions is information, not noise.
The one asset the state did not sell tripled its revenue and is now on the table. The 2004 formula with 2026 numbers: $600 m × 30 years = $18 bn of nominal gross revenue. That figure is revenue, not profit; where a concession price settles is decided by the discount rate and operating cost:
| Assumption (gross revenue $600 m/yr, 30 yrs) | Present value | Note |
|---|---|---|
| 10% discount, no growth | ~$5.7 bn | The 2013 bid ($5.72 bn, 25 yrs) sits on this street |
| 10% discount, 3% revenue growth | ~$7.4 bn | Same street as the $8 bn proposed in 2004 |
| 8% discount, 3% growth | ~$9.6 bn | With an investable currency/guarantee framework |
| 12% discount, no growth | ~$4.8 bn | Currency risk entirely with the operator |
Illustrative; calculated on gross revenue. Maintenance and operating cost, the margin of the KGM service contract, the tariff formula and the currency mechanism move the price down or up. The model will be calibrated with actual traffic and revenue series obtained from ÖİB.
Twenty-two years on, the concession price still sits on the $5–8 bn street; revenue has tripled and the term has fallen from 40 to 30 years. The difference is that whoever bought 40 years of $200 m in 2004 would today buy 30 years of $600 m at the same price. That is the sentence we take to Brookfield.
The model was born in a parliamentarian's office, in an unplanned twenty-minute conversation. The Ministry of Public Works was looking for funding for 12,000 km of dual carriageway promised at the election; the Treasury could not breathe under the IMF programme; domestic capital was waiting to "receive", not to "give". On the table sat what previous governments had left: the two Bosphorus bridges and the Edirne–İskenderun toll motorway network — roughly $170–200 m of annual revenue, growing 8–10% a year.
The formula was built on the spot: annual revenue × years = cash resource. An $8 bn need divided by $200 m of annual revenue: 40 years. The structure: an SPV owned jointly by the Highways Directorate (KGM) and the funds; KGM transfers the operating rights of all toll roads to the SPV for 40 years; the funds pay $6.4 bn in cash for 80%; the state builds its dual carriageways with the proceeds; the SPV contracts maintenance and operations back to KGM under a fixed-margin service agreement; new toll projects also run through the SPV. Title stays with the state. Funding outside the IMF programme — called a "lifebuoy" that day.
The funds were in Ankara the next day. Outcome: model and IRR accepted; a 3–6 month due diligence funded by the funds (~$8 m); an MOU if the data verified within ±5%. We were received at three ministries. Then the matter moved to the big room.
When the host said the matter would proceed "by tender, designed around the bid", the table closed — on one side, ministers had confirmed minutes earlier that an SPV partnership carried no tender obligation; on the other, the American funds' compliance standards would not permit the sentence "a tender designed for you" even to be translated. The lesson is structural: the structure is not taken into the big room until it is finished in the side room; and when it goes in, it must be tender-proof. Section 06 is that lesson made institutional.
The letter of 19 October 2004 (Appendix A) extended the same logic to the whole Privatisation Administration portfolio: an Asset Management Company buys 20% of each asset, 80% stays with the state, value is optimised under professional management, and on sale to a strategic buyer 20% of the amount above the baseline valuation is the incentive fee. First on the list: toll roads and bridges.
Same day, same team, a second letter: 19 October 2004, from Perry, Farallon and Hellman & Friedman to SDIF Chairman Ahmet Ertürk. The subject was the corporate non-performing loan book left to the SDIF by the 2001 banking crisis — roughly $4 bn of principal (about $23 bn in claims with accrued interest and penalties). The structure was built on the Asset Management Company skeleton the SDIF itself had outlined: investors 80% ($320 m in cash against $3.2 bn face), SDIF 20% ($800 m face contributed at a notional $80 m) — ten cents on the dollar; on top, a profit share in the SDIF's favour of 15% of total distributions above $1 bn and 30% above $1.5 bn. An all-cash offer with no financing contingency; target closing 31 March 2005.
This time it went all the way to signature. An exclusivity letter on 3 December 2004; a confidentiality undertaking on 21 December; a Memorandum of Understanding on 22 December 2004 — Ahmet Ertürk for the SDIF, Richard Perry and Alp Erçil for Perry, William F. Duhamel for Farallon. Six weeks of due diligence from 3 January 2005, exclusive negotiation over the SDIF's corporate claims, SDIF representation on the board including the audit committee, restructuring of loans where economic (Exhibit C). It is the document in which a Turkish state institution accepted the KAMCO model at signature level. After the bridge-and-motorway meeting in the big room, this table closed too; the model was not rejected — the table was removed.
The 2004 SDIF book consisted of corporate loans taken over from banks in the 2001 crisis under Law No. 4743 — uncontested title, clean-capital buyers. The SDIF assets excluded from every lane of this note are those seized after 2016 on security grounds. The two inventories carry the same institution's name; they are not the same thing.
| 2004 — proposed model | 2026 — Presidential Decision No. 11750 | |
|---|---|---|
| Assets | Bosphorus + FSM bridges, Edirne–İskenderun axis | 15 July + FSM bridges, 8 motorways, 2 ring motorways (Niğde–Pozantı, Gaziantep Ring, Bursa Ring added) |
| Ownership | Stays with KGM; operating rights to the SPV | "No transfer of ownership" |
| Method | Revenue-share partnership, concession rights | "Transfer of operating rights, lease, rights in rem short of ownership, revenue partnership model" — as a whole or in groups |
| Term | 40 years | 30 years |
| Revenue base | ~$200 m/yr → $8 bn need | ~$600 m/yr (2025) → $18 bn nominal over 30 years (Treasury: "revenue, not profit") |
| Authority | ÖİB (letter addressed to President Metin Kilci) | ÖİB mandated to prepare and execute; KGM operates until completion |
| Timeline | "Process within 30 days" | To be completed by 31 December 2031 |
| External constraint | IMF programme; concession proceeds fell outside it | None |
In between: the December 2012 tender drew a $5.72 bn high bid for 25 years (Koç–UEM–Gözde), cancelled in February 2013 as "insufficient"; new roads then moved to Treasury-guaranteed BOT/PPP. The 2026 decision and the ÖİB statement ("the priority is not revenue but service quality, the pace of maintenance and investment, and operating efficiency") point back to the 2004 logic — manage the value, then sell.
Every asset we proposed in the letter (except the toll roads and Manavgat) was sold in the following years. Below, each sale price is converted to a 100% equivalent and set against today's market value or a registered subsequent transaction. The table is the model's 22-year natural experiment.
| Asset | Sale | Price → 100% eq. | Today / next hand | Multiple |
|---|---|---|---|---|
| Mey İçki (Tekel spirits) | 2004, 100%, domestic consortium | $292 m | $810 m (TPG, 2006) → $2.1 bn (Diageo, 2011) | ×7.2 |
| Port of Mersin | 2007, 36-yr concession, PSA–Akfen | $755 m | $2.17 bn (implied by IFM's 40% purchase, 2017); $453 m capex in between | ×2.9 |
| Tüpraş | 2006, 51%, Koç–Shell | $4.14 → $8.1 bn | ~$16.5 bn market cap | ×2.0 |
| Erdemir | 2006, 49%, OYAK | $2.77 → $5.6 bn | ~$5.6 bn | ×1.0 |
| Electricity distribution (Enerjisa's 3 regions) | 2008–2013 (Başkent, AYEDAŞ, Toroslar) | $4.18 bn | ~$2.8 bn (Enerjisa Enerji, incl. retail) | ×0.7 |
| Türk Telekom | 2005, 55%, Oger | $6.55 → $11.9 bn | ~$4.6 bn; TVF bought the 55% back in 2022 for ~$1.65 bn | ×0.4 |
| Petkim | 2008, 51%, SOCAR | $2.04 → $4.0 bn | ~$1.1 bn | ×0.3 |
| Tekel tobacco | 2008, BAT | $1.72 bn | not listed | — |
| Bridges and motorways | not sold | 2004 revenue ~$200 m/yr | 2025 revenue ~$600 m/yr | ×3 (revenue) |
Market values as of September 2026 at ~48 TL/$. Total privatisation proceeds for all 21 distribution regions were roughly $13 bn. The "today" column excludes dividends; a buyer's total return adds the dividend stream (material for Tüpraş, Türk Telekom and Erdemir). Interim "+5 year" values are given only where a transaction evidences them.
Figure 1 — Assets on the 2004 list: value today or next-hand transaction as a multiple of sale price (100% equivalent, USD). Gold: value created on the second sale. Navy: below the first-hand price in dollar terms.
Türkiye's public assets follow a life cycle: privatise → leverage → distress → the state buys back → privatise again. In the first round we sat at the SDIF's table and signed; the third round is starting now.
The "proven in South Korea" reference in the 2004 letter is post-1997 KAMCO: buy distressed assets at roughly a third of face, manage and resolve, sell, and recover more than was paid. Newbridge's Korea First Bank operation belongs to the same school. Korea's sovereign wealth fund, KIC, was founded in 2005 for a different purpose.
The scale is audited, not assumed: the fund's 2024 report shows TRY 12.7 trillion of assets (up 36%), TRY 2 trillion of equity and TRY 371.4 bn of profit for the period across 32 companies in seven sectors — with no disposal programme, no second-hand sale discipline and no external management partner (Exhibit D). Ankara has already been asked to appoint one: our March 2026 brief to the Presidency proposes exactly that mandate (Exhibit E). Türkiye also knows the model: the 2004 SDIF letter opens with "the AMC structure you outlined" — the SDIF had drawn the skeleton itself. The translation today: the TVF holds the KAMCO inventory; it does not have the KAMCO operator. Reabsorbed assets have been gathered onto one balance sheet, but the "manage → optimise value → sell on the second hand" step has brought in neither foreign capital nor an external management partner. The AMC structure proposed to ÖİB in 2004 is stronger when proposed to the TVF today, because the inventory already sits on one balance sheet and the counterparty is one.
Oaktree is not a growth fund but a cycle fund: "buying resilient cash flows at panic prices" works precisely at the transition from reabsorption to the third round. Brookfield is one of the world's largest private owners of long-duration, revenue-producing infrastructure — toll roads, ports, distribution — with motorway portfolios in India, Brazil, Chile and Peru. The two are one group. The three lanes defined in Türkiye Desk Note No. 2 map onto three stations of the cycle: Lane A onto the moment of distress, Lane B onto the reabsorbed inventory, Lane C onto third-round concessions.
A lodge in Ankara, next door to the Ministry of Finance: front garden facing the private sector, back garden facing the state. The side room of 2004, which was an accident, becomes an institution in 2026.
Each exhibit opens as a reading page in English and Turkish, with the original document embedded and downloadable at the foot of that page.
| Date | Document | Content |
|---|---|---|
| 7–8 October 2004 | Istanbul meetings | Perry, Farallon and H&F team with the SDIF; the SDIF presents its own AMC skeleton |
| 19 October 2004 | Indication of interest — Perry / Farallon / H&F → A. Ertürk | ~$4 bn face of corporate NPLs; investors 80% ($320 m cash / $3.2 bn face), SDIF 20% ($800 m face at a notional $80 m); profit share 15% above $1 bn, 30% above $1.5 bn; all-cash, no financing contingency; target closing 31 March 2005 |
| 3 December 2004 | Exclusivity letter — Perry / Farallon → SDIF | Exclusivity to 31 March 2005; countersignature block for the SDIF |
| 21 December 2004 | Confidentiality undertaking | Perry and Farallon as "Bidder"; Istanbul courts |
| 22 December 2004 | Memorandum of Understanding — SDIF · Perry · Farallon | Signed by Ahmet Ertürk (SDIF Chairman), Richard Perry, Alp Erçil, William F. Duhamel. AMC 80/20 under Law No. 4743; six-week due diligence from 3 January 2005; exclusive negotiation; SDIF representation on the board and audit committee; incentive arrangement for senior management; restructuring where economic; target closing 31 March 2005; costs borne by Perry/Farallon |
| 2005 | — | Process halted after the bridge-and-motorway meeting; the Türkiye file closed after 16 months and ~$2 m of costs |
The original English letters, Turkish translations, the signed MOU and the undertaking are on file. Adviser: Mehmet Narin, Asia Strategic Advisory Partners Ltd.