The $104 billion Oregon fund returned 3.5% for the quarter against a 6.6% peer median, and private equity was the only major class to decline. Over one year the private book returned 1.1% against a benchmark at 25.4%. The structure that produced top decile results since inception is now the reason the fund trails its peers.

Aon modeled Oregon's liquidity under a deep recession followed by prolonged stagnation and found illiquid exposure rising from 53% to 77% of the fund. Its conclusion is that liquidity remains sufficient in every scenario tested. The number worth sitting with is what sufficient looks like at 77%, and how much of the gap comes from today's overweight.

Oregon holds close to 50% of a $104 billion fund in private markets against a 40% target, and its consultant says plainly that the position now limits what the policy portfolio can be changed to. The recommended fix raises the illiquid target rather than cutting the exposure, which is what a decade-old overweight does to an allocation study.

Meketa's final asset-liability study gives the $104 billion Oregon fund a 7.5% Credit allocation where policy currently has none, funded mainly by cutting fixed income from 25% to 20%. Private equity comes down a point. Credit stops being a line item inside the bond book and becomes a class the council allocates to directly, with its own benchmark and its own range.

The $65.4 billion Illinois Municipal Retirement Fund's private real assets sleeve returned 2.26% net against a 1.36% benchmark, but a $172 million farmland fund inside it returned 66.23% for the quarter while sitting down 8.34% over one year and 3.20% annualized over ten. The $5.26 billion real estate book that holds the actual capital has returned negative 0.09% annualized over three years and sits 2.2 points below target.

Alternatives contributed 81 basis points of manager effect against 62 basis points for the entire $65.4 billion Illinois Municipal Retirement Fund. Private equity returned 4.75% against a benchmark of negative 2.65%, because both the marks and the index are lagged one quarter and those two periods moved in opposite directions. Over three years the same construction turns the advantage into a 708 basis point shortfall.

Callan's second quarter review of the $65.4 billion Illinois Municipal Retirement Fund records a 0.07% net return on private credit against a 2.32% benchmark, while the plan's own bank loan and high yield mandates returned 1.70% and 2.13%. Over five years the sleeve has returned 4.37% against 7.94%. The plan is also measured against a benchmark that assumes a private credit program roughly twice the size of the one it holds.

Connecticut's private equity program returned 6.5% over three years while its Russell 3000 plus 250 basis points policy benchmark returned 18.8%. The plan's answer is that a public market index measures a long-term asset class poorly over short horizons, an argument that is sound and also convenient.

Connecticut returned 15.1% for the fiscal year, and global equity produced 12.3 points of it. Private equity and private credit contributed 0.6 points each, from an allocation holding 51% in global equity against a 37% policy target and underweight every private asset class.

Connecticut will make no further real estate commitments this calendar year and is cutting its target allocation to 8% from 10%. The pause comes while the plan's own review reports outperformance against NCREIF at every measured horizon, which makes this a capacity decision rather than a market call.

Buyout managers ended last year holding nearly $4 trillion of unrealized value across about 32,000 companies, with entry multiples at 14.0x. Connecticut's review calls the backlog a constraint on manager capacity and cash flows, and the plan has cut its own venture allocation to a third of the benchmark weight.

Data infrastructure is 17% of Connecticut's infrastructure NAV, and data centers are roughly 8% of its real estate book, reported inside an Other category. The plan has named cross-portfolio data center concentration an increased priority, and no combined exposure figure appears anywhere in the materials.

Connecticut's private credit portfolio holds 63% in senior credit against 26% for the Hamilton Lane benchmark, and it outperformed at three, five and 10 years while carrying that overweight. The plan is committing most of this year's private credit capital back into the segment absorbing retail redemption pressure.

Connecticut's $73.7 billion pension system runs its private markets program through a short list of managers. The top 10 relationships hold about 88% of infrastructure exposure, and the private equity roster has fallen to 34 managers from 50. Every strategic review names re-ups with existing managers as the preferred use of capital.

OCERS sold 17 legacy fund-of-funds and energy positions for $231.4m, describing them as having limited remaining upside. Those positions are marked at 2.0x to 2.7x and have returned $1.38 per dollar called. The programs receiving the proceeds have returned 19 cents and 3 cents.

Active infrastructure funds account for 32.4% of OCERS' real assets contributions and 9.8% of its distributions. Policy now targets 4% infrastructure against 2% energy, and every recent commitment has gone to infrastructure or energy transition.

OCERS' opportunistic real estate sleeve returned 5.0% gross and 1.5% net over five years, losing to the core index at twice the leverage. Value-added, on the same fee logic, beat that index by more than two points.

Townsend's glossary caps core real estate leverage at 30%. Three funds OCERS classifies as core run at 50.6%, 53.2% and 61.7%, and the two most levered are the two worst performers in the sleeve.

OCERS publishes gross and net returns for 198 alternative investment funds side by side. The median fund gives up 3.6 points of annual return to fees, 23 give up more than ten, and eight turn a positive gross return into a loss.

Two DBL Partners vehicles hold 75.8% of OCERS' co-investment net asset value at 19.2x and 13.3x, and neither has distributed a dollar. Strip them out and the program falls from 4.9x to roughly 1.5x.