"CRI Manhattan has 167% coverage in collateral. If there's more collateral than debt, why is that a problem?"
Because paper coverage and money in the account are different things. Early due means that the debtor is no longer honoring the combined flow — the fund no longer receives the coupon and now needs execute the guarantees to recover the capital. This process is legal, takes months or years, and the 167% are an estimate of collateral value before to go to auction, with cost of execution, defiance and time running against. Meanwhile, the asset remains marked on the market by pressing on the equity value. High coverage is what separates a Manhattan CRI from a total cap — that's relative good news. But it's not a cashier, and it's not today.
O VGHF11, Valera Hedge Fund, is one of the most pulverized "paper" FIIs in Brazil: 378 thousand unit holders, R$ 1,38 billion equity and a multistrategy portfolio that mix units from other FIIs (56,7% PL), CRIs (28,8%), stakes in real estate SPEs (15%) and a fringe of FIDCs and shares. In May, we lifted the fall of ~19% from unit in nine studs and the B3 trade demanding explanations — fact that you can revisit in our analysis of may/2026. . At that moment, the discussion was of price: Was the market right or exaggerated by overthrowing the unit?
The May management report changed the axis of the conversation. Now it's not about screen oscillation — it's about ground. . And what the document shows is uncomfortable: a relevant credit was for execution, the equity value has been in continuous decline for five months and the mattress that supported the dividends practically disappeared. Let's go in pieces.
The Manhattan 161S: what actually happened
The CRI Manhattan 161S is a backed-up credit operation in a SPE of Manhattan Constructora. In June 2026, the fund manager decreed the anticipated maturity the operation and began the execution of guarantees. Translating: the debtor has failed to comply with a contract clause (delay, breach of covenant or deterioration of the real estate operation behind), and the Valera has pulled the trigger that makes the entire debt enforceable immediately, leaving for the taking of guarantees.
The fund manager shall inform that the guarantee cover is 167% — i.e. real estate and receivable collateral data would be worth 1,67 once the debtor balance — and expects to recover the value in full. This is the reassuring part, and it matters: a CRI with 167% of collateral is in a very different situation than a chirophant credit without collateral. In theory, there is room for absorbing auction gap and costs and still return the main one.
The problem is what that phrase hides. Execution of real estate guarantee in Brazil is slow: between notification, possible judicial dispute, disposal of assets and effective receipt, it can be many months — sometimes years. All this time, three things happen at the same time: the background does not receive the coupon that the CRI paid, the asset follows in the portfolio subject to rescheduling, and the recovery of the 167% is a expectation, not a closed number. If the real estate market of the SPE gets worse, or if the execution faces litigation, that mattress may shrink. It is the highest individual credit risk of the portfolio today, and the lack of provision so far does not mean that it does not come — it means that the fund manager still bets on full recovery.
The mathematics of equity value
The VP by unit does not fall because of Manhattan — it falls for a broader and more technical reason. Look at the sequence of the last five months:
| Month | VP/unit | Difference in the month |
|---|---|---|
| Jan/2026 | R$ 8,75 | — |
| Feb/2026 | R$ 8,70 | - R$ 0,05 |
| Mar/2026 | R$ 8,62 | - R$ 0,08 |
| Apr/2026 | R$ 8,53 | - R$ 0,09 |
| May/2026 | R$ 8,37 | - R$ 0,16 |
It's five months of fall, adding up -R$ 0,38 per unit (-4,3%), and the pace is Accelerating — the loss of May was the largest in the series. The main pressure comes from the market marking of IPCA indexed CRIs. . Much of the VGHF’s credit portfolio yields IPCA plus a spread; when the long interest curve opens (the NTN-B rises by rate), the present value of these securities falls, even if no debtor has given a call. In May, the negative MTM was -R$ 25,1 Mi, versus -R$ 19,5 Mi in April — a worsening of R$ 5,6 Mi in a single month.
Here's the distinction that separates the investor who understands from what only reads the headline: Marking the market is not money coming out of the cashier. . The fund didn't pay anyone R$ 25,1 Mi. If he holds the CRIs to maturity and the debtors pay, he gets the face value and the "negative" MTM falls apart. It's a loss. accounting, no carried out. . So the negative accounting result (sea -R$ 5,2 Mi, Apr -R$ 8,3 Mi, May -R$ 14,1 Mi) does not mean that the bottom is burning box in the same proportion.
But there is one thing the unit cannot ignore: the VP that falls Yeah. Real in the sense that it matters to who buys or sells the unit. When you get VGHF the "29% discount on VP", you're buying something whose VP shrinked 4,3% in five months. If the interest curve does not close, the discount that seemed to be a security margin becomes a mirage that adjusts down to each report. The asset is no longer cheap — the benchmark is that it is falling together.
The dividend dilemma: where does the R$ 0,07 come from?
This is the most delicate part. VGHF pays R$ 0,07 for unit eight months in a row (November 2025 to June 2026). But the arithmetic of May does not close alone: the accounting result was -R$ 14,1 Mi and the accumulated reserve to distribute is in just R$ 167 thousand — approximately R$ 0,001 per unit, i.e. practically zeroed.
How, then, did the fund keep R$ 0,07? The answer is that the negative accounting result is dominated by MTM (which is not a box). What the fund distributes is the cash result — the CRI coupons actually received, the income of the IFIs in portfolio, the interest on transactions. This cash flow still exists and still covers the R$ 0,07 of the month. The MTM overthrows the accounting profit and the VP, but does not prevent immediate distribution.
The risk is structural and cumulative. The reserve to be distributed is the "stock" that allows to soften weak months — it distributes a little more than what was won, withdrawing from the reserve. With the reserve emptied to R$ 167 thousand, this shock absorber is over. From here, the DPS depends almost entirely on the generated cashier in the month itself. . If the June cash flow doesn't cover R$ 0,07 — and it tends to squeeze, because the Manhattan coupon has stopped entering with early maturing — the fund manager will only have two exits:
- Cut DPS to the level that the cashier effectively sustains. It's the honest way out, but it penalizes the unit in the short term.
- Distribute beyond the generated, drawing from the main. This keeps the R$ 0,07 on the screen, but corrupts the VP even faster — the fund would, in practice, be returning capital of the transvestite earner.
The level of R$ 0,07 is already the result of a long shrinkage: the VGHF paid R$ 0,13/unit in 2021 and R$ 0,10 in 2024. The dividend’s trajectory tells the same story as the PV — a fund that has been compressing distribution year after year. With the reserve zeroed and Manhattan out of flux, the maintenance of the R$ 0,07 was no longer comfortable.
Value selling assets to rebalance: active management or problem management?
The report shows the fund manager going through the wallet: sold the CRI VFDL and the CRI Oscar Freire 50S, and reduced positions in BTHF11 and VGIR11. . At first sight, it's active management — turning your wallet, making gains, adjusting exposure. But the context matters: selling CRIs and FIIs units at a time of open interest curve and negative accounting result is also the classic description of who is lifting box.
The two readings are not exclusive, and the probable truth lies in the middle. Valora is actively administering the liquidity of a fund whose principal credit has just stopped paying and whose reserve has dried up. This is what a competent fund manager should do — there is nothing wrong with selling assets to keep the house in order. The warning signal is another: when the sale of assets becomes the recurring source of cash to support the DPS, the fund has entered a regime that is not sustainable indefinitely. It is worth following in the next reports if the spin continues and if there is sale of positions with loss only to generate liquidity — that would be the sign that portfolio management has become grip management.
The Conflict of Interest That Worsens
There is a governance point that deserves attention and is in the wrong direction. About PL 14,6% the VGHF is allocated to funds of Valera itself — and this exhibition It's gone up. (was ~12% in February), even with the reduction of VGIR11. This sets rate in two points: the VGHF unit pays the VGHF management fee and, indirectly, the rate of the funds Value invested by VGHF. The house charges twice for the same structure.
Worse: the VGHF holds a unit subordinate the Pre-Subordinated CRI Value — the tranche that it receives last in the payout cascade and that, in 2026, received no income. . That is, part of the exposure to the house's own funds is in an instrument that, this year, did not generate return to the VGHF unit, while the Valora continues to capture rate on both ends. Increasing exposure to the house itself in a period of cash tightening is the kind of decision that an attentive quotaist should question.
The leverage that nobody sees on the screen
A technical detail closes the frame of fragility. The net cash from the bottom is negative in -R$ 31,4 Mi: VGHF has R$ 31,4 Mi in reverse committed operations — sale of CRIs with a repo commitment —, and this liability exceeds the gross cash. In practice, it is a cost leverage of CDI plus ~0,84% per year. It is not alarming on its own in a R$ 1,38 Bi fund, but, added to a zero reserve and a running credit, it is yet another piece that reduces the fund manager's slack.
Verdict: KEEPING, with reservations — note 5,9/10
VGHF11 is not a case of a broken background. The portfolio is large, pulverized, and the worst credit (Manhattan 161S) has 167% coverage — there is real margin for full recovery. The discount of P/VP 0,71 is genuine in relation to accounting assets. But the thesis changed color: what was a predictable income fund turned a fund in defensive transition, with VP falling five months ago, distribution reserve exhausted, a relevant credit out of the flow and exposure to the fund manager itself rising at the worst time. The R$ 0,07 DPS is vulnerable — not cut but dependent on the cashier of the month covering what the reservation no longer covers.
For those who make sense: investor who already carries the position, understands the mechanics of the MTM (know that the VP falls by marking, not by cap) and is willing to wait for the interest cycle to turn and the execution of Manhattan to settle — betting that the discount on the VP becomes return when the curve closes. The new porter must do so with the awareness that he is buying paper volatility, not broke income.
For those who don't make sense: who seeks to divide predictable and safe in the coming months — the R$ 0,07 can be reviewed. And who confuses "P/VP 0,71" with guaranteed security margin: if the VP continues to fall, the discount is mobile and can evaporate report to report. The number that counts is not today's discount, it's the VP trend — and it points down.
The message of the reanalysis is direct: the VGHF is not collapsing, but it is more fragile than the headline of the "FII with discount 29%" suggests. The next two variables to be monitored are objective — if the June DPS remains in R$ 0,07 and where the cashier who pays it comes from, and if the VP stabilizes when the interest curve stops opening. As long as these two questions have no answer, the "discount" is a promise, not a certainty.
