The Concentration You Cannot See: Why Manager Count Is Not Diversification

An investor holding five private credit funds will usually describe their position as diversified. Five managers, five credit committees, five origination networks, five sets of underwriting standards. On the face of it, no single lender's judgement can determine the outcome.

It is a reasonable description of the structure. It is often a poor description of the risk.

Diversification across managers and diversification across borrowers are different things, and only one of them determines what happens when a large borrower fails. Manager diversification is the one investors can see and the one that gets marketed. Borrower diversification is the one that binds, and in most private credit portfolios it is neither measured nor disclosed.

The administration of the Bathla Group has made that distinction concrete.

Why the same borrowers keep appearing

Adding managers to a portfolio feels like adding independence. In Australian real estate credit it frequently does not, for reasons that are structural rather than accidental.

There is a finite population of developers operating at scale in any metropolitan market. The largest of them are, by most measures a credit committee applies, the most attractive counterparties available: established track record, repeat delivery, meaningful pipelines, the capacity to absorb facilities large enough to be worth writing. A manager assessing that borrower on its merits will frequently conclude it is a good credit. So will the next manager, independently and on the same evidence. Origination compounds it, because managers in this market draw from overlapping broker and introducer networks and the same transactions are shopped to multiple lenders.

The result is a portfolio-level concentration that no individual manager created, none of them can observe, and none of them has breached a limit to produce. It is also close to invisible from outside. Funds typically disclose sector, geography, loan type, weighted average LVR and term. Very few publish borrower-level detail, and almost none publish it in a form that can be aggregated with another manager's.

What look-through exposure actually does

The error most investors make here is specific, and it runs in one direction.

The assumption is that spreading capital across five managers cuts single-name exposure to roughly a fifth of what any one manager holds. That assumption only holds if the managers hold different borrowers.

Look-through exposure is the allocation-weighted average of each fund's exposure. Where three managers each hold the same borrower at 8 per cent of their respective books, and an investor holds those three funds equally, look-through exposure to that borrower is 8 per cent. It has not fallen towards 2.7 per cent. It has not moved at all. Every one of those managers sits comfortably inside a 10 per cent single-borrower limit, and every one of them can accurately report compliance.

Two points follow, and both are commonly stated the wrong way round.

Look-through exposure can never exceed the largest single-fund exposure. It is bounded above by it. The failure is not that concentration multiplies at the portfolio level. The failure is that the expected dilution never arrives, while the investor continues to believe it has, having paid several fee layers to obtain it.

Adding managers reduces look-through exposure to a given borrower only if the new managers hold less of that borrower than the existing weighted average. Where every manager lends into the same top tier of names, adding managers leaves the aggregate close to unchanged. Diversification by count is not diversification by exposure.

There is a separate effect that does compound, and it sits with the borrower rather than the investor. As the same name is written by more lenders, that borrower's total leverage across the market rises, which raises the probability of failure for everyone holding it. That is a different question from what the failure costs.

The size of the position is the whole argument

Assume a distressed resolution recovering 60 cents in the dollar, a loss severity of 40 per cent. The figure is illustrative rather than an estimate of any outcome, and it excludes enforcement costs, unrecovered accrued interest and the time cost of a multi-year workout, all of which push realised severity higher.

On that assumption, and assuming the full exposure resolves at that severity:

The upper end of that table is not hypothetical. Published guidelines for at least one Australian real estate credit fund rated by a major research house permit a single borrower to represent up to 20 per cent of net asset value, and a single investment up to 15 per cent. Limits at that level are not concealed. They are disclosed, and most investors read past them.

Identical loan. Identical sponsor. Identical underwriting. Concentration changes nothing about whether a loan was well underwritten. It changes everything about what a bad loan means for the person who owns it.

Bathla as the case in point

Administrators from Teneo were appointed on 25 August 2026 to Universal Property Group Pty Limited, Raj and Jai Construction Pty Limited and associated companies within the Bathla Group, according to documents lodged with ASIC. Universal Property Group reported approximately $3.2 billion in liabilities as at 30 June 2025, the majority of it reportedly owed to private credit funds (ABC News, 25 August 2026). Reported lenders include Centuria Bass, Credit Connect, PAG Asia Capital, CVS Lane Capital Partners, Balmain, Ray White Capital, Keyview and La Trobe Financial, with Alceon having exited an approximately $670 million exposure in January (The Urban Developer, 27 August 2026).

Bathla was not an obscure borrower funded by marginal lenders. It was a large-scale developer that a number of experienced credit teams independently assessed and independently approved. That is evidence that the recurrence described above is a live feature of this market. It is not, on its own, evidence about any particular investor's aggregate position, which public reporting cannot answer.

Each of those lenders could see its own position precisely. What none could see with the same precision was the borrower's total leverage across the whole lending market. What their investors could not see was how much of that one name they held once exposures across several funds were added together. The first is a credit problem, affecting how likely the loss is. The second is a portfolio construction problem, affecting what it costs, and it is the one almost nobody is solving.

It is worth being precise about one thing the administration does not change. The sponsor is a large part of any credit assessment and highly relevant to execution, but the sponsor is not the collateral. A developer entering administration changes the counterparty and who manages the projects. It does not, of itself, change the value of the security.

Most multi-manager credit portfolios have something they describe as a diversification policy. Fewer have one that would have constrained the outcome above. Five properties, and a limit needs all of them.

• It has to bind at the portfolio level. A limit applied inside a single fund constrains that fund and says nothing about what an investor holds in aggregate. The only limit that protects the investor is applied across every underlying manager simultaneously, as a single constraint.
• It has to aggregate to the ultimate borrower group. Large developers operate through many entities, and project-level special purpose vehicles are the norm in development finance. The Bathla administration itself covers a main corporate entity, a construction entity and a large number of associated companies. If the aggregation does not reach through to the group, it is measuring something other than the risk.
• It has to treat builder concentration separately. One builder can be engaged across several unrelated borrowers and several unrelated managers, which makes a builder insolvency a correlated event hiding behind counterparty names that appear independent. Borrower-level aggregation alone will not detect it. This is the control least commonly applied in this market and the one this cycle is most likely to test.
• It has to be measured on the shortest cycle the underlying reporting permits, and that cycle has to be disclosed. A multi-manager portfolio can only aggregate what its managers report, and the standard reporting cycle in this market is quarterly. A manager claiming continuous look-through aggregation is either receiving something the rest of the market is not, or describing an intention. The defensible position is a disclosed cycle, direct interrogation between cycles when events warrant, and an as-at date attached to every figure published.
• The level has to be set where a total loss on any single name is survivable. That is the table above, and it is the property from which the other four take their purpose.

None of this substitutes for credit work. A concentration framework applied over poor underwriting simply distributes bad loans more evenly. It is worth being equally precise about where an allocator's contribution sits: it is sizing discipline. It is not underwriting outcomes, which belong to the managers who write the loans. Conflating the two produces a claim that cannot survive a bad loan, and every portfolio in this asset class will have bad loans.

Four questions worth asking any private credit manager

1. What is the largest single borrower group exposure, measured on a full look-through basis, and as at what date?
2. Is builder concentration aggregated separately from borrower concentration, and if so, to what limit?
3. What proportion of the book is currently in extension or workout, and how is "default" defined for that purpose? If the manager is an allocator, is that definition standardised across underlying managers or does it vary?
4. On what cycle is look-through aggregation actually performed, and what is the reporting lag?

A manager reporting no extensions or workouts at all in current conditions is making a claim that warrants testing. Extensions and workouts are an ordinary feature of construction lending, particularly in this interest rate environment, and transparency on them is a governance obligation rather than a reputational risk.

How we operate

Applied to our own structure, so that readers can hold us to the same standard.

The Ekam Real Estate Credit Fund caps look-through exposure to any single borrower group at 5 per cent of total portfolio exposure, aggregated across every underlying manager rather than assessed manager by manager. A separate 5 per cent cap applies to any single builder, aggregated the same way.

Look-through aggregation is performed on a quarterly cycle using reporting received from each underlying manager, and is interrogated directly with managers between cycles where events warrant. We do not describe this as continuous monitoring, because manager reporting in this market is not continuous, and a control should be described as it operates rather than as it might ideally operate.

Fifteen specialist managers were active in the Fund as at 30 June 2026. That number is an outcome rather than a target. We have reviewed a considerably wider universe.

Our look-through exposure to Bathla Group entities was approximately 2 per cent of total portfolio exposure based on manager reporting as at 30 June 2026, held across two of the fifteen managers, and has since been confirmed directly with both managers. It sits well within the 5 per cent cap. It was recorded before the administration was announced rather than assembled afterwards.

One of those managers has restricted investor redemptions. That is a separate question from concentration, and a larger one than the 2 per cent implies. Part two addresses it directly.