Judo Capital 40% Crash|3 Bad Loans vs Best-Ever Margin

Record Fall on a Record Margin

Judo Capital Holdings fell 40 per cent on Thursday to its lowest price since 2023, wiping $400 million from its market value in a single session. At the same moment, the bank was reporting its strongest-ever net interest margin — above 3.2 per cent — and loan growth that pushed gross advances beyond $14.4 billion. The contradiction sits inside one number: three specific loans, across three unrelated sectors, deteriorated in the final weeks of the financial year. That concentration, not the business model, is where the explanation lies — but whether it stays there is the unresolved question the market priced violently today.

The announcement came as an ASX update on Thursday morning. CEO Chris Bayliss arranged an emergency analyst call to explain a $20 million increase in specific provisions, one week before financial year-end and just six weeks after the bank reaffirmed its profit guidance at the lower end of its $180–190 million range. During that call, the stock fell 40 per cent.

Profit before tax guidance was revised to $163–169 million for FY26, an 8 per cent miss against the $180.7 million consensus. The cost of risk — the annual loan-loss charge as a proportion of the book — rose to $116–122 million, from $40.1 million in the first half alone. That doubling of loss costs in six months is the figure that drove the session, not the revised profit number.

What makes the fall harder to place is the other side of the ledger. Net interest margin lifted above 3.2 per cent versus prior guidance of 3.15 per cent. The cost-to-income ratio is tracking below the first half's 48.5 per cent. Deposit costs remain below historical averages. The operating business is not distressed — it is running at its best reported metrics. The problem is that three borrowers are not.

What Three Loans in Three Sectors Actually Tell You

Judo's containment argument rests on a specific claim: the three deteriorating exposures are in different sectors, different states, and reflect borrower-specific developments that emerged after a full customer-by-customer review completed earlier in the year. If that framing holds, the loss is a known quantity and the repricing is complete. If it does not, the review's failure to surface these loans six weeks ago raises a harder question about what else the portfolio contains.

The timing is the fault line. The April trading update — which 12 of 13 covering analysts used to reaffirm Buy ratings — included a portfolio review specifically designed to flag elevated-risk sectors: agriculture, construction, and manufacturing, all sensitive to fuel costs and inflation. That review found no material deterioration beyond existing provisions. The three loans that subsequently failed are described as having deteriorated due to "recent, borrower-specific developments" that post-date the review. The market is being asked to accept that three unrelated businesses, in three sectors, became uncollectable in the six weeks after a clean audit.

SmartCompany's coverage read this as evidence of a "broader economic slowdown affecting small business borrowers." The Market Index analysis took the opposite position, arguing the 10.5 times trailing price-to-earnings ratio — the cheapest the stock has ever traded — still prices in roughly 30 per cent earnings growth given the unchanged FY27 guidance floor. Two articles, same event, opposing conclusions.

The buried assumption in the containment thesis is that the review methodology can detect stress before it becomes a provision charge. The three loans that broke suggest the opposite: credit deterioration in the SME space may be moving faster than quarterly portfolio reviews can identify. At 3 per cent of gross loans now 90-days-past-due or impaired — up from 2.65 per cent just three months prior — the movement in the asset quality ratio is not marginal. The question investors cannot resolve today is whether 3 per cent is the ceiling or a floor.

What the articles do not carry is position-level data on which institutional class was selling Thursday's volume. What is observable is the direction: 12 of 13 analysts held Buy ratings entering the session; the stock reached 83 cents intraday before recovering to 91 cents at the close. That spread between analyst positioning and execution-level selling is the conflict the pool records, and it is not resolved by the close.

FY27 Guidance Gap and What the Big Banks Priced

The FY27 guidance cut carries a different weight than the FY26 revision. A year-end provision charge is a one-time adjustment; a forward earnings target cut signals that the cost-of-risk environment is expected to remain elevated into the next financial year. Judo's FY27 profit before tax guidance of $210–220 million sits 16 per cent below the $255 million consensus — a gap that cannot be explained by three specific exposures already provisioned.

That gap implies management expects the macro environment to continue generating credit losses above historical norms through FY27. Given that Judo lends exclusively to small and medium enterprises — a cohort that carries no mortgage offset, limited asset backing, and direct exposure to interest-rate pass-through — the FY27 cut is a forward reading on the SME credit cycle, not a backward adjustment for known losses.

The session's other data point connects. National Australia Bank, the largest of the big four by business lending exposure, fell 3.4 per cent on Thursday. Westpac, ANZ, and Commonwealth Bank fell between 1 and 2 per cent. Bell Potter's Richard Coppleson linked the banks' sell-off to the morning's employment data — unemployment fell to 4.4 per cent, well above forecasts, cementing a 48 per cent market-implied probability of an RBA rate hike by year-end. A rate increase into an SME credit cycle already showing 3 per cent impairment at Judo would further compress borrower margins and lift the probability of additional provisions across the sector.

The Australian's headline — "Judo's warning for the big banks — fears over hidden risks lurking" — is not commentary. It is the channel through which the day's credit deterioration signal transmits: if a specialist SME lender with a relationship-led model and a recent portfolio review missed three emerging bad loans, the question about what a diversified major bank's SME book contains is structurally identical and has a larger absolute exposure.

30 June Is the Confirmation Date

The verification event is four trading days away. Judo's financial year closes on 30 June, and the full balance sheet at that date will confirm whether the 90-days-past-due and impaired ratio holds at 3 per cent — the number management guided today — or moves higher.

If the ratio closes at or below 3 per cent with no additional provisions beyond the three named exposures, management's containment argument gains the evidence it currently lacks. The FY27 guidance of $210–220 million would then be interpretable as a conservative reset from a known event, and the 10.5 times trailing earnings multiple would represent a genuine value entry for a bank still growing its book at a pace its competitors cannot replicate in the SME segment.

If the ratio closes above 3 per cent, or if additional exposures emerge in any sector beyond the three named, the idiosyncratic framing collapses. The portfolio review's failure to identify Thursday's loans six weeks prior means a second failure to identify further deterioration in the same window would be a methodological finding, not a coincidence — and FY27 guidance would need to be revised again.

The counter-evidence in the pool is real: NIM at 3.2 per cent is a genuine operating improvement, deposit costs are below historical averages, and loan volume growth to $14.6 billion by 30 June remains on track. The business is generating revenue efficiently. The risk is that three-sector simultaneous deterioration suggests the credit environment is exerting stress across Judo's borrower base faster than the model's relationship-led monitoring can surface it.

For a holder, the monitoring variable is the 30 June closing 90DPD ratio — not the FY26 profit number, which is already bracketed at $163–169 million. For a watch-list candidate considering entry at 10.5 times earnings, the trigger is whether that ratio holds or breaks. The stock's cheapest-ever valuation is only a floor if three is the final count.