Judo’s August buying cluster versus the big-bank trade

Judo Capital Holdings JUDO CAPITAL HOLDINGS LIMITED is not being bought in a vacuum. Australia’s business lending market has held up better than the broader mood, with business credit demand rising 10.6% year to June 2026, even as policy tightening, geopolitics and higher energy prices kept the macro tone cautious. That matters because Judo is a pure-play SME lender, not a universal bank with a dozen other ways to hide a weak patch. When the cycle turns, you see it in the loan book faster.

The comparison that matters is the majors. ANZ, NAB, Commonwealth Bank and Westpac still dominate business lending, and they still command the market’s default trust. Judo is the smaller, more specialised operator, built around relationship lending, faster decisions and competitive pricing funded through deposits and term funding. It has less than 1% of the Australian business loan market overall, and roughly 2% of the SME segment by analyst estimates. That is a long way from the incumbents. It is also why the stock can rerate quickly when the market believes the model is working.

InsiderTrades data puts director-level buys at sweet-spot names in a bucket with a 90-day win rate of 52.7% and an average return of 3.67%. That is the historical backdrop, not a forecast for Judo, and it is the right way to read it here. The point is not that the next 90 days must rhyme with the last 5,829 cases. The point is that this is the kind of size and role where insider buying has historically been least ignored.

Hornery buys after the June damage

The filing that anchors this story is David Stephenson Hornery, Judo’s chairman and co-founder, who bought 50,000 shares at A$1.00 on 20 August, a A$50,000 purchase in local terms and about EUR 30,440 on a euro-normalised basis. He then filed again on 28 August. That is not a token gesture from a passive director. It is an operating founder adding stock after the company has already been through a hard reset.

The timing matters because Judo’s share price had already been hit by the June trading update, which flagged elevated impairments from a small number of idiosyncratic exposures and helped trigger a roughly 40% drop. The August purchases came after FY26 results were out on 18 August, not before them. That sequence gives the buys a different flavour. Hornery was not buying into a vacuum of missing information. He was buying after the market had seen the damage, the recovery in the numbers, and the still-open questions around credit quality.

Jennifer Douglas also bought in the same window, with 29,973 shares at A$0.99 on 19 August, and the broader cluster included CEO Chris Bayliss. The company had 9 distinct insiders trading the name in the same direction over the past quarter, with 12 recent declarations in the dossier window. That is the kind of pattern that deserves attention because it is broader than a lone director nibble. It is also still a cluster inside a single name, not a sector-wide stampede.

Judo’s FY26 numbers versus the majors’ scale

Judo’s FY26 result was not a rescue story. It was a business that kept growing through a difficult patch. Gross loans and advances reached A$14.7 billion, up 18%. Deposits rose to A$12.2 billion, up 24%. Net interest margin improved to 3.13%, up 20 basis points. Cost-to-income improved to 45.3%. Pretax profit rose 34% to A$168.1 million, and net profit after tax rose 29% to A$111.1 million. Those are real numbers, not a marketing deck.

Against the majors, though, the scale gap is still the whole game. The big banks can absorb a bad pocket in business lending and keep moving. Judo cannot. That is why the market punished the June update so hard, and why the August rebound in the results matters. The company showed operating leverage, funding growth and a margin that still looks respectable for a lender in this part of the market. But a specialist lender with a concentrated book does not get the same benefit of the doubt as a diversified bank. You pay for that in volatility.

The valuation gap reflects that. Recent data puts Judo around 0.59x price-to-book and roughly 10x earnings, which is compressed relative to the sector after the June sell-off. The majors trade on richer book multiples because they own the deposit franchise, the distribution, and the market’s trust in their underwriting. Judo does not have that luxury. What it does have is a smaller base, a narrower focus, and a path to grow faster if credit stays manageable.

The insider cluster and what it says about the register

The cluster is not the story by itself, but it is the sharpest piece of evidence in the filing set. InsiderTrades data scores this as a director-level buy in a small or mid-cap name, with a euro-normalised filing value near EUR 30,440 and a market value fraction under 0.01%. That combination matters because it sits in the part of the market where insider activity has historically been less fully priced in. The score rationale also notes that the filing came from an operating director and sat inside a wide cluster. That is the sort of setup our scoring rewards most.

The register detail is more useful than the headline count. Hornery is not a detached board member. He is a co-founder and chairman. Bayliss is the CEO. Douglas is a director. When those names buy in the same window, you are not looking at a random one-off. You are looking at a board and management team that has just lived through a sharp drawdown, a public credit scare and a results release that showed the business still growing. That does not tell you the next quarter will be clean. It does tell you where the internal temperature sits.

The comparison with the majors is useful here too. Big-bank insiders rarely move the needle with small on-market buys because the institutions are so large and the ownership structures so diffuse. At Judo, the same size of purchase carries more weight because the company is smaller and the board is closer to the operating story. A A$50,000 buy is not life-changing money for a chairman. It is still a visible vote when the stock has already been cut down and the company has just put up a decent year.

What the cohort history adds, and what it does not

The historical cohort read is modest, and that is exactly why it is useful. For director-level buys at sweet-spot names, InsiderTrades data shows a 90-day win rate of 52.7% and an average return of 3.67%. That is not a heroic number. It is not supposed to be. It says that this bucket has been slightly better than a coin flip over the next 90 days, with enough positive drift to matter when the setup is otherwise credible.

The important part is the bucket, not the mythology. Judo sits in the size range where insider buying has historically been more informative than at the mega-caps, but less dramatic than at tiny illiquid names where a small trade can distort the picture. That is why the cohort stat belongs inside the comparison with the majors. The majors are too big for a small buy to matter much. Judo is small enough for the buy to matter, but large enough that the business still has institutional relevance.

InsiderTrades strategy data also points to a live out-of-sample framework on the restricted EU venue universe, with tokens for 0.81, 26.4 and 51.5. Those figures survive only in that restricted, short-window regime and do not survive search-aware deflation, so they are a screen, not an alpha claim. I am not leaning on them here because the company-specific story is already doing the work. The point is simply that the framework exists, and that the filing sits in a bucket where the historical evidence has not been trivial.

Judo’s execution against the credit scare

The real comparison with the majors is not just valuation. It is execution under stress. Judo’s June update exposed how quickly a specialist lender can be marked down when a few exposures go wrong. The market did what it always does with lenders after a credit surprise. It assumed the next surprise might be worse. That is why the stock fell roughly 40% and why the August results had to do more than just look tidy.

They did enough to matter. Loan growth was still there. Deposits were still growing faster than loans. Margin improved. Costs improved. Profit grew. The company also guided FY27 pretax profit to A$210 million to A$220 million, with above-system loan growth, broadly stable net interest margin, continued cost-to-income improvement and return on equity expected near 8%. That is not a blowout target. It is a credible path back toward steadier compounding if credit stays contained.

The majors do not need that kind of narrative repair. They already have the market’s trust. Judo has to earn it back. That is why the insider cluster matters more here than it would at a bank with a fortress deposit base and a diversified earnings mix. The board is buying after the company has shown it can still grow through the noise. The market will still want to see whether the June impairments were an isolated hit or the first sign of a more stubborn credit cycle.

What to watch while the majors keep their premium

The next comparison point is not another headline buy. It is whether Judo keeps delivering the operating numbers that justify a rerating from a compressed base. Watch the loan book, the deposit mix, and the margin. Watch whether the cost-to-income ratio keeps moving the right way. Watch credit quality, because that is where the June shock came from and where the market will keep looking for trouble. The majors can absorb a wobble. Judo has to prove the wobble was contained.

The valuation gap to the big banks is still the obvious argument for the bulls. Judo trades on a much lower book multiple than the majors, and the market has already priced in the June scare. If the company keeps growing deposits and loans, keeps the margin stable, and avoids another credit surprise, the stock has room to narrow that gap. If impairments reappear, the discount will look deserved. There is not much middle ground in a lender like this.

That is why the August buying cluster deserves to be read as part of the setup, not as the setup itself. Hornery, Douglas and Bayliss bought after the damage, after the results, and after the market had already made its judgment. The majors still look safer. Judo looks cheaper. The question is whether the cheaper name can keep executing well enough to make the discount look excessive rather than earned, and the next hard test is the company’s credit performance into the next reporting window.

This is not investment advice.