The Hong Kong stock exchange appears to have concluded that the most valuable part of an ETF is not the listing fee. It is the index.
HKEX has now co-branded benchmarks with Korea Exchange and with Bursa Malaysia, licensed them to the Hong Kong arms of five mainland asset managers, and sits back while the products arrive.
This is the first index HKEX and Korea Exchange have ever put their names on together, and the construction is unusually candid about why it exists. Thirty companies, fifteen from each market, with the Korean side lifted wholesale from the KRX Semiconductor Top 15 Index and the Hong Kong side drawn from names eligible for Southbound Stock Connect.
Then the weights. Hong Kong carries 62 per cent, Korea 38 per cent. Not because anyone believes that is the correct balance of the semiconductor world, where two of the largest players are Korean, but because the index is built to qualify for Southbound ETF Connect. The methodology says so in plain language. There is even a tripwire: if the Korean names drift above 38 per cent, an ad hoc rebalance drags them back down.
The better detail is what happens when Hong Kong runs short of semiconductor companies. The methodology anticipates this, and permits topping the list up with what it calls Supplementary Securities, drawn from electronic components, communications equipment and computer hardware and storage. In practice the Hong Kong half currently includes a laptop manufacturer, a phone manufacturer, an acoustic components business, and a camera lens maker, most of them sitting obediently near the individual weight ceiling.
So a semiconductor index in which the largest single market is the one with the fewest semiconductor companies, padded out with consumer hardware to reach its quota. The exchange brands it, the Korean exchange co-brands it, a German firm calculates it, and an asset manager licenses the result and collects the fee. Everyone has a job except, arguably, the portfolio manager.
Here is where it gets funny. Entirely different partner market, entirely different remit, no sector focus at all, sixty constituents rather than thirty, thirty from each exchange in a display of perfect headcount symmetry.
And the weights are 62 per cent Hong Kong, 38 per cent Malaysia. The same split, to the percentage point, as the semiconductor index. Same 12 per cent individual security cap. Same ad hoc rebalance if the partner market creeps above 38 per cent.
Two indices, two unrelated partner markets, two unrelated investment cases, one identical number. That is not a view on relative opportunity. It is a compliance constant wearing an index methodology as a disguise. Both benchmarks are also backdated to the same base date, several years before either of them existed, and both start at a base value of 10,000. The house style is at least consistent.
Worth saying that most of the coverage reported the Korea split as 60/40. The methodology on the exchange's own site says 62/38. If you are building products off either benchmark, the second number is the one that matters.
The same logic, one corporate step tighter. Hang Seng Investment Management has listed a fund tracking the Hang Seng Artificial Intelligence Theme Index, which selects Hong Kong-listed companies on the basis of computing power, algorithms, model development and cloud infrastructure.
The index provider and the fund manager are, of course, in the same family. There is nothing improper about it and the thematic definition is more disciplined than most. It is simply another instance of the same trade: whoever owns the benchmark owns the economics, and everything downstream is assembly.
New York has arrived at the same conclusion from the opposite direction. The move here is not to build an index but to license someone else's and slice the market as thinly as the plumbing allows. This one runs on an independent provider's US banking benchmark, and it did not arrive alone: a world consumer staples version (PBCS, NYSE) and a world ex-US industrials version (PBWN, NYSE) listed alongside it.
The name is the giveaway. These are not portfolios. They are components, sold to somebody who intends to do the actual portfolio construction themselves. An entire product range built on the premise that the asset allocation decision belongs to the buyer. Which is either admirable humility or the industry quietly conceding that the value has moved elsewhere.
Amid all of this, National Bank Investments listed the next rung on its target-maturity bond ladder, a range that already covered every year from 2026 to 2031. Investment-grade credit, held to a stated maturity, income paid along the way, capital returned at the end.