DSEX closed at 5,336 on May 24 — up 125 points across a four-day pre-Eid rally. Fifteen of Bangladesh’s 36 listed banks are now in Z-category. Both statements are true at the same time, and that is not a contradiction the market should be able to produce. The reason it does comes down to a single line in the index methodology: Z-category stocks are excluded from DSEX by design. Remove the worst performers from a measurement and the measurement looks healthier.
The question is whether the index you watch every day actually reflects the market you invest in. Right now it does not — and to understand why, you need to understand how the three DSE indices are actually built.
How DSEX Is Built
The DSE Broad Index was launched on 28 January 2013 by S&P Dow Jones Indices with a base value of 1,000. It uses free-float market capitalisation weighting, which means only shares available for public trading are counted. Sponsor and director holdings, government stakes (excluding ICB), and locked-in shares are stripped out before each constituent’s weight in the index is calculated.
The formula is straightforward in structure. Take every eligible stock’s price, multiply by its free-float share count, sum those products across all constituents, divide by an index divisor, and multiply by the base value of 1,000. The divisor is the mechanism that keeps the index continuous when constituents change. Without it, every addition or removal would produce a discontinuous jump in the index value that had nothing to do with actual price movement on the exchange.
Eligibility requires a minimum float-adjusted market capitalisation of Tk 10 crore and trading on at least 10% of sessions over the prior six months. The index covers roughly 97% of total equity market capitalisation on the DSE. Z-category stocks, mutual funds, debentures, and bonds are excluded by methodology.
That exclusion clause is where the May 2026 distortion enters.
DS30: The Blue Chip Filter
Launched alongside DSEX, the DS30 selects 30 stocks under considerably stricter rules. Constituents need a minimum float-adjusted market cap of Tk 50 crore, a three-month average daily trading value of Tk 50 lakh, and positive net income across the most recent four quarters combined.
Sector concentration is capped to prevent any single industry from dominating the blue chip index. No more than five stocks from any of six sectors — banks, financial institutions, insurance, real estate, pharmaceuticals, and fuel & power — can appear in DS30. The combined limit across those six sectors is 20.
DS30 rebalances semi-annually in January and July. The DSE Index Committee reviews every listed stock against the criteria. The July 2024 rebalancing dropped 10 firms — a routine event by the methodology, but a structural one for the stocks involved.
DSES: The Shariah Subset
The DSEX Shariah Index, launched on 20 January 2014, is a subset of DSEX. Every DSES constituent must already be in the parent DSEX index and additionally pass a rules-based Shariah screen.
The screen has two parts. The business activity filter excludes conventional banks, conventional insurance, alcohol, tobacco, gambling, and pork-related businesses. The financial ratio filter requires debt-to-total-assets below 33%, interest income below 5% of total revenue, and interest-bearing cash holdings below 33% of total assets.
That second filter is unintentionally significant in May 2026. The same debt and interest screens that disqualify highly-leveraged conventional banks also incidentally screen out the kind of balance sheets most vulnerable to non-performing loan crises. DSES does not contain the conventional banks now sliding into Z-category — because it never could under its own rules.
Why Rebalancing Moves Prices
Index changes are not bookkeeping events. They are forced flows.
When a stock enters DSEX or DS30, every passive fund, ETF, and mandate-bound institution tracking the index must buy it to match weightings. When a stock is removed, the same buyers become forced sellers. Research from S&P Dow Jones on global indices shows added stocks typically see 2–5% abnormal returns around rebalancing dates, with partial reversal in subsequent weeks.
In Bangladesh, the effect is amplified because mandate restrictions on Z-category holdings are stricter than in most markets. Margin loans are prohibited, settlement extends to T+3, and many institutional investors are barred from holding Z-stocks entirely. When 13 banks were downgraded between 30 April and 4 May 2026, the forced selling was not theoretical. It happened across institutional books in a matter of days.
What the Index No Longer Sees
Which brings us back to the paradox we opened with. DSEX rose into Eid in part because the banking sector’s worst performers were no longer in it. The 15 banks now in Z-category — including Islami Bank, AB Bank, UCB, Rupali, and Premier — represent a meaningful share of total banking market capitalisation. They are still trading. Investors still own them. Their crisis is real and ongoing.
The index simply does not see them anymore. That is the cost of survivorship bias in a methodology that excludes the failing. When the market you watch starts diverging from the market you hold, the index is doing exactly what it was designed to do — and exactly what you should not let it tell you alone.