Follow the Money #1: EPF trimmed six major holdings. CIMB was the exception

Follow the Money, Issue #1. Changes in EPF's disclosed substantial shareholdings on Bursa, read straight from the filings. A weekly series.

Directors and substantial shareholders must disclose changes in their interests. So must EPF, the fund guarding the retirement savings of millions of members. We aggregated EPF's Section 138 notices from 1 to 17 July and removed the changes distorted by IPOs or new share issuance, which are not clean trades. Across the seven clean, self-consistent names left in the period, six were net reductions. CIMB was the only addition.

THE PATTERN: six net reductions, one addition.

EPF recorded net reductions across banks, REITs and Tenaga:

• Tenaga: about 25.9m shares sold (holding 24.86% to 24.41%), about minus RM371m
• RHB Bank: about 20.7m sold (34.54% to 34.04%), about minus RM175m
• Public Bank: about 28.1m sold (16.71% to 16.57%), about minus RM147m
• Alliance Bank: about 5.8m sold (8.86% to 8.52%), about minus RM28m
• Sunway REIT: about 10.4m sold (16.02% to 15.72%), about minus RM23m
• Pavilion REIT: about 7.8m sold (12.97% to 12.78%), about minus RM13m

CIMB was the only clean addition during the period:

• CIMB: about 43.6m shares bought (17.70% to 18.11%), about plus RM339m

So of the big banks, EPF trimmed Public Bank, RHB and Alliance, but added CIMB. At current prices, the net share reductions across the six names are equivalent to roughly RM757m, compared with roughly RM339m added to CIMB. These are approximate value equivalents at current prices, not the actual amounts EPF paid or received.

A BIG NUMBER IS NOT ALWAYS A REAL MOVE: You may see it noted that EPF's stake in Johor Plantations jumped to about 12%, or that it added around 44m shares of SD Guthrie. Read those carefully. Johor Plantations only listed in late June, so that stake is cornerstone allocation and early trading, not a fresh conviction bet, and EPF has in fact been selling it since. SD Guthrie's share count rose while EPF's percentage barely moved, which means the company issued new shares, not that EPF was buying. We excluded both, because a big disclosed number is not the same as a real trade. Separating the two is the whole point of this series.

HOW TO READ THIS: These are small changes in percentage terms, because EPF is enormous, and they are small relative to its total holdings. But the direction was consistent across six separate names, and the ringgit amounts are real. What it means is a judgment call. The disclosures show activity, not intent, and we are showing you what the fund did, not why. Valuation, the rate cycle, mandate rebalancing and liquidity could all play a part.

METHODOLOGY: For each stock we summed every disclosed acquisition and disposal in EPF's Section 138 substantial-shareholder filings from 1 to 17 July 2026, giving the net shares traded in the window. Percentages are the holdings before and after, as stated in the filings. Ringgit values are approximate value equivalents at current prices, not trade proceeds. Names with a corporate action (IPO, placement, new issuance) in the window are excluded, because the change is not a clean trade. These are substantial-shareholding disclosures under the Companies Act 2016, which apply at or above 5 percent and to changes while a substantial shareholder. They do not show every EPF position or every EPF trade. Source filings are on Bursa under each company's announcements.