Chapter 5

Sum-of-the-Parts

At Rp775 (24 July 2026, market capitalisation about Rp2.3tn), PBSA looks cheap on the headline: 7.3x reported earnings, 2.5x book, a 7.7% dividend yield. But a third of that reported profit is an unrealized securities mark. Once the cash-and-securities stack is valued separately, the market is paying roughly 10x normalized construction earnings — a full multiple for a single-customer contractor — and the securities cushion that flattered the headline P/E has already thinned by about 40% in one quarter.

What the price is

Source: price per IDX daily close 24 Jul 2026; 3,000,000,000 shares and equity Rp929.23bn per FY2025 Annual Report [1]; Rp60 dividend per FY2025 results [2].

PBSA has 3,000,000,000 shares listed, of which 15,113,700 now sit in treasury after the early-2026 buyback, leaving roughly 2.985bn outstanding [3]. At Rp775 that is a market capitalisation near Rp2.33tn. Reported FY2025 net profit was Rp320.08bn and EPS Rp106.61, so the trailing multiple is 7.3x — the kind of number that flags a value screen [4]. The Rp60 per-share dividend declared for FY2025 lifts the yield to 7.7% [5]. On book value, equity of Rp929.23bn puts the stock at 2.5x [6].

The reported P/E is the wrong denominator. FY2025 net profit carried a Rp111.53bn unrealized fair-value gain on the securities portfolio and a Rp5.68bn realized gain; stripping both leaves construction-only net profit of about Rp202.9bn (Earnings Quality) [7]. On that base the whole company trades at 11.5x, not 7.3x.

Source: reported EPS Rp106.61 and net profit Rp320.08bn, FY2025 Financial Highlights [8]; normalized construction net Rp202.9bn derived from the FY2025 income statement [9].

Two stacks, not one number

The cleaner way to value PBSA is to separate what it owns from what it does. The company runs with effectively no interest-bearing debt, so its financial assets — cash plus the fair-value securities portfolio — are a stack that can be handed to shareholders or lost to a market move, independent of the construction business. At 31 December 2025 that stack was Rp473bn: Rp145.22bn of cash and Rp327.85bn of short-term investments [10]. Subtract that from the Rp2,325bn market capitalisation and the market is implicitly paying about Rp1,852bn for the construction operation itself — 9.1x its normalized Rp202.9bn of earnings, or 7.6x its Rp243.02bn of operating income [11].

That was the year-end picture. The most recent balance sheet is worse for the cushion. By 31 March 2026 the cash line had fallen to Rp64.43bn and the securities portfolio to Rp207.92bn — a combined Rp272bn, down Rp201bn in a single quarter [12]. On the current stack the same arithmetic values the construction business at about Rp2,053bn, or 10.1x normalized earnings and 8.4x operating income.

Source: market cap at Rp775 x 3.0bn shares; financial stack = cash plus short-term investments, FY2025 AR [13] and Q1 FY2026 statements [14]; construction stub = market cap less the stack.

The green block is the part of the price a buyer can, in principle, underwrite with a balance sheet. It shrank from a fifth of the market capitalisation to about an eighth in three months, as the quarter's Rp71.62bn portfolio loss and negative operating cash flow drained both cash and marks (Capital and Control). The low headline P/E rested partly on that cushion; the cushion is now materially smaller.

Source: cash and short-term investments, FY2025 AR [15] and Q1 FY2026 statements [16].

Is ten times fair for the construction stub

A single-customer contractor with a roughly 12-month backlog does not obviously deserve a full-market multiple, and 10x normalized earnings is not cheap for that risk profile. The case for it rests on two numbers. First, the construction business earns an exceptional return on the capital actually tied up in it: with the Rp473bn financial stack removed, operating equity at end-2025 was about Rp456bn, and Rp202.9bn of normalized profit on that base is a return above 40% [17]. The business is advance-funded and asset-light; it needs little capital to grow. Second, it is growing quickly: Q1 FY2026 operating income rose 59.8% year on year to Rp81.34bn on revenue up 46.8% [18]. Measured against a forward run-rate rather than trailing FY2025, the same price implies a lower multiple.

The evidence points to a construction stub priced for quality and growth rather than for its concentration risk. The strongest fact against that read is that the growth and the returns both depend on one conglomerate's palm-oil downstream capex (Palm Oil Capex): every customer above 10% of revenue and the entire disclosed order book is Sinar Mas / Golden Agri, and the backlog turns over annually with no multi-year visibility. A single lost capex cycle would compress both the earnings base and the multiple at once. What would change the read is order-book replenishment outside that group, or a durable second end-market — neither of which the filings yet show.

The price's own history is the clearest anchor

With no sell-side coverage and no consensus target, the most honest external reference is how the market itself has valued PBSA. Two years ago the stock traded around Rp370, a market capitalisation near Rp1.1tn — 5.2x the FY2024 earnings and 1.4x book. In the H2-2025 spike it reached a Rp1,500 close (Rp4.5tn), 14.1x earnings and 4.8x book, before collapsing [19]. Today's 7.3x reported and 2.5x book sits between the two, still roughly 40% above the pre-spike multiple.

Source: closing prices and market capitalisation from Share Performance Highlights, FY2025 AR [20]; earnings and equity per Financial Highlights [21].

The multiple has re-rated even after a 64% fall from the peak. That the price still trades on momentum as much as fundamentals is visible in the tape: it moved from Rp765 to Rp825 and back to Rp775 in the three sessions to 24 July 2026 on no disclosed news. A buyer here is not just paying for the construction business; they are paying a post-spike multiple set by a thin 16% float.

What the range depends on

Because the financial stack is close to a known quantity, the valuation is most sensitive to the multiple placed on the construction stub. Each additional turn of that multiple on Rp202.9bn of normalized earnings is worth about Rp68 per share. The table below holds the current Rp272bn financial stack fixed and varies the construction multiple.

Source: derived — construction value = Rp202.9bn normalized earnings x multiple; implied price = (construction value + Rp272bn financial stack) / 2.985bn shares [22].

At the current Rp775, the market is placing the construction business at almost exactly 10x. An 8x multiple — still a fair figure for a concentrated contractor — implies about Rp635, roughly 18% below the price; a growth-premium 12x implies about Rp907. The second sensitivity is the securities portfolio itself: at Rp207.92bn it is Rp70 per share, and Q1 FY2026 showed that number can swing by a third of its value in one quarter [23]. A reader who marks the portfolio conservatively and the construction stub at a concentration discount lands well below Rp775; one who capitalises the recent construction growth lands above it. The gap between those two views, not the headline 7.3x, is the valuation question.