Chapter 9

Return Without Re-rating

Figures converted from IDR at historical FX rates — see data/company.json.fx_rates. Ratios, margins, and multiples are unitless and unchanged.

From the end of 2021, Saratoga's net asset value per share compounded at 2.8% a year while a shareholder lost 11.4% a year. From the end of 2020 the same two figures are 14.8% and 20.4%. In both cases the difference is the discount, not the portfolio. This chapter prices what a holder earns if the discount never moves again, and finds the distribution decision worth up to 5.8 points a year of it.

What the last five years paid

Two things can be measured across the same period: what the portfolio did per share, and what an owner of the shares actually collected. They are not the same number, and neither is the other's proxy.

Saratoga publishes net asset value per share back to 2002, restated for the 2021 five-for-one split. The series runs $0.101 (2016), $0.115, $0.081, $0.118, $0.164, $0.291, $0.292, $0.234, $0.246 and $0.267 at the end of 2025 [1][2]. The 2024 figure was published as $0.247 in the 2024 report and appears as $0.246 in the 2025 one; the arithmetic below uses the figure each year's own report carried.

The share price over the same window: $0.048 at the end of 2020, then $0.196, $0.164, $0.107, $0.130 and $0.095 [3][4][5]. Dividends paid in each of those years, split-adjusted, were $0.0015, $0.0039, $0.0049, $0.0014 and $0.0009 [6][7][8][9].

Sources: derived from net asset value per share [1][2], year-end closing prices [3][4][5] and dividends paid [6][9]; each series reinvests the dividend at its own value.

The shape matters more than the endpoints. The green line quadruples in a single year — 2021 — and then falls for four of the next five periods. The blue line rises in four years out of five and never doubles. An owner who bought at the end of 2020 beat the portfolio by fifty-four points of cumulative return; an owner who bought a year later trailed it by fifty. Same assets, same management, same twelve months of separation.

Where the return came from

A holder's return in any year comes from three places: what the portfolio did per share, the dividend measured against the price paid, and the change in the ratio of price to net asset value. The first two are what an investor thinks they are buying. The third has been the largest of the three in four of the five years since 2020; in 2023 it was a close second to a 19.8% fall in net asset value per share.

Source: derived from the net asset value per share series [1][2], year-end closes [3][4][5] and dividends paid [6][7][8][9]; the three components compound rather than add.

In 2025 the portfolio added 11.7% per share and the shareholder lost 23.7%, because the price-to-net-asset ratio fell by nearly a third. In 2024 the portfolio added 10.5% and the shareholder made 28.8%, because the ratio rose by 15%. The income column never exceeded 3.2 points in any of the five years: on the pre-2026 payout, the dividend was not a material part of the answer.

The 2026 part-year runs the other way. $0.0062 a share was paid on 12 June [10] and the shares closed at $0.1056 on 27 July [11], a total return of 17.9% from the $0.0948 close, of which 6.5 points is the dividend — more income in seven months than in the previous five years combined.

Over a longer frame the two measures separate further. Net asset value per share was $0.092 at the end of 2013, the year of listing, against $0.267 at the end of 2025 — 10.7% a year for twelve years [12]. The shares were sold in the June 2013 offering at $0.385, or $0.077 after the 2021 split [13]; at $0.1056 that is 3.7% a year over thirteen years before dividends. The corpus carries the dividends paid from 2020 onward — $0.019 a share across seven payment years — but not those for 2014 to 2019, so a complete since-listing total return cannot be built here. All of these figures are in US dollars converted from rupiah; the currency's depreciation over the period is a further deduction for a dollar-based holder, and sits in Inside the Portfolio.

Return when the gap does not move

Every chapter before this one has examined why the gap exists. None has priced what happens if it simply stays. That is the more useful question for a buyer at $0.1056, because it is the case that requires nothing to go right.

The arithmetic is exact and short. If the ratio of price to net asset value is unchanged from one year to the next, a holder's return is the portfolio's own return on net assets plus the dividend per share multiplied by the difference between one divided by the price and one divided by net asset value per share. At $0.1056 against the $0.2666 published at 31 December 2025 [14], that difference works out at 3.43 percentage points for every $0.0060 of annual dividend per share.

So the discount, held constant, is not a drag on the forward return at all. It is a multiplier on the distribution. A dollar of net asset value paid out arrives in the holder's hand as a dollar; the same dollar left inside is capitalised by the market at $0.40. Every dollar distributed moves about $0.60 of value across that boundary, and nothing about the portfolio has to change for it to happen.

Source: derived from the $0.1056 close of 27 July 2026 [11], published net asset value per share of $0.2666 [14], the 2025 and 2026 dividends [9][10] and the 2025 consolidated statement of cash flows [15]; annual holder return at an unchanged price-to-net-asset ratio.

The rows are anchored, not invented. Zero is the case where the marks go nowhere. 1.7% is what net asset value per share actually compounded at over the four years from the end of 2021. 10.7% is the twelve-year rate since listing. 6.0% sits between them. The columns are the FY2024 dividend of $0.0009, the FY2025 dividend of $0.0062, and $0.0101 — the whole of the holding company's recurring cash income for 2025, which the audited cash flow statement puts at $137.6 million once dividends received of $150.6 million and interest of $9.7 million are reduced by interest paid, employee costs, tax and other operating payments [15].

Two readings sit in that grid. On the four-year portfolio rate and the current dividend, a holder earns about 5.2% a year with the discount frozen — below the 5.45% to 7.88% Saratoga itself pays on its rupiah bank borrowings [16]. On the twelve-year rate and full distribution of recurring cash, the same holder earns 16.5%. The distance between those two cells is not the discount. It is the combination of what the portfolio does and what the board decides to send out.

What the distribution decision is worth

Distributing the entire $137.6 million of 2025 recurring cash income rather than none of it moves roughly $83.2 million of market value to shareholders at the current ratio — $0.0061 a share, 5.8% of the price, recurring annually. It is also entirely discretionary: the stated dividend policy specifies no ratio, only that distributions require sufficient profits or positive retained earnings and must respect liquidity, capital adequacy and future investment requirements [9], and the company sets formal annual targets only for operating expenses [18].

The same arithmetic settles a question left open by Realisation Record. At a fixed price-to-net-asset ratio, a buyback and a dividend deliver almost exactly the same value. Spending $0.0062 a share retiring stock at $0.1056 lifts net asset value per share from $0.2666 to $0.2766, worth $0.1096 at the same 39.6% of net assets; paying it as a dividend leaves $0.1031 of share plus $0.0062 of cash, or $0.1093. Three ten-thousandths of a dollar apart. Saratoga has bought back nothing in either 2024 or 2025, distributing 5,305,000 and 8,031,900 treasury shares to employees instead [17], but the $84.1 million it did distribute in June 2026 captured what a buyback of the same size would have captured. The instrument is close to neutral; the amount is not.

The strongest fact against reading all of this as a case for maximum distribution is that the company has a use for the money and says so. Management targets US$100 million to US$150 million of new investment a year [10], and deployed $165.9 million in 2025 against $357.9 million in 2024 [15]. Capital deployed at cost and marked at cost converts into $0.40 of share price on day one, so it has to compound to roughly two and a half times its outlay before it matches a rupiah simply handed over — but that is a demand for a good return, not an argument that no return is available. The judgement here is that the payout is worth more to a holder than the deployment record supports, and it would change if a deployment produced a realisation above its carrying value; the record of realisations sits in Realisation Record.

The second counter-fact is that the discount has never actually been constant. It ran 70.6% at the end of 2020, 32.6% a year later and 64.4% at the end of 2025 [3][5][1]. A frozen ratio is a modelling convenience that has never held for two consecutive years. It is used here because it isolates what the holder earns without needing the market to change its mind — and because, as the decomposition table shows, movement in that ratio has swamped everything else in both directions.

What would move the answer

Each of these is checkable in a specific document, against a specific prior figure.

Dividends received in the 2026 cash flow statement. $150.6 million in 2025, down from $263.3 million [15]. This line sets the ceiling on any repeat of the $84.1 million distribution without recourse to disposals or borrowing.

The anchor the 2027 meeting uses. The May 2026 resolution set the payout at 19.13% of a reported profit of $439.2 million [10], a profit that was mostly fair-value gain. A percentage of dividends received and a percentage of reported profit give very different answers in a year when the marks fall, and 2023 was such a year.

Purchases of investments in the cash flow statement. $165.9 million in 2025, $357.9 million in 2024 [15]. A return to the 2024 rate alongside a maintained dividend means new borrowing, as it did in 2024 when $219.0 million was drawn.

Any treasury purchase beyond incentive-plan replenishment. Fifteen million shares remain in treasury against roughly five million distributed a year [17]. A purchase larger than that requirement would be the first time the discount is used rather than described.

A named entry in the divestment row of the Investment Milestones chart. The row records exits in 2023 and 2024 and names none for 2019 through 2022 or for 2025 [1]. A realisation above carrying value would raise the portfolio-return row of the grid and narrow the ratio at the same time.

The register. PT Unitras Pertama held 4,289,610,000 shares, or 31.62%, at 31 December 2025, [16], against 4,438,610,000 a year earlier [17]. Further sales from the 89% block enlarge the float; they also supply stock into a thin market.