Transition Capacity of an Emerging-Market National Oil Company: A Framework-Based Assessment of Indonesia

2.1 Two dimensions of financial resilience

The framework applied here treats financial resilience as having two distinct dimensions rather than one. The level dimension asks how much capacity to service the debt stock a firm sustains on average, measured as operating cash generation relative to total debt, in the tradition of cash-flow-based early-warning measures developed for emerging markets [8]. That tradition has since been extended to non-bank financial intermediaries in the same market setting [9] and to the transmission of distress into equity returns through the pandemic period [10]. The event dimension asks how likely the firm is to cross a threshold of financial pressure, measured as a binary distress state. The distinction is not cosmetic. In the reference panel described in Section 3.4, state ownership leaves the level dimension statistically unchanged while being associated with approximately 72% lower odds of crossing the distress threshold, as reported in Table 1. The wider repositioning of this literature from financial distress toward financial resilience has been reviewed systematically elsewhere [11]. A variable can therefore be irrelevant to the average and decisive at the extreme. The explanation for this pattern is a difference in exposure rather than a difference in management. A sovereign owner that sets domestic prices, adjusts fiscal terms, and absorbs part of the price cycle attenuates the transmission of commodity price movements before those movements reach the income statement [12-19]. Attenuated transmission compresses the lower tail of the outcome distribution without shifting its centre, which is precisely the pattern observed in the panel. The attenuation is partial rather than complete, however, and Section 4.1 documents an earnings contraction of roughly one quarter in the case firm's own accounts within the observation window.

2.2 Structural versus adaptive resilience

These findings support a distinction between two forms of durability. Structural resilience is conferred by institutional position: a sovereign owner able to provide implicit guarantees, preferential financing, and fiscal terms that absorb part of the price cycle before it reaches company earnings [12-19]. It works by attenuating shocks. Adaptive resilience is accumulated through managerial choice, through conservative leverage, liquidity buffers, portfolio flexibility, and investment sequencing [20, 21], and works by absorbing shocks after they arrive. In dynamic-capabilities terms [22, 23], adaptive resilience is a reconfiguration capability, while structural resilience is an endowment that lowers the need for reconfiguration.

For a wholly state-owned firm, the governance implication is uncomfortable. The same insulation that protects the firm financially attenuates the market discipline that would otherwise force transition, a tension already visible in the mixed evidence on the operating efficiency of state-owned producers [24]. Structural resilience buys time; whether that time is used for transformation is a matter of ownership design rather than of market pressure. This is the central reason why the recommendations offered here are addressed to the shareholder rather than to management.

2.3 Measurement in very large firms

Two measurement results condition the assessment that follows. First, conventional distress criteria do not transfer to firms of this scale. In the reference panel of Section 3.4, among the 799 firm-years for which the full ratio set is available, an interest-coverage criterion identifies only 14 distress events and a non-positive-earnings criterion 13, while a net-leverage criterion identifies 90. The three together identify 103 events, or 12.9% of firm-years, and every event flagged by the coverage criterion is already flagged by one of the other two, so the coverage criterion contributes nothing of its own. The median coverage ratio on an earnings before interest, taxes, depreciation and amortisation (EBITDA) basis is 19.0 times, so a coverage-based trigger is effectively silent through the industry's worst episodes. Net debt above four times EBITDA is therefore the appropriate trigger for this population.

That threshold is not chosen arbitrarily. It is written explicitly into supervisory guidance on leveraged transactions in the euro area [25] and into the interagency guidance issued by the United States banking agencies [26]; it is used in the same range by the Financial Stability Board in its assessment of leveraged lending vulnerabilities [27]; and it corresponds to the point at which published rating methodology moves a financial risk profile from significant to aggressive. The academic basis is equally established. Covenants written on debt to EBITDA are among the most common terms in private credit contracts, and their breach has measurable real effects on corporate investment [28, 29]. The ratio of debt to operating cash flow has been used as a measure of debt overhang with a demonstrated causal effect on investment [30]. Threshold-based classification of distressed firms is standard in the literature on zombie firms [31, 32], which conventionally applies an interest-coverage trigger; the substitution of a leverage-based trigger here is required by the emptiness of the coverage criterion in this population rather than chosen for convenience. Second, the discriminatory power of conventional financial ratios weakens in a population of financially homogeneous giants, so a firm's position relative to peer percentiles is more informative than its absolute ratio values.

2.4 Absorption capacity and the non-linear boundary

The absorption of an additional transition cash burden is not linear in the size of that burden. The mechanism is a threshold effect. While free cash flow remains positive, a burden is met from operations and leaves the balance sheet unchanged. Once free cash flow turns negative, cash reserves deplete and the shortfall becomes new borrowing, which raises net leverage toward the distress threshold. The point at which that transition occurs is therefore the quantity of interest for a shareholder setting a mandate.

The construct applied here is defined at three levels, and the distinction between them is substantive rather than presentational. Gross headroom is EBITDA less financing costs less baseline capital expenditure. Post-tax headroom deducts income tax expense, which is not discretionary. Post-distribution headroom further deducts the dividends actually remitted to the shareholder, which are discretionary. The absorption boundary is the annual transition burden, expressed as a proportion of total assets, at which post-tax headroom reaches zero; that boundary is reported throughout as the planning quantity. Below it a transition programme is self-funding from operations; above it the programme necessarily draws on reserves or on new debt. The three-level definition matters because the gap between the post-tax and post-distribution lines is not an accounting detail but the shareholder's own decision variable, and Section 5 returns to it. The construct is deliberately mechanical rather than behavioural, and it is conservative in the sense that firms in practice respond by re-phasing capital expenditure and adjusting distributions.

4.1 Financial position

Table 2 summarises the financial position and the derived framework metrics for the six years from 2020 to 2025. In 2025, the firm generated EBITDA of US\$11.4 billion on revenue of US\$70.9 billion, against total assets of US\$91.6 billion and interest-bearing debt of US\$23.0 billion. Cash and equivalents stood at US\$14.6 billion, so net debt was US\$8.4 billion. Capital expenditure was US\$6.0 billion on the investing cash flow measure, and financing costs were US\$1.2 billion at a reported cost of debt of 4.56%. Income tax expense was US\$2.7 billion, and dividends of US\$2.5 billion were remitted to the shareholder.

The six-year series answers a question that a single year cannot. On every framework metric, the position in 2025 lies within the range observed since 2020 and close to the centre of that range: a resilience ratio of 0.50 against a range of 0.39 to 0.60, net leverage of 0.73 times against a range of 0.33 to 1.26, and a liquidity buffer of 15.9% against a range of 14.0% to 21.7%. The assessment year is therefore neither a cyclical peak nor a trough, and the metrics reported below are not an artefact of cyclical position.

Two features of the series deserve separate emphasis because they carry directly into the absorption analysis. Earnings are volatile: EBITDA fell by 24.8% between 2023 and 2024, and the 2020 trough was 47.0% below the 2023 peak. And the effective tax rate ranges from 35.3% to 63.6% with a mean above 40%, which reflects the fiscal terms applying to upstream operations in this jurisdiction and which is the reason the distinction between pre-tax and post-tax headroom is material rather than technical.

4.2 Position relative to global peers

Table 3 places the 2025 metrics against the reference distribution described in Section 3.4. The picture is more differentiated than aggregate commentary usually allows, and the two dimensions of the framework point in different directions.

Three readings follow. On the event dimension, the firm is comfortably positioned: net leverage of 0.73 times is roughly half the panel median and less than a fifth of the distress threshold, and the liquidity buffer is an outlier at roughly five times the panel median. On the level dimension, the picture is unremarkable: the resilience ratio of 0.50 falls below both the full-sample median and the median for state-owned firms, indicating that the firm generates less annual operating cash flow per unit of debt than the typical large producer. Leverage and asset turnover sit almost exactly at panel norms.

Table 2. Financial position and derived framework metrics, 2020-2025

Notes: Compiled from the company's Annual Reports for 2020 to 2025. Amounts in US$ million. Derived metrics computed by the authors from the reported figures. EBITDA is as published by the company. Capital expenditure is the sum of investing cash outflows for oil and gas and geothermal properties, fixed assets, and exploration and evaluation assets. ICR is operating income divided by finance costs; the EBITDA-based coverage ratio is reported separately because only the latter is comparable with the reference panel. EBITDA = earnings before interest, taxes, depreciation and amortisation, ICR = interest coverage ratio, EBIT = earnings before interest and taxes.

Table 3. Framework metrics relative to the global reference distribution

Table 4. Free cash flow headroom ladder, 2020-2025

This combination is what the framework predicts for a structurally resilient firm. Ordinary operating performance coexists with a strong position against extreme outcomes, because the protection operates on exposure rather than on operations. The very large liquidity buffer merits separate emphasis. It is the single most distinctive feature of the firm's balance sheet, and the panel estimates in Table 1 identify liquidity as the strongest firm-level protector in the tail of the distribution even though it does not discriminate in the middle. The firm therefore holds, in unusual measure, precisely the asset that matters most where structural protection would be tested.

4.3 Transition commitment profile

The firm's stated transition position is a Net Zero Emission aspiration for Scopes 1 and 2 by 2060 or earlier, operationalised through a roadmap of eight acceleration initiatives spanning geothermal generation, solar, biofuels and low-carbon fuel processing, battery and electric-vehicle ecosystem development, and carbon capture, utilisation and storage. Installed renewable and new-energy capacity reached approximately 3,226 megawatts in 2025, up from approximately 2,502 megawatts a year earlier, an increase of roughly 29%.

That aspiration is not adopted in isolation. It sits inside Indonesia's national net zero emission commitment, the country's position on which has been assessed directly [33], and the factors that determine the attainable national emission level have been identified empirically for the same economy [34]. A corporate commitment of the kind examined here is therefore best read as the firm-level expression of a national target rather than as an independent strategic choice, which is a further reason why the shareholder's mandate, and not management's ambition, is the operative variable in what follows.

The direction of travel is consistent with the wider national research agenda on sustainable energy deployment, where feasibility assessments of solar, wind, and hybrid systems for Indonesian infrastructure have moved from concept to costed engineering-economic analysis [35], and with the broader evidence on the determinants of environmental, social and governance performance in Indonesian and regional firms [36, 37]. Two features of the firm's own position are notable. The commitment is material rather than token by the standards of the global industry, where clean-energy spending has remained a small share of the sector's capital expenditure [38] and cumulative firm-level ratios have been reported in the low single digits [39]. At the same time, the annual report does not disclose transition capital expenditure as a separate, audited line item within capital goods investment, so the share of capital actually allocated to low-carbon assets cannot be verified from published accounts. This is the disclosure gap in concrete form: the commitment is visible, its financing is not.

4.4 Absorption capacity

Applying the definition in Section 2.4 to the 2025 accounts, gross headroom is US\$11.4 billion of EBITDA less US\$1.2 billion of financing costs less US\$6.0 billion of capital expenditure, or US\$4.2 billion, equivalent to 4.61% of total assets. Income tax expense of US\$2.7 billion reduces this to US\$1.5 billion, so the absorption boundary on the post-tax definition adopted here is 1.63% of assets per year. Dividends of US\$2.5 billion remitted to the shareholder reduce headroom further, to negative US\$1.0 billion, or negative 1.13% of assets. Table 4 reports the full ladder for all six years.

Three observations follow from Table 4. The post-tax boundary has ranged from 0.92% to 5.58% of assets per year since 2020, and its 2025 value of 1.63% lies in the lower half of that range. The post-distribution line was positive in every year from 2020 to 2024 and negative in 2025, which means that in the assessment year the firm did not generate sufficient cash from operations to meet capital expenditure, financing costs, income tax and the dividend simultaneously; the difference was met from the cash balance and from new borrowing. And the gap between the two lines is not stable. Dividends remitted have ranged from 4.9% to 68.7% of profit for the year over the six years, a variation of more than thirteen-fold, so the quantity available for transition investment is determined more by the annual distribution decision than by operating performance.

Two sensitivities are reported rather than left for the reader to discover. On the accrual basis used in the company's own management discussion, which records capital goods investment of US\$6,612 million for 2025, the corresponding boundaries are 3.93% gross and 0.95% post-tax; the choice of capital expenditure measure therefore places the post-tax boundary in a range of roughly 0.95% to 1.63% of assets per year. Separately, finance income of US\$1.0 billion in 2025 nearly offsets the US\$1.2 billion of finance costs deducted here, so measuring the deduction net of finance income would raise each level of the ladder by approximately US\$1.0 billion and would bring post-distribution headroom close to balance. The gross measure is retained because it is consistent with the denominator of the coverage ratios and because it is the conservative choice. Under every variant, the post-tax boundary remains far below the gross figure of 3.93% that a pre-tax calculation would report.

4.5 The fragility of absorption capacity

The boundary is contingent on earnings, and the contingency is severe. Table 5 reports post-tax headroom across a grid of earnings shocks and transition burdens, with income tax scaled at the 2025 effective rate. The grid is calibrated on the firm's own history rather than on an assumed elasticity. The 25% column corresponds to the contraction of 24.8% actually recorded between 2023 and 2024, and the 40% column remains inside the peak-to-trough movement of 47.0% observed between 2023 and 2020.

The result is stark. A 10% earnings reduction lowers the boundary from 1.63% to 0.91% of assets. A 25% reduction, a shock the firm experienced in the year before the assessment year, removes the boundary altogether: post-tax headroom becomes negative, at negative 0.16% of assets, so any transition burden at all must be funded from reserves or from new debt. At a 40% reduction, post-tax headroom is negative US$1.1 billion before any transition spending whatsoever. Figure 1 shows the base and shocked schedules together.

The disproportion between the shock and the response is the central quantitative finding of this assessment. Capital expenditure, financing costs and, under this fiscal regime, a substantial part of the tax charge are inflexible in the short run, while earnings are not, so absorption capacity is a highly geared quantity. The cash reserve of US$14.6 billion is what separates a strained programme from a curtailed one: at a burden of 2% of assets under a 25% earnings reduction, the reserve would sustain the shortfall for approximately seven years.

Two implications follow directly. Transition mandates should be expressed as multi-year commitments with explicit contingency provisions rather than as fixed annual percentages, since the affordable percentage is itself a function of the price cycle and, as Section 4.4 shows, of the distribution decision. And the liquidity buffer identified in Section 4.2 should be understood as transition capacity held in reserve rather than as idle cash available for distribution.

Table 5. Post-tax headroom across earnings shocks and transition burdens

tu_pian_1.png

Figure 1. Post-tax free cash flow headroom against the annual green investment burden, at base 2025 earnings and under a 25% earnings reduction

Four limitations qualify this assessment. The reference panel covers listed firms, so its application to a wholly state-owned company is an assumption rather than a result, although the direction of any bias favours the conclusions reached here, since the excluded firms are plausibly those with the strongest sovereign linkage. EBITDA is reported by the case firm on a basis defined by its own performance framework rather than by a standard audited definition, and no restatement was attempted. The absorption calculation is mechanical: it holds capital expenditure, financing costs and managerial behaviour constant, whereas firms in practice re-phase investment and adjust distributions, so the boundary should be read as a planning quantity rather than as a forecast. And the distress classification depends on a single threshold; the four-times convention is supported by the supervisory guidance and the covenant literature cited in Section 2.3, but the sensitivity of the panel estimates to alternative thresholds is not reported here.

Three extensions follow. The most valuable is to assemble a disclosure-based panel of unlisted national oil companies, which is feasible for the substantial number that publish audited statements, and to estimate the two-dimensional framework directly on that population, which would allow the insulation mechanism to be tested where it is presumed strongest rather than inferred from listed proxies. A second extension is to decompose the insulation mechanism into its fiscal, contractual and hedging components, which would identify which features of state ownership perform the protective function and therefore which are replicable by contract rather than requiring ownership. A third is to apply the absorption framework comparatively across emerging-market national oil companies, which would establish whether the boundary computed here is typical or particular, and whether the dominance of the distribution decision documented in Section 5.2 generalises. Each extension gains urgency from the wider transition-risk agenda, in which the pace of demand decline remains genuinely uncertain [45], the strategic options open to incumbents are contested [46], and the resulting exposures propagate to lenders, investors and sovereign balance sheets [47, 48].

This paper forms part of the first author's doctoral research at IPB University. The panel estimates reported in Section 3.4 were computed for this paper from published consolidated financial statements and are reproducible from the sources described there. The authors thank the reviewer for comments that materially improved the measurement basis of the assessment.