alirezarezvani/claude-skills / finance/skills/stock-analysis
stock-analysis
UnverifiedAnalyze, research or value a listed stock or company (US, global, NSE or BSE) with verified data, ratios, scores and valuation. Use whenever asked if a stock is a buy, what it is worth, or how a company looks.
references/sectors/rail-freight.md
View on GitHub# Railroads and rail freight networks — sector playbook
Use this when: the company owns or operates a freight rail network or earns its margin from moving freight over one — US and Canadian Class I railroads, Mexican and other national freight carriers, short lines and regional holding companies, and in India the listed rail-linked logistics operators (container train operators, private freight terminal and multimodal players) that run over Indian Railways infrastructure.
A railroad is not a transport company with trucks that happen to run on steel. It is a capital-intensive network utility with duopoly geography, a common carrier obligation, real pricing power, and a maintenance burden that never stops and is only partly visible in the depreciation line. The economics are closer to a regulated pipeline or a tower network than to a trucking or dry-bulk shipping business, and analysing it with a trucking cyclicality framework produces confident nonsense in both directions — it under-rates the pricing durability and under-states the capital intensity.
Two facts govern everything below. First, the sector's own efficiency metric — the **operating ratio** — is inverted relative to every margin metric in the generic toolkit: lower is better, and a falling OR is only good news if the network is still being fed. Second, a railroad can manufacture years of margin and free cash flow simply by under-spending on track, ties, ballast, bridges and locomotives, and the damage shows up in service metrics long before it shows up in the P&L. Most of this playbook is about telling genuine efficiency apart from deferred maintenance.
If the company builds rail infrastructure, rolling stock or signalling rather than operating a network (wagon and coach manufacturers, EPC contractors, RVNL/IRCON-type entities), it is a capital goods / infrastructure EPC business — use the relevant playbook, not this one. If it leases rolling stock and funds itself as a financier (IRFC-type entities), route to `references/sectors/nbfc.md`.
## Contents
- [Why the generic ratio set fails here](#why-the-generic-ratio-set-fails-here)
- [The metrics that actually matter](#the-metrics-that-actually-matter)
- [How to value companies in this sector](#how-to-value-companies-in-this-sector)
- [Peer set construction](#peer-set-construction)
- [Sector-specific red flags](#sector-specific-red-flags)
- [Cycle and structural context](#cycle-and-structural-context)
- [India vs global notes](#india-vs-global-notes)
- [Checklist](#checklist)
## Why the generic ratio set fails here
Most generic ratios are computable here, unlike in banking. The problem is that several are *mis-scaled*, one is *inverted by convention*, and the two that matter most — FCF and ROCE — are the easiest in the entire market to flatter. Report the sector's own metric alongside, or instead of, each of these.
**OPM / EBITDA margin — computable, but the wrong convention and the wrong denominator discipline.** The sector speaks in **operating ratio = operating expenses / operating revenue**, so OR ≈ 100% − OPM. Report OR. Two reasons this is not merely cosmetic: (a) every peer, every management guidance statement and every historical series is stated in OR, so converting to OPM breaks comparability with the company's own record; (b) OR moves mechanically with **fuel surcharge revenue**, which inflates both revenue and expense — a rising diesel price worsens OR arithmetically while dollar profit is unchanged or better. Always ask for OR ex-fuel-surcharge, and for the adjusted OR excluding restructuring, land sales and casualty items. EBITDA margin is worse than useless here: depreciation is a real, large, recurring economic cost of consuming a physical network, typically running to a high-single-digit to low-teens percentage of revenue. A railroad that "generates EBITDA" while its rail is wearing out has generated nothing.
**ROCE / ROIC on reported book — structurally overstated, sometimes grossly.** The capital employed denominator is historical-cost track, grading, tunnels, bridges and land, much of it laid down decades or a century ago and, in the case of land and grading, never depreciated and never revalued. Replacement cost of a mainline network is a large multiple of its carrying value. A railroad showing 14% ROCE on book may be earning materially less on the capital it would take to recreate the asset. This matters analytically in one direction: it makes ROCE useless for cross-sector ranking and useful only against the company's own history and direct peers with similar network vintage. It also means a *newly built* corridor will show a punishingly low ROCE for a decade against an old network's high one, with no difference in operating quality.
**Free cash flow and FCF yield — the single most manipulable number in the sector.** Capex is huge (structurally mid-to-high teens as a percentage of revenue, often more) and the great majority of it is *sustaining*, not growth. Because track renewal can be slipped a year or three without an immediate revenue consequence, "record free cash flow" is frequently just a capex holiday. FCF yield is informative **only after** you have separately identified maintenance capex — from the capex programme disclosure, from track miles of rail and ties installed, from the capex/depreciation ratio, and from the fleet age — and re-run FCF at a normalised sustaining level. Treat a headline FCF yield computed on reported capex as unverified.
**D/E, net debt/EBITDA and interest coverage — mis-signed if read against a generic band.** Stable, contracted-ish, inflation-linked cash flows over hard, mortgageable assets support far more leverage than an industrial average. Investment-grade Class I railroads run sustained net debt/EBITDA in the low-to-mid 2x range by design, much of it raised to fund buybacks, and this is a deliberate capital structure rather than distress. The signal to look for is not the level but the *use*: leverage rising while capex/revenue falls is a warning; leverage rising to fund network expansion at returns above WACC is not.
**Current ratio, quick ratio, working capital and inventory days — near-irrelevant.** A railroad carries essentially no saleable inventory (materials and supplies are track components and fuel), bills customers on short terms, and routinely runs a current ratio below 1 with committed revolver backup. Reading that as a liquidity problem is a category error. The real liquidity question is commercial paper access and revolver headroom in a downturn, not the ratio.
**Operating lease and equipment-rent treatment — check before comparing anything.** Locomotives, cars and terminals may be owned, leased or hired from other carriers (car hire and equipment rents flow through operating expense and therefore straight into OR). A carrier that leases heavily reports a worse OR and a lighter balance sheet than an identical one that owns, and lease capitalisation rules differ across accounting regimes and vintages. Normalise for lease and equipment-rent intensity before any cross-carrier OR or ROIC comparison.
**Asset turnover — structurally low, and not comparable to anything else in transport.** Revenue to gross assets in the 0.2–0.4x region is normal. A trucking company at 1.5x looks "more efficient" and is simply a different business: it rents the road from the taxpayer. Never rank a railroad against a trucker, a 3PL or a freight forwarder on turnover, margin or capital intensity.
**P/B — understates by construction.** Book equity reflects historical-cost infrastructure net of decades of depreciation, less large cumulative buybacks. P/B of 4–6x can be perfectly consistent with the stock trading below replacement value of the network. Use replacement value directly instead (see valuation).
**Revenue growth as an unqualified positive — must be decomposed before use.** Reported revenue change = volume × mix × core price × fuel surcharge × FX. A quarter of "8% revenue growth" that is 6 points fuel surcharge is a margin-neutral, value-neutral event. Conversely, revenue *falling* because coal volumes rolled off while core price rose 4% can be a strengthening franchise. Always decompose.
**EV/EBITDA used alone — misses the capital intensity that defines the sector.** Two railroads at the same EV/EBITDA can have completely different capex/revenue and therefore completely different owner earnings. Pair it with EV/EBIT and with EV/(EBITDA − maintenance capex).
**Depreciation itself needs reading, not accepting.** Railroads depreciate long-lived track assets over very long lives, often on group or composite methods, and the capitalise-versus-expense boundary for rail grinding, tie replacement, ballast cleaning and locomotive rebuilds is a genuine accounting choice that moves operating ratio by real amounts. A carrier that capitalises more of its renewal work reports a better OR and a worse capex line than one that expenses it, with identical physical activity. Check the property accounting policy note and the depreciation study history before comparing OR across carriers or across an accounting-policy change.
**ROE — flattered twice over, and not the metric to lead with.** Book equity is depressed by historical-cost accounting on the asset side and by years of buybacks on the liability side, so ROE can be spectacular at a company earning only a modest return on the capital genuinely at work. Some large networks have driven book equity down far enough that ROE ceases to mean anything at all. Use ROIC against WACC and treat ROE as a by-product of the capital structure.
**EPS growth taken at face value — check the share count first.** Sustained buybacks funded partly by leverage are standard capital allocation in this sector. EPS can compound for years while operating income, volume and network quality are flat. Report operating income growth, revenue ton-mile growth and per-share growth separately so the reader can see which is which.
**Dividend yield and payout ratio — read against FCF at normalised capex, not against EPS.** A payout ratio that looks conservative on earnings can be unaffordable once sustaining capex is set at an honest level, particularly for a carrier that has been deferring renewal.
**EV/Sales and PEG — no analytical content here.** Revenue per unit of enterprise value varies with commodity mix and length of haul rather than with quality, and PEG on reported EPS rewards exactly the buyback-and-defer pattern this playbook is designed to detect.
## The metrics that actually matter
Ranges are **indicative only**. They vary by market, network geography, commodity mix, gauge and electrification, regulatory regime, cycle stage and period, and any of them can be wrong for a specific company in a specific year. Comparison against a tight peer set and against the company's own 5–10 year history overrides every absolute band below. Bands quoted for "Class I" refer to large North American mainline freight railroads and do not transfer to short lines, Indian operators or state systems.
| Metric | Definition / how to compute | Indicative healthy range | Why it matters |
|---|---|---|---|
| **Operating ratio (OR), adjusted and ex-fuel** | Operating expenses / operating revenue, in percent. Compute three versions: reported, adjusted (excluding restructuring, land/asset sale gains, casualty and legal charges), and ex-fuel-surcharge (strip surcharge revenue and fuel expense from both sides). | Class I best-in-class 55–60%; 60–65% good; 65–72% average-to-improving; >75% sub-scale or structurally disadvantaged. Short lines 70–82%. State systems carrying a passenger cross-subsidy can sit near or above 95%. | The sector's defining efficiency metric, and the number management is compensated on — which is exactly why it must be interrogated. Lower is better. A 100 bps improvement on a large network is very large in absolute profit, so OR trajectory drives the multiple more than absolute margin does. But OR improves for good reasons (longer trains, better asset utilisation, higher price) and bad ones (deferred maintenance, shedding low-margin but network-feeding volume, one-off land gains in the revenue line). Never accept an OR improvement without attributing it. |
| **Volume: revenue ton-miles (RTM) / net tonne-kilometres (NTK), and carloads / units** | RTM = tons of revenue freight × miles hauled. Carloads (and intermodal units) count shipments. Track both plus **gross ton-miles (GTM)**, which includes locomotive and car weight and is the correct denominator for productivity and fuel metrics. India: NTKM and originating tonnage; container operators report TEUs and rakes. | Judge against industrial production, the company's own history and rail-served market growth — not an absolute band. Volume shrinking while price rises is normal in secular-decline commodities and not automatically bad. | Volume and price are the two revenue drivers and they behave completely differently. Carloads tell you shipment count (drives cost — crews, switching, terminal work); RTM/NTK tells you the work performed (drives revenue and fuel). A network moving fewer carloads at longer length of haul can grow revenue while cost falls. Divergence between carload growth and RTM growth is a mix signal you must explain before scoring anything else. |
| **Commodity mix and coal exposure** | Revenue and volume split by: intermodal, coal, grain/agricultural, chemicals and petroleum, automotive, metals/minerals, aggregates and construction, forest products. Compute coal as % of revenue *and* % of operating profit — coal is usually higher-margin than its revenue share implies. | Coal below ~10–12% of revenue is a structurally lower-risk mix for a US network today; above 20% requires an explicit run-off model. Intermodal above ~40% of volume means truck-competitive pricing dominates the revenue line. | The single largest determinant of both cyclicality and pricing power. Each commodity has a different competitive alternative: coal has almost none (captive utilities), so it carries high rates and high margin but is in secular decline; intermodal competes directly with truckload, so its price is capped by diesel-and-driver economics; grain is weather- and export-driven and lumpy; autos and chemicals track industrial cycles. Two railroads with identical OR and identical volume growth can have entirely different forward economics purely on mix. |
| **Revenue per carload / per RTM, and core pricing** | Revenue per carload = freight revenue / carloads. Decompose the change into mix, length of haul, fuel surcharge and **core price**. Core pricing is the rate change excluding fuel and mix — compare it to a rail-specific cost inflation index (in the US, the STB's Rail Cost Adjustment Factor family; elsewhere, wage plus materials inflation). | Core pricing sustained 100–250 bps above rail cost inflation is the signature of genuine network pricing power. At or below rail inflation over several years means the moat is being competed or regulated away. | This is the moat test, and it is the most important single line of the analysis. A network with duopoly geography and captive shippers can price above its own cost inflation year after year; a network that has to match truck economics cannot. Note the asymmetry: revenue per carload can rise purely on longer haul or a shift to denser commodities, which is mix, not power. Insist on the company's own core-price disclosure and sanity-check it against revenue-per-unit trends by commodity. |
| **Train velocity and terminal dwell** | Average train speed (mph or kph, system-wide, excluding time in terminals) and average terminal dwell (hours a car spends in a yard between movements). In the US these are filed weekly with the STB and published by the AAR; elsewhere, dig them out of investor presentations. | Class I system velocity roughly 20–26 mph in fluid conditions; terminal dwell roughly 20–28 hours. Direction and volatility matter far more than the level, which is geography-dependent. | The two headline fluidity metrics, and the earliest observable evidence of whether an OR improvement is real. A network running faster with shorter dwell is genuinely more productive: the same physical assets do more work. A network whose OR improved while velocity fell and dwell rose has bought margin with service, and will pay it back in lost volume, customer complaints, regulatory attention and eventually a costly recovery programme. Read these two metrics *before* you read the margin. |
| **Car miles per day, cars online, trip plan compliance** | Car miles per day = total car miles / (cars online × days). Cars online = active cars on the system. Trip plan compliance = % of shipments delivered within the published plan, reported for intermodal and manifest separately. | Car miles per day rising over time is the productivity signal; cars online falling *while volume holds* is genuine asset release. Trip plan compliance in the 70–90% band is typical; a sustained fall below the company's own norm is a service failure. | These distinguish the two ways of shrinking cars online. Good: better velocity means each car does more work, so fewer are needed — car miles per day rises. Bad: volume left the network, so cars are idle or stored — car miles per day is flat or falling while cars online falls. Both look identical in a headline "we removed 15,000 cars" slide. Trip plan compliance is the customer's view of the same network and is what shippers testify about to regulators. |
| **Train length, train starts and crew productivity** | Average train length (feet or cars), total train starts, GTMs per employee, and crew starts per thousand GTM. Also locomotive count in service vs stored, and horsepower per trailing ton. | GTMs per employee rising steadily is the core productivity series. Train length increases are real productivity up to the point where they exceed siding lengths and terminal capacity. | Running fewer, longer trains is the primary lever behind modern rail margin improvement, and it is genuine — one crew, fewer locomotives, more freight. But it has a hard physical limit and a service cost: very long trains cannot fit passing sidings, block crossings, take longer to build and break, and increase the chance a single failure strands a large amount of freight. Train length rising while trip plan compliance falls is the trade-off going the wrong way. |
| **Fuel efficiency and fuel surcharge mechanics** | Gallons (or litres) of diesel per thousand GTM. Separately: the fuel surcharge programme — index used (highway diesel or WTI), trigger threshold, and the lag between the price move and the billed surcharge (commonly around two months). | Class I roughly 1.0–1.2 gallons per thousand GTM, improving slowly with new locomotives, distributed power and idle-reduction. Rail is several times more fuel-efficient per ton-mile than truck — that gap is the structural competitive advantage. | Fuel is one of the largest single expense lines, and the surcharge lag creates a mechanical margin swing that has nothing to do with performance: when diesel rises fast, the surcharge under-recovers and margin dips; when it falls fast, the railroad over-recovers and margin looks great. Analysts routinely mistake this for operating improvement or deterioration. Always separate the *efficiency* trend (gallons per GTM, a real long-term productivity and ESG lever) from the *recovery* effect (timing noise). |
| **Labour cost and agreement exposure** | Compensation and benefits as % of revenue and as % of operating expense; headcount vs volume; average headcount cost; and the status of collective agreements — expiry dates, national bargaining rounds, arbitration or government-imposed settlements, crew-consist rules. | Labour typically the largest or second-largest expense, often roughly a quarter to a third of operating expense. Headcount falling faster than volume over a multi-year period is productivity; headcount cut below the level needed to recover from a demand upturn is a future cost. | Rail labour is heavily unionised, agreements are multi-year, and in some jurisdictions strikes are governed by special statutes (the US Railway Labor Act, with Congressional intervention as a live possibility; Canadian federal mediation and back-to-work legislation). The cost is therefore lumpy and partly outside management control: a national settlement can reset the wage base in one step, and back-pay accruals distort a quarter. Crew-consist and one-person-crew rules are a multi-hundred-basis-point OR question decided in bargaining and regulation, not by management. |
| **Capex as % of revenue, and the maintenance vs growth split** | Total capex / revenue; capex / depreciation; and the disclosed split between maintenance-of-way and infrastructure replacement (rail, ties, ballast, bridges, signals) versus growth and productivity projects (new sidings, terminals, capacity, locomotives). Cross-check with physical disclosure: track miles of rail replaced, ties installed, average locomotive and car fleet age. | Structurally high: mid-to-high teens as % of revenue is the normal running rate for a mainline network, and can exceed 20% during a capacity or regulatory-driven programme. Capex / depreciation sustained near or below 1.0x for several years is a deferral flag. | This is where the sector's most important judgement sits. Maintenance capex is very large and largely irreducible — the network physically degrades under load whether or not the market is good. Management has wide discretion over *timing*, which converts into reported FCF. The physical disclosures are the honest check: ties and rail installed per year, and fleet age, cannot be spun the way a capex budget can. A company cutting rail and tie programmes while celebrating record FCF is borrowing from its own future, and the repayment comes as slow orders, derailment risk and a step-up in spend three to five years later. |
| **ROIC vs WACC across a full cycle** | NOPAT / (net debt + equity), or the company's own defined ROIC — check the definition, since treatments of operating leases, deferred tax and goodwill differ. Compare to an estimated WACC and track the spread over at least one full freight cycle. | A mature Class I network earning a sustained few hundred basis points above WACC is a genuinely good business; ROIC in the low-to-mid teens against a high-single-digit WACC is the modern benchmark. Peak-year ROIC alone proves nothing. | The real test of whether the network creates value, and the only metric that simultaneously captures pricing power, cost control and the capital intensity that OR alone ignores. It is also the metric regulators watch: in the US the STB makes an annual **revenue adequacy** determination, and railroads earning persistently above their cost of capital face greater regulatory pressure on rates. That makes ROIC a two-edged number here — high is good for shareholders and increases regulatory risk, which is a genuinely sector-specific dynamic worth stating in the report. |
| **Free cash flow after honest sustaining capex; FCF conversion** | Operating cash flow − *normalised maintenance* capex. Then FCF conversion = FCF / net income, and cash return to shareholders (dividends + buybacks) vs FCF. | Conversion near or modestly above 1.0x on a steady-state network is credible. Conversion well above that for several years, alongside falling capex/revenue, is deferral rather than performance. Distributions persistently exceeding FCF are being debt-funded — say so. | FCF is the number the equity story is usually told in, and the number most easily manufactured here. Recomputing it at normalised sustaining capex is the core act of diligence in this sector. Also check whether buybacks are funded from cash generation or from incremental leverage; both are legitimate, but only one is repeatable. |
| **Incremental margin / operating leverage** | Change in operating income / change in revenue, year over year, ideally ex-fuel and ex-one-offs. | Roughly 50–65% incremental margin on volume growth in a fluid network is typical; decremental margins on the way down are usually gentler because a large share of cost is fixed but a meaningful share (crews, fuel, car hire) is variable. | Quantifies the network effect. A high incremental margin tells you the next ton over an existing network is very profitable — which is why volume defence matters more than it looks, and why shedding "low-margin" traffic under a margin programme can be value-destructive if that traffic was contributing above its variable cost and feeding density. |
| **Service metrics and regulatory exposure** | Filed/published performance data (velocity, dwell, cars held, first-mile/last-mile service, missed switches), customer complaints and formal proceedings, rate case activity, demurrage and accessorial revenue as % of total, and safety data — derailments and accident rates per million train miles. | Demurrage/accessorial income should be a small, stable share of revenue. Accident rate flat-to-improving. Any formal regulatory service proceeding is a material event. | Service quality is the leading indicator of both volume and regulation. Deteriorating fluidity brings shipper complaints, regulator hearings, emergency service orders and, historically, structural rule changes (rate relief, switching access) that permanently reduce pricing power. A jump in demurrage and accessorial revenue often means the *railroad's own* congestion is being billed to customers — flattering revenue while the underlying network worsens, and reliably attracting regulatory attention. |
| **Land, right-of-way and non-freight income** | Acreage and route miles owned vs leased; trackage and haulage rights granted or received; real estate, fibre/telecom easement, pipeline crossing and lease income; land sale gains in the P&L. | Non-freight income small and consistent; land sale gains should be excluded from adjusted OR and from any run-rate earnings. | The right-of-way is an irreplaceable asset — contiguous corridors through built-up land cannot be assembled again at any realistic cost, which is the ultimate source of the duopoly. It also carries optional value (easements, data conduit, redevelopment) that never appears at fair value on the balance sheet. But land gains dropped into the revenue or other-income line flatter OR and earnings; strip them, then value the optionality separately if it is material. |
## How to value companies in this sector
Railroads are enterprise-value businesses with real debt, real assets and long-lived cash flows. Unlike financials, the whole EV toolkit applies — but every multiple must be read against the OR trajectory, the ROIC-WACC spread and the honesty of the capex line, or it is just a number.
**Primary — EV/EBITDA and EV/EBIT, read against operating ratio trajectory and ROIC.** Class I railroads have historically traded in a broad EV/EBITDA band around the low-to-mid teens, with the position inside that band determined almost entirely by (a) where OR is versus where the market believes it can get to, and (b) whether ROIC exceeds WACC and is widening. A railroad at 68% OR trading at a discount is not cheap if there is no credible plan to get to 60%; a railroad at 58% OR is not expensive if the improvement is real and price is still compounding above rail inflation. Always carry **EV/EBIT alongside EV/EBITDA**, because depreciation here is a genuine economic cost and the D&A gap between an old, heavily depreciated network and a recently rebuilt one is large.
**Capital-intensity-adjusted multiple.** Compute **EV / (EBITDA − normalised maintenance capex)**. This is the honest version of the cash multiple and it is where capex deferral gets caught: a company whose EV/EBITDA looks cheap and whose capex-adjusted multiple looks expensive is under-investing. Do this before quoting any FCF yield.
**DCF with an explicit, separately-stated maintenance capex line.** The right structure for a network: model volume by commodity (with coal on an explicit decline path), core price separately from fuel surcharge, OR trajectory with an argued endpoint, and **maintenance capex as its own line that does not fall when the model wants free cash flow**. Long asset lives and stable demand make a DCF more reliable here than in most cyclicals, provided the terminal-year capex is at least depreciation and realistically above it. Sensitivity on core price versus rail cost inflation is the most informative sensitivity in the model.
**Replacement value of the network.** Estimate what it would cost to rebuild the route structure — right-of-way acquisition, grading, tunnels and bridges, track, signalling, yards and terminals — and compare to EV. For dense corridors through developed land the answer is usually "you could not, at any price," which is the quantitative expression of the moat and a hard floor under long-run value. Use it as a cross-check and a downside anchor, never as a target price: replacement value says nothing about whether current returns justify the asset.
**FCF yield — only after maintenance capex is honestly identified.** State explicitly in the report what sustaining capex you assumed, how you derived it (capex/depreciation history, physical renewal disclosures, fleet age, management's own maintenance-vs-growth split) and what the yield would be at that level versus at reported capex. A FCF yield quoted off a deferral year is a number that will not repeat.
**P/E — usable, but cycle- and OR-aware.** Rail earnings are cyclical on volume and mechanically noisy on fuel surcharge lag. Normalise: mid-cycle volume, core price at trend, fuel surcharge neutral, no land gains. A trough-volume P/E looks high and a peak-volume P/E looks low; both mislead in the usual direction.
**Sum-of-the-parts where the entity is mixed.** Many rail-linked companies are not pure networks: an operator may combine rail haulage, terminals and warehouses, a coastal shipping arm, a trucking fleet and a 3PL contract-logistics business, each with different capital intensity and different multiples. Value the rail-linked network business on its own metrics, the asset-light logistics arm on a services multiple, and terminals/real estate on an asset or yield basis. Reporting one blended EV/EBITDA across that mix is one of the most common errors in this sector, particularly in India.
**Scenario the operating ratio explicitly.** Because a single point of OR on a large network is a very large absolute profit number, the valuation is far more sensitive to the OR endpoint than to volume assumptions. Build three cases — OR holds, OR improves by the amount management guides, OR reverts because the improvement was service-funded — and show what each does to equity value. If the bull case requires an OR the network has never achieved and the fluidity metrics are deteriorating, say so in one sentence rather than burying it in a sensitivity grid.
**Control and M&A value.** Rail transactions are priced on network fit — the value of a contiguous corridor to an adjacent owner, of a short line to the Class I it feeds, of a terminal to an operator that already controls the lane — rather than on standalone multiples. Merger approval is a genuine gating risk in North America, where the competitive standard is demanding and conditions can include granting access to the very corridors that create the moat. Do not capitalise an unapproved merger's synergies.
**Cost of capital deserves more care than usual.** Long asset lives, high fixed costs and a duopoly position argue for a lower equity risk premium than the industrial average; regulatory exposure to rate and access rules argues the other way, and the effect is asymmetric because regulation bites hardest exactly when returns are highest. For cross-border networks, model each country's cash flows and discount rate separately rather than blending — a network spanning several currencies and regulatory regimes is not one asset.
**Cross-checks.** EV per route mile / per track kilometre against comparable transactions; EV per carload or per TEU of annual throughput; dividend and buyback sustainability against FCF at normalised capex; and a reverse-valuation test — solve for the OR and the core-price spread implied by today's price, then ask whether the network's own history and fluidity metrics support it.
**Do not use:** P/B as a primary method (book value bears no relation to network worth); EBITDA multiples quoted without a capex adjustment; FCF yield on reported capex; trucking or shipping sector multiples as a benchmark; or any comparison of margin, asset turnover or capital intensity against asset-light freight brokers and forwarders.
## Peer set construction
A valid comparable shares the *asset model, regulatory regime, geography and commodity mix* — not the word "rail". The commonest error in the sector is comparing an operator that runs trains on someone else's tracks with a railroad that owns the tracks.
**Splits that must never be mixed:**
- **Infrastructure-owning railroads vs train operators running on third-party track.** The first owns the right-of-way, bears the maintenance capex and captures the network rent; the second buys haulage and is a margin-taker with a fraction of the capital intensity, a completely different OR, and no pricing moat. This is the single split that matters most, and it is exactly the line between a US Class I and a listed Indian container train operator.
- **Class I mainline railroads vs short lines and regional holding companies.** Short lines run higher operating ratios, feed the Class I networks, have far shorter hauls, and are often valued on acquisition-roll-up logic rather than network economics.
- **Freight-only vs mixed freight-and-passenger networks.** A network carrying a socially-priced passenger obligation has policy-set tariffs, a cross-subsidy running through the freight rate, and an operating ratio that is not comparable to a commercial freight carrier's at any level.
- **Commodity-mix cohorts.** A coal-and-bulk-heavy network, an intermodal-heavy network and a chemicals/manifest-heavy network face different pricing regimes, different cyclicality and different asset needs. Match the mix before comparing OR.
- **Regulatory regimes.** US STB-regulated, Canadian, Mexican, EU open-access (where track and operations are legally separated and multiple operators compete on the same infrastructure), Indian state-monopoly-with-private-participation. Open-access markets have structurally lower operator margins and no geographic duopoly — never benchmark them against a vertically-integrated North American network.
- **Networks with different length of haul and density.** A long-haul, low-density transcontinental network and a short-haul, high-density regional network have structurally different revenue per carload, crew cost per ton-mile and terminal intensity. Length of haul alone can explain most of an apparent revenue-per-unit gap.
- **Rail operators vs the rail supply chain.** Wagon and locomotive manufacturers, signalling and EPC contractors, wagon-leasing financiers and rolling-stock lessors are different sectors with different playbooks. Their fortunes correlate with rail capex, not rail freight economics.
- **Pure networks vs diversified logistics groups.** If rail is under roughly half of EBITDA, it is a logistics company with a rail arm; do the SOTP and compare the segments separately.
**Also align:** fiscal year end (Indian companies April–March, most North American railroads calendar); accounting regime (Ind-AS vs US GAAP vs IFRS — lease treatment and capitalisation policies for track renewal differ, and capitalise-vs-expense choices on rail grinding, ties and rebuilds move OR by real amounts); gauge, electrification and double-stacking capability, which cap the productivity ceiling; network density and average length of haul; and the degree of vertical integration into terminals and last-mile trucking.
**Build the peer comparison on operating metrics first, financials second.** Velocity, dwell, car miles per day, GTMs per employee, gallons per thousand GTM and capex per route mile are more comparable across networks than margins are, because they are physical. Two carriers with the same OR and very different fuel efficiency or crew productivity are not equally good businesses — one has better price or mix hiding worse operations. Rank on the physical metrics, then explain the financial gap.
**Two comparisons that are always wrong and are made constantly:**
- **Railroad versus trucking company.** A trucker rents its right-of-way from the taxpayer, turns its asset base several times faster, has near-zero barriers to entry, and no pricing power beyond the spot market. Comparing margin, asset turnover, capex intensity or ROCE across the two says nothing. They meet only at the competitive boundary — truckload rates cap intermodal pricing — and that is a demand relationship, not a valuation comparison.
- **Railroad versus dry-bulk or container shipping.** Shipping is a global, near-commodity, order-book-driven business where capacity can be added by anyone with a shipyard slot and returns are destroyed by newbuild cycles. A rail network cannot be replicated at all. Rail cyclicality is industrial-demand cyclicality; shipping cyclicality is supply cyclicality. The frameworks share almost nothing.
Aim for 4–8 peers. There are very few large listed railroads globally, so a thin peer set is normal — when it is, lean harder on the company's own 5–10 year history and say so explicitly in the report. Where the peer set has to cross regulatory regimes to reach a usable size, benchmark operating metrics (velocity, fuel efficiency, GTMs per employee) rather than margins and multiples, and state the limitation.
## Sector-specific red flags
- **Operating ratio improving while velocity falls, dwell rises and trip plan compliance deteriorates.** The defining failure mode of the sector. Margin has been taken out of service quality, and the bill arrives as lost volume, regulatory intervention and an expensive recovery. Always read fluidity metrics in the same breath as OR.
- **Capex falling as a percentage of revenue, or capex/depreciation drifting below ~1.0x, while FCF hits records.** Deferred maintenance dressed as cash generation. Corroborate with physical disclosures: rail miles and ties installed, bridge programme, average locomotive and car fleet age. A shrinking renewal programme on a network with flat volume is a three-to-five-year problem being created today.
- **Headcount and locomotive cuts that leave no surge capacity.** Furloughing crews and storing locomotives improves OR immediately and looks like discipline — until volume returns and the network cannot absorb it, producing a service meltdown, emergency hiring at premium cost, and customer defection to truck. This has happened repeatedly and is the classic PSR over-shoot.
- **Yard, siding or line closures presented purely as cost savings.** Closing a hump yard, lifting a siding or embargoing a branch permanently removes optionality and capacity that cannot be cheaply restored, and it is irreversible in a way headcount is not. Ask what the closed asset was doing and what happens to that traffic when volume recovers.
- **Volume shed deliberately to improve the ratio.** Walking away from low-rate but density-feeding traffic raises OR arithmetically and can lower absolute profit and long-run network value. Check whether revenue per carload rose because price rose or because the cheap freight was pushed off. Ask what the incremental margin on the departed traffic actually was.
- **Shifting equipment off balance sheet to flatter capital metrics.** Selling and leasing back locomotives or cars raises ROIC and cuts capex while raising equipment rents inside the operating ratio. It is a financing decision dressed as an operating one; reverse it before comparing to peers.
- **Rising demurrage, storage and accessorial revenue.** Frequently a symptom of the railroad's own congestion being billed to shippers. It flatters revenue and OR while the network deteriorates, and it is a reliable precursor to complaints and regulatory scrutiny.
- **Revenue growth that is mostly fuel surcharge.** Decompose every reported growth number into volume, mix, core price and fuel. Surcharge-driven growth is margin-neutral at best and OR-dilutive arithmetically.
- **Core pricing at or below rail cost inflation for several years.** The moat is eroding — either to truck competition, to a resurgent competing route, or to regulatory pressure. This is the quietest of the red flags and the most consequential.
- **Concentration in a single commodity, single customer or single corridor.** A coal-heavy network facing utility retirements, a grain network exposed to one export programme, an automotive-heavy corridor exposed to one manufacturer's plant decisions. Concentration converts a customer-level event into a network-level one.
- **Slow orders rising, or the disclosure of them disappearing.** Slow orders — speed restrictions imposed on sections of track pending repair — are the most direct physical evidence of deferred maintenance. They lengthen transit times, consume crew hours and compound congestion. Where disclosed, track the trend; where not disclosed, treat velocity deterioration on a stable volume base as the proxy.
- **Fleet age rising while capex falls.** An ageing locomotive fleet consumes more fuel per GTM, fails more often (each failure stranding a whole train), and eventually forces a bunched replacement cycle at a time not of management's choosing. Same logic for the car fleet and for bridges.
- **Safety and derailment trend deterioration.** Beyond the human and environmental cost, a serious incident triggers regulatory response, litigation, remediation cost, insurance repricing and reputational damage with shippers and legislators, and is often the visible consequence of the deferred maintenance the numbers were hiding.
- **Buybacks funded by leverage while the network under-invests.** Leverage rising, capex falling, EPS rising on a shrinking share count — an arithmetic improvement in per-share metrics with no improvement in the business. Track capex/revenue and net debt/EBITDA on the same chart as EPS.
- **Adjusted OR carrying persistent "one-off" adjustments.** Restructuring charges every year, recurring land sale gains booked above the operating line, casualty reserve releases. Recompute OR on a clean, consistent basis over five years.
- **Regulatory escalation.** Rate case losses, an adverse reciprocal switching or open-access rule, a revenue adequacy finding used against the carrier, service-related emergency orders, or a merger condition imposing open access on a corridor. Each permanently changes the pricing equation; none of them appears in this year's ratios.
- **Management incentives tied predominantly to operating ratio.** Directly encourages the deferral and volume-shedding behaviours above. Read the remuneration disclosure and check whether service, safety and volume metrics carry real weight.
- **A change in capitalisation policy, depreciation lives or a new depreciation study, arriving alongside an OR improvement.** Moving renewal spend from expense to capex, or extending asset lives, improves OR and earnings with no change in the physical railroad. Check the accounting policy note whenever a step-change in margin has no corresponding step-change in an operating metric.
- **Withdrawal or redefinition of a previously disclosed operating metric.** A carrier that stops publishing trip plan compliance, changes how velocity is measured, or moves to a "new methodology" for dwell in the same period the number turns is telling you something. Rebuild the series on the old definition where possible.
- **Growth capex justified by volume that has not been contracted.** New terminals, sidings and capacity projects should have identifiable committed volume or a specific bottleneck behind them. Speculative capacity on a network with flat volume earns a return below WACC for a very long time.
- **Pension, casualty and environmental liabilities under-reserved.** Long-lived unionised networks carry large legacy obligations — personal injury and occupational claims, environmental remediation on old yards and rights-of-way, and defined-benefit pensions. These sit off the operating ratio and surface as cash calls. Check the reserve trend and the discount-rate assumption.
- **Contract structure hiding the pricing story.** Long-term contracts repricing on a lagged index look like stability until the index turns; a large share of revenue coming up for renewal in one year is a concentrated repricing event. Ask what proportion of the book reprices when, and on what index.
- **Interchange and haulage disputes with connecting carriers.** A network dependent on a competitor's terminal, bridge or trackage rights for a major lane has a cost and service exposure it does not control. Check trackage-rights agreements and their renewal terms.
- **A "one-time" service recovery programme that recurs.** Recovery cost booked as unusual in consecutive years means the network is running at or beyond capacity as a normal state, not suffering an event.
- **India-specific — dependence on a policy-set haulage rate.** For operators running over Indian Railways, a haulage charge revision, an empty-flow or terminal-access policy change, or a diesel/electric traction cost pass-through decision can reset the entire margin structure in one notification, with no warning in the financials.
- **Customer countervailing power building.** Large shippers responding to poor service by building private fleets, signing trucking capacity, relocating plants off-rail or lobbying for access remedies. Each is a permanent, not cyclical, loss of pricing power, and each is visible in trade press and regulatory filings long before it appears in volume.
- **India-specific — terminal and land economics carrying the earnings.** If a rail-linked operator's profit increasingly comes from terminal handling, warehousing or land monetisation rather than from freight movement, it is becoming a real-estate and infrastructure business. That may be fine, but it must be valued as one, and the rail volume story should stop being the headline.
- **India-specific — EXIM concentration and empty running.** A container train operator whose volume is concentrated on one port pair, or whose empty-running ratio is deteriorating, is carrying a cost problem that revenue per TEU alone will not reveal.
## Cycle and structural context
**The freight cycle is industrial, not consumer.** Volumes track industrial production, housing starts, auto builds, agricultural harvests and export programmes rather than retail sentiment, and intermodal tracks the container import cycle with a port-to-inland lag. Because a large share of cost is fixed, profit is more cyclical than volume, and OR worsens on the way down before management can adjust crew and locomotive levels. Judge a network by its through-cycle OR and ROIC, never by a peak or trough year.
**Seasonality and weather are real operating variables, not noise to be smoothed.** Winter operations on northern networks reduce train length (air brake limitations in extreme cold), slow velocity and raise cost; spring flooding, hurricanes and wildfires close corridors outright; the grain harvest and export programme concentrate demand into specific quarters. A quarter-on-quarter OR deterioration in a severe winter is not a franchise event. Conversely, a mild year flatters the comparison. Always compare like quarters and check what the weather did.
**Service recovery has its own cycle, and it is expensive.** Once a network goes fluid-to-congested, recovery requires hiring and training crews (a months-long lead time), returning stored locomotives to service, and temporarily running the railroad less efficiently to clear backlogs. Cost rises before volume returns, so OR worsens for several quarters after the demand recovery has begun. Analysts who read that as structural deterioration sell the recovery; analysts who read a congestion-driven revenue spike as strength buy the peak.
**Coal is in structural decline and must be modelled as run-off, not cycle.** Utility retirements, gas substitution and renewables have permanently reduced domestic thermal coal volumes in developed markets; export metallurgical and thermal coal is more durable but price- and trade-policy-dependent. Coal's disproportionate contribution to operating profit means its decline compresses margin faster than it compresses revenue. Do not model a coal recovery into a terminal value.
**Intermodal is the growth vector and the truck-competitive one.** Rail's structural fuel efficiency per ton-mile gives it a large cost advantage over trucking on long hauls, which is why intermodal has taken share for decades. But its price ceiling is set by truckload economics: when truck capacity is abundant and spot rates collapse, intermodal loses both price and share; when trucking is tight or driver supply is constrained, intermodal price and volume both rise. Model intermodal against the truck cycle, not against GDP. Service reliability is the binding constraint on further share gain — shippers pay for predictability, and a network with poor trip plan compliance cannot convert truck freight regardless of price.
**Density is the network effect, and it is the reason volume defence matters.** A rail network's cost per ton-mile falls with traffic density on a corridor: the same track, signalling, crew base and terminal absorb more freight. That is why the incremental margin is high, why losing a lane to truck is worse than the revenue loss suggests, and why a network that has thinned its traffic to improve a ratio has quietly raised its own unit cost. Density also explains why the strongest corridors are almost impossible to attack — a challenger would need volume to be cost-competitive and needs to be cost-competitive to win volume.
**Network effects and duopoly geography are the moat.** Freight rail in North America is effectively a set of regional duopolies with interchange between them; most large shippers have access to one or two railroads, and only a minority are genuinely dual-served. This geography — not brand, technology or scale in the ordinary sense — is what produces sustained above-inflation pricing. It is also why the regulatory question is always about *access*: reciprocal switching, trackage rights and merger conditions all attack pricing power by manufacturing a second option for captive shippers.
**Land and right-of-way are irreplaceable and under-stated.** Contiguous corridors through developed land cannot be reassembled. That is simultaneously the source of the moat, the floor under asset value, and the reason regulators treat railroads as quasi-utilities rather than ordinary competitors.
**Precision scheduled railroading (PSR) — the central operating debate.** PSR runs freight on fixed schedules with longer, less frequent trains, points-based operation rather than hub-and-spoke switching, and aggressive removal of cars, locomotives, yards and headcount. Applied well it delivered large, durable OR improvements and genuinely faster asset cycles. Applied dogmatically it removed the buffer capacity that absorbs demand surges and weather, degraded first-mile/last-mile service, pushed customers to truck, and drew regulatory intervention. When analysing a PSR programme, ask three questions: did velocity and car miles per day improve alongside the OR (real productivity) or deteriorate (extraction); did volume grow or shrink over the same period; and did capex per route mile hold. Margin gains that pass all three are structural. Margin gains that fail them are a loan.
**The common carrier obligation is an asymmetry worth naming.** Where it applies, a railroad must provide service on reasonable request over its lines — it cannot simply refuse unattractive freight the way a trucker can decline a load. That obligation is the political price of the franchise, it constrains how far a margin programme can go in shedding traffic, and it is the hook through which regulators intervene when service fails. A margin story that depends on walking away from customers is running against a legal duty as well as against network density.
**Regulatory posture moves with service quality.** The historical pattern is consistent: deregulation and consolidation produce a period of strong pricing and returns; service failures and shipper complaints then produce regulatory push-back on access and rates. Always check what is currently in flight at the regulator before extrapolating today's core pricing.
**Safety and labour rulemaking is a slow-moving cost variable with step changes.** Crew-size mandates, hours-of-service rules, train-length and inspection requirements, hazardous-materials routing and signalling mandates all arrive through rulemaking rather than negotiation, and each can reset the cost base or the productivity ceiling in one step. They tend to tighten after high-profile incidents, which is precisely when a carrier with a deteriorating safety record is least able to absorb them. Check what is in the rulemaking pipeline before extrapolating a productivity trend that depends on crew or train-length flexibility.
**Consolidation is nearly complete in North America and only partly begun elsewhere.** Decades of merger activity have left a small number of very large networks, each with regional pre-eminence and interchange dependence on the others. Further large combinations face a demanding competitive standard and conditions designed to preserve or create shipper choice, which caps the upside from consolidation-driven pricing. Short-line roll-ups remain a live source of value, but the economics there are acquisition-multiple arbitrage and feeder-traffic capture, not network rent.
**Inflation is broadly a friend, with a lag.** Rail contracts and tariffs commonly escalate on cost indices, the asset base is already built at historical cost, and the replacement cost of competing capacity rises with inflation — so a network with genuine pricing power passes cost through and widens the gap to a truck alternative whose driver and equipment costs rise faster. The lag is the risk: labour settlements and fuel move before indexed rate increases catch up, so a rapid inflation acceleration compresses margin for several quarters before it helps.
**Longer-term structural questions to score explicitly:** decarbonisation (rail is the low-emission long-haul mode, a genuine tailwind, offset by exposure to fossil freight); battery and hydrogen locomotives and their capex implications; autonomous trucking as a potential long-run threat to intermodal's cost advantage; nearshoring and shifting trade lanes redrawing which corridors matter; and, in India, the modal-share reversal that dedicated freight corridors are explicitly designed to achieve.
## India vs global notes
The structural difference is fundamental and must be stated at the top of any Indian rail analysis: **there is no listed Indian equivalent of a Class I railroad.** The network is owned and operated by the state. Listed Indian companies participate in rail freight as *users* of that network, not owners of it, so almost none of the network-rent economics above transfers directly to them.
| Dimension | India | US / global |
|---|---|---|
| Who owns the network | Indian Railways, a departmental undertaking under the Ministry of Railways — a vertically integrated state monopoly over track, signalling, traction and most rolling stock. Not listed, not investable directly. Its own operating ratio is published in the Railway Budget documents and has historically hovered near or above the mid-90s, because freight tariffs cross-subsidise below-cost passenger fares | Investor-owned Class I railroads own their right-of-way, track and terminals outright, are freight-only, and are among the most profitable large transport assets in the world |
| How listed companies participate | Container train operators licensed to run container rakes over IR track (categories by geography, since the 2006 liberalisation), private freight terminals and inland container depots, wagon investment and leasing schemes (general-purpose wagon investment, liberalised wagon investment, private freight terminal policies), automobile freight train operators, and multimodal/3PL groups with a rail arm | Direct ownership of the network; short lines and regionals feed the Class I system under interchange agreements |
| The economics of a listed Indian rail-linked operator | Buys **haulage** from Indian Railways at policy-set rates and resells door-to-door service. Owns rakes, containers, terminals and trucks, not track. Therefore: far lower capital intensity than a Class I, no right-of-way moat, no core pricing power over the underlying haul, margin squeezed between IR haulage revisions and customer rates, and a business that is fundamentally a **terminal, asset-utilisation and multimodal margin business** — closer to a logistics operator than a railroad. Never apply a Class I operating ratio, capex/revenue or ROIC framework to it | Network rent, common carrier obligation, captive shipper pricing, regulated rate reasonableness, enormous sustaining capex |
| Key operating metrics for Indian operators | TEUs and rakes handled; EXIM vs domestic mix; realisation and margin **per TEU**; empty running ratio (both directions); rake turnaround days; double-stacking share (constrained by overhead electrification on many routes); terminal throughput and land bank; haulage cost as % of revenue; port-pair concentration | RTM/GTM, carloads, OR, velocity, dwell, core pricing, gallons per thousand GTM, capex/revenue |
| Dedicated freight corridors (India-specific) | The Eastern and Western DFCs are purpose-built high-axle-load freight-only corridors intended to raise average freight speed, allow double-stack containers and longer/heavier trains, and reverse rail's declining share of national freight (the rail modal coefficient has fallen over decades to well under a third). Their commissioning changes transit times, cost per TEU and route economics for private operators — model this as a structural shift, not a cyclical one, and check corridor-by-corridor which of an operator's lanes actually benefit | No direct analogue; capacity comes from incremental double-tracking, siding extensions and terminal investment on existing corridors |
| Regulator / policy | Ministry of Railways sets freight tariffs, haulage charges, terminal access policy and wagon scheme terms administratively — policy risk arrives as a circular, not a proceeding. No independent economic rail regulator with tariff-setting power comparable to the STB. SEBI governs disclosure for the listed operators | US STB: common carrier obligation, rate reasonableness proceedings (stand-alone cost and simplified tests, with a revenue-to-variable-cost threshold before jurisdiction attaches), annual revenue adequacy determination, reciprocal switching rules, merger review with a competition-enhancing standard. Canada has interswitching and maximum grain revenue entitlement; the EU mandates infrastructure/operations separation and open access |
| Accounting and reporting | Ind-AS; ₹ crore and lakh; fiscal year April–March; quarterly results with investor presentation and concall. Promoter holding matters, and several rail-linked entities have government promoters, which brings policy alignment and policy risk simultaneously. Operating KPIs are disclosed less consistently than in North America — build the series from concalls and presentations | 10-K / 10-Q on EDGAR under US GAAP; calendar year; weekly carload and performance data filed with the STB and published by the AAR, giving near-real-time volume and service visibility that has no Indian equivalent |
| Valuation convention | EV/EBITDA and P/E on the operator, with lease and terminal assets examined explicitly; SOTP where rail, trucking, warehousing and freight forwarding coexist; land and terminal real estate valued separately. Replacement-value-of-network logic does **not** apply — the operator does not own the network | EV/EBITDA and EV/EBIT against OR trajectory and ROIC; DCF with explicit maintenance capex; replacement value of the network as a downside anchor |
**One further India-specific caution.** Because the listed operators sit downstream of a state monopoly, their earnings quality depends on a policy relationship rather than on a contractual or regulated one. There is no rate case to bring and no independent tribunal to appeal a haulage revision to. That argues for a lower terminal multiple than the operating metrics alone would suggest, and for explicit scenario work on haulage cost per TEU. Equally, the same relationship can deliver step-change benefits — corridor commissioning, wagon scheme liberalisation, terminal policy easing — that no financial model would have forecast from history. Score policy in both directions rather than assuming continuity.
**Also note (India):** rail-linked capital goods and EPC companies (wagon and coach builders, track and signalling contractors, project execution arms) move with the government's rail capex budget, not with freight volumes, and belong in a different playbook. Rolling-stock financing entities are lenders, not operators. Passenger-side listed entities (ticketing, catering, tourism) share a name and almost nothing else with freight economics — do not put them in a rail freight peer set. Wagon-leasing and wagon-investment scheme participants earn a haulage rebate or assured-rake economics rather than freight margin, so their return profile is closer to a lease yield than to an operating business, and the analysis should say which of the two the company actually is. Where a group spans several of these — operator, terminal owner, wagon investor, road fleet — insist on segment disclosure before writing any consolidated multiple.
## Checklist
- [ ] Establish first whether the company **owns the network** or runs trains on someone else's — the entire framework depends on the answer.
- [ ] Confirm the entity is a freight network or operator, not a rail EPC, rolling-stock manufacturer, lessor or passenger-services company — route elsewhere if so.
- [ ] Build the physical operating-metric comparison before the financial one; explain any margin gap the physical metrics do not support.
- [ ] Check slow orders, locomotive and car fleet age, and bridge programme status as direct evidence on maintenance adequacy.
- [ ] Identify interchange, trackage-rights and terminal dependencies on connecting carriers, and their renewal terms.
- [ ] Model each country's cash flows and discount rate separately for a cross-border network.
- [ ] Report operating ratio, not OPM; compute reported, adjusted and ex-fuel-surcharge versions over at least five years.
- [ ] Decompose revenue change into volume, mix, length of haul, core price and fuel surcharge before scoring growth.
- [ ] Compare core price growth to rail cost inflation over 3–5 years — this is the moat test.
- [ ] Pull the commodity mix; compute coal as % of revenue and % of profit; model coal as secular run-off.
- [ ] Read velocity, terminal dwell, car miles per day and trip plan compliance *before* reading the margin; check whether OR improvement was accompanied by better or worse fluidity.
- [ ] Check train length, train starts, GTMs per employee and locomotives in service vs stored — is productivity real, or is capacity simply gone?
- [ ] Separate fuel *efficiency* (gallons per thousand GTM) from fuel *surcharge recovery lag*; strip the timing effect from margin commentary.
- [ ] Normalise for owned-versus-leased equipment and car hire/equipment rents before any cross-carrier OR or ROIC comparison.
- [ ] Compute capex as % of revenue and capex/depreciation; find the maintenance-vs-growth split and corroborate with rail miles, ties installed and fleet age.
- [ ] Re-run FCF at a normalised sustaining capex level and state the assumption explicitly; never quote FCF yield off reported capex alone.
- [ ] Compute ROIC and compare to WACC across a full cycle, not a single year; note the regulatory implication of persistently high returns.
- [ ] Compute EV/EBITDA, EV/EBIT and EV/(EBITDA − maintenance capex); read all three against the OR trajectory.
- [ ] Check incremental margin on volume growth and ask what the shed traffic's contribution margin actually was.
- [ ] Check whether buybacks and dividends are funded by cash generation or by rising leverage while capex falls.
- [ ] Review safety and derailment trends, service complaints, demurrage/accessorial revenue share, and any live regulatory proceeding.
- [ ] Check the safety and labour rulemaking pipeline for anything that resets crew size, train length, inspection or hours-of-service economics.
- [ ] Confirm intermodal assumptions against the truckload cycle, not against GDP.
- [ ] Read the remuneration disclosure: is management paid on OR alone, or also on service, safety and volume?
- [ ] List any yard, siding or line closures and state what capacity and optionality was permanently removed.
- [ ] Strip land sale gains, restructuring and casualty items; rebuild a clean five-year OR and earnings series.
- [ ] Check whether the common carrier obligation applies and whether the margin plan is compatible with it.
- [ ] Look for evidence of customer countervailing action — private fleets, modal shift, access lobbying — in trade press and regulatory filings.
- [ ] India: identify haulage-charge dependence, EXIM vs domestic mix, realisation per TEU, empty running ratio, rake turnaround, double-stacking eligibility and DFC exposure by lane.
- [ ] India: confirm the entity is a freight operator, not a rail capital goods, EPC, leasing/finance or passenger-services company — route elsewhere if so.
- [ ] Peer set: same asset model, regulatory regime, geography, commodity mix and consolidation basis — stated explicitly; where the peer set is thin, lean on own-history and say so.
- [ ] Read the property accounting and depreciation policy notes; flag any capitalisation-policy or asset-life change coinciding with a margin improvement.
- [ ] Separate operating income growth, volume growth and EPS growth in the report so buyback-driven per-share growth is visible.
- [ ] Check legacy liabilities — pension, personal injury/casualty, environmental remediation on yards and rights-of-way — and their reserve trends.
- [ ] Ask what share of the revenue book reprices in the next 12 months, on what index, and what happens if that index turns.
- [ ] Adjust quarter comparisons for weather and harvest seasonality before calling a trend.
- [ ] Scenario the operating ratio explicitly (holds / guided improvement / service-funded reversion) and show the equity value impact of each.
- [ ] Never benchmark against trucking, shipping or asset-light forwarders on margin, turnover or capital intensity — state this in the report if a screener has done it for you.
- [ ] State where in the freight and truck cycle this sits, and what regulatory change is in flight.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |