Reserves, PV-10, and the upstream NAV model

Why upstream needs its own valuation framework

Take a software company with a hundred million dollars of EBITDA, apply a peer multiple, and you have a defensible answer. Do the same to an oil and gas producer and you have made a hidden assumption that will not survive a follow-up question: that the company can keep producing at its current rate. It cannot. Every barrel it sells is a barrel it no longer owns, and the wells producing today decline on a curve that is steep in the first year and flattens after. A producer that spends nothing watches its production fall, so the EBITDA in the denominator of your multiple is not a run rate. It is the top of a slope.

That is the entire argument for the net asset value model, and interviewers ask about it because it is the cleanest test of whether a candidate has thought about the asset or memorized a methodology. The NAV values the company by projecting what each category of reserves will produce over the life of the assets, applying a price deck, subtracting the costs and capital required to get those barrels out, tax-effecting the result, and discounting it back. There is no terminal value, because the reserves eventually run out.

The reserve categories

Everything in the model keys off the reserve report, and the report classifies volumes by how certain they are to be produced economically. Getting these categories straight is table stakes.

Proved reserves are those that engineering and geological data demonstrate with reasonable certainty are recoverable under existing economic and operating conditions. Proved splits into three sub-categories. Proved developed producing volumes are flowing right now from completed wells. Proved developed non-producing volumes sit behind pipe or await a connection, so the well exists but the volumes are not currently being sold. Proved undeveloped volumes require capital that has not yet been spent, on locations that have been identified but not yet built.

Probable reserves are less certain than proved but more likely than not to be recovered. Possible reserves are less certain still. The rough convention is that proved corresponds to about a ninety percent confidence level, proved plus probable to about fifty percent, and proved plus probable plus possible to about ten percent, though the governing definitions come from regulatory and engineering standards rather than from intuition.

Only proved reserves appear in standardized regulatory disclosure, which is why probable and possible volumes are a matter of company-provided or bank-estimated data rather than something you can pull cleanly from a filing.

The reserve report and who writes it

The reserve report is the central document in upstream valuation and in energy M&A and A&D deals, and it is prepared or audited by independent petroleum engineering firms rather than by the company alone. Understanding what the report is and is not saves you from a common mistake.

The report estimates recoverable volumes and their associated economics under a specified set of price and cost assumptions. Change the price assumption and the reserve volumes themselves change, not just their value, because volumes that are uneconomic at a low price do not qualify as reserves at all. This is why a price collapse produces negative reserve revisions across the sector, and why a price recovery produces positive ones, without any physical change in the rock. Candidates who understand this can answer one of the sector's better questions: how can a company's reserves fall when it made no operational mistake.

The report is also produced on a defined effective date and goes stale. In a transaction, the buyer's engineers will typically re-engineer the properties with their own assumptions rather than accepting the seller's report at face value, and the gap between the two views is often where the negotiation actually happens.

PV-10 and the standardized measure

You will be asked about PV-10, and the correct answer has two halves.

PV-10 is the present value of estimated future net revenues from proved reserves, discounted at ten percent, computed using a mandated historical average price and before deducting income taxes. Its virtue is comparability: because every producer computes it identically, differences between companies reflect the assets rather than the assumptions. The standardized measure is the closely related figure that does deduct income taxes.

The second half is why PV-10 is not a valuation. The mandated price is a trailing average that may bear no relationship to the forward curve, so PV-10 can be far above or below what the assets are worth. It includes only proved reserves, so it ignores probable and possible volumes and any undeveloped acreage that has not been assigned reserves. It is before tax in the case of PV-10. It excludes corporate general and administrative costs and says nothing about the capital structure. Bankers use it as a reference point and a sanity check, then build a NAV with their own price deck for the answer that goes in the deck.

Building the NAV, step by step

Start with the production profile. For each reserve category, lay out expected volumes over the life of the assets, applying the appropriate decline curve. For producing wells this is largely mechanical from existing performance. For undeveloped locations you are modeling a development schedule: how many wells get built per year, when each comes online, and what type curve each follows.

Apply a price deck. Most models use the forward strip for the near years and a long-term assumption beyond it, since the curve becomes illiquid and unreliable further out. Then adjust the benchmark price to a realized price by subtracting basis differentials, which reflect the location of the production relative to the pricing hub, and quality adjustments, which reflect the specific grade of crude or the liquids content of the gas stream. This step matters more than candidates expect, because a producer in a basin with constrained takeaway capacity can receive a price materially below the headline benchmark. The price deck choice, and how hedging interacts with it, is covered in commodity hedging and price decks.

Subtract costs. Lease operating expense is the per-barrel cost of running the wells. Production and severance taxes are levied on revenue and vary by jurisdiction. Gathering, processing, and transportation costs come out either here or as part of the realized price adjustment, depending on convention, and you should be consistent so you are not deducting them twice. Corporate general and administrative expense is typically handled at the corporate level rather than per barrel.

Subtract development capital. Undeveloped reserves only become cash flow after someone spends money to build them, and the model has to include that spending at the point in time it occurs. Also include the cost of plugging and abandoning wells at the end of their lives, which is a real liability and one that a careless model omits.

Tax-effect the result, then discount. Each reserve category gets discounted at its own rate, and the discounted values sum to the asset value.

Finally, bridge from asset value to equity value. Add the value of undeveloped acreage that has not been assigned reserves, valued on a per-acre basis from comparable land transactions in the same basin. Add any other assets, such as an equity stake in a midstream affiliate or an interest in a gathering system. Subtract net debt. Add or subtract the mark to market value of the hedge book if hedge settlements are not already reflected in the cash flow lines, and be careful not to count them twice. Divide by fully diluted shares to get NAV per share.

Discount rates and risking

The most common follow-up is why you use different discount rates by category, and the answer is that they carry different risk. Producing reserves are already flowing, require no further capital, and carry mostly price and decline-rate uncertainty. Undeveloped reserves require capital that has not been committed, execution that has not happened, and sometimes infrastructure that does not exist, so there is a genuine chance the volumes never materialize on the assumed schedule or at the assumed cost. A single blended rate would systematically overvalue the undeveloped portion.

In practice, models apply a lower rate to producing volumes and a higher rate to undeveloped ones, and many also apply a separate risk factor that reduces the undeveloped volumes outright rather than relying on the discount rate to carry all of the adjustment. The reason to separate the two mechanisms is that a discount rate penalizes distant cash flows more than near ones, while the actual risk on an undeveloped location is closer to a binary probability that it gets built at all. Probable and possible reserves, when included, get risked far more aggressively, and many bankers exclude them from the base case entirely and present them as upside sensitivity.

Where the NAV breaks and what interviewers push on

The model has real weaknesses, and volunteering them is what separates a strong answer from a recited one.

It is extremely sensitive to the price deck. Move the long-term price assumption by ten dollars and the equity value can move by a multiple of that percentage, because the cost structure is largely fixed and the change flows almost entirely to equity. This is why NAV output is usually presented as a sensitivity table across prices and discount rates rather than as a single number, and why an interviewer asking "what is the company worth" in this sector should get a range with the assumptions stated.

It depends on a development schedule that management controls and can change. If a company decides to build its undeveloped locations more slowly, the discounted value falls even though the resource is unchanged.

It says nothing about the quality of management's capital allocation beyond what you assume. Two companies with identical reserves can produce very different outcomes if one consistently adds reserves at a lower finding and development cost than the other.

And it is not the right tool for the rest of the sector. A refiner, a service company, and a midstream operator have no reserve base, so they get valued conventionally, which is why the multiples used in each sub-sector differ so much. That comparison is laid out in how energy companies are valued, and the accounting policy differences that make comparables tricky are in energy accounting: successful efforts versus full cost.

The practical resolution is that bankers run the NAV as the primary methodology for producers and put trading comparables, precedent transactions, and quick screens like EV per flowing barrel of oil equivalent per day alongside it as cross-checks. When the NAV and the comps disagree sharply, that disagreement is itself the interesting finding, and usually points at something specific: an unusual decline profile, a hedge book the market is pricing differently, or an acreage position the market is not valuing.

Practice question

Walk me through how you would build a NAV for an exploration and production company.

I'd start with the reserve report and lay out a production profile for each reserve category, so proved developed producing volumes on their existing decline curves, and proved undeveloped volumes on an assumed development schedule of how many wells come online each year. Then I'd apply a price deck, usually the forward strip for the near years and a long-term assumption after that, and adjust the benchmark down to a realized price using basis differentials and quality adjustments, because a producer in a constrained basin can receive meaningfully less than the headline price. From there I'd subtract lease operating expense, production and severance taxes, and transportation, then subtract the development capital required to bring the undeveloped locations online and the eventual cost of plugging and abandoning the wells. I'd tax-effect that and discount each category at its own rate, with producing reserves at the lowest rate because they're already flowing and need no more capital, and undeveloped reserves at a higher rate plus an explicit risk factor because they might never get built. Summing those gives asset value. Then I'd add undeveloped acreage on a per-acre basis from comparable land deals, add any other assets like a midstream stake, subtract net debt, adjust for the hedge book if it isn't already in the cash flows, and divide by fully diluted shares. There's no terminal value, because the reserves deplete.

What the interviewer is listening for: The absence of a terminal value is the single detail they most want to hear, since it proves you understand the asset is finite rather than having memorized the model's name. They are also listening for differentiated discount rates by reserve category with a reason attached, and for the realized-versus-benchmark price adjustment, which is the step candidates most often skip.