What You Actually Need to Know About Oil And Gas Economics
The basics are simple on paper and almost nobody uses them correctly. A derrick operator or completion engineer might calculate reserves in their head, but when it comes time to justify a capital expenditure to someone in a finance department, the economics become the battlefield. That is where most people stumble. At its core, the discipline involves estimating the cash flows from hydrocarbon production over the life of a field or well, discounting them back to present value, and comparing the result against other potential investments. The standard metrics are NPV (net present value), IRR (internal rate of return), and payback period. You will also encounter EUR (estimated ultimate recovery), decline curve analysis, and break-even price per barrel. These terms appear in every boardroom presentation, but the way they are actually calculated in practice varies wildly between companies and even between analysts at the same firm. I have seen two economists working the same geological data come up with NPVs that differed by over 40 percent, and neither was clearly wrong. The difference came down to which operating cost assumptions they loaded into the model and how they handled the tax structure for deepwater versus onshore shale plays.
The industry standard model architecture has remained stubbornly Excel-based for over three decades. There are specialized platforms like SAP, CMG, or Schlumberger's Eclipse, but for actual economic evaluation, nearly everyone still builds a spreadsheet. That is not because it is ideal. It is because the flexibility of a well-constructed Excel model beats a black-box application every time, and because the audit trail matters when you are defending a multi-million dollar project approval. Here is the part most tutorials skip. The discount rate is where decisions actually get made. A 10 percent discount rate and a 12 percent discount rate can flip a project from economic to uneconomic depending on when the cash flows materialize. Early-maturing shale wells suffer more from a higher discount rate than a long-lead-time deepwater project, even if the absolute dollar amounts look similar. Junior analysts often treat the WACC as a fixed input. Senior people know it is a lever that gets tuned to match the company's strategic appetite for risk in a given year. I ran into a specific problem a few years back involving a tight oil play in the Permian. The geological team had provided a stochastic output with P10, P50, and P90 case volumes, but my finance counterpart wanted a single deterministic number for the board deck. When I pulled the NPV using the P50 case, the project barely cleared the hurdle rate. The P90 case was negative. The P10 case was very attractive. The standard approach would be to just pick the P50 and move on.
What I did instead was build a quick Monte Carlo overlay on the spreadsheet using Latin Hypercube sampling across the key uncertain variables — decline rate, commodity price path, and operating cost escalation. I set it so the commodity price correlated with the production volume since higher pressure reservoirs generally produce more in the early years. The result was a probability distribution rather than a single NPV number. We ended up reporting that there was a 68 percent likelihood of exceeding the hurdle rate and a 23 percent chance the project would return less than the cost of capital. The project got approved, but with a reduced initial drilling budget and a requirement to reassess after the first five wells came online. This approach took about four hours to build and validate, compared to the usual two days spent arguing over which deterministic case to present. It also forced the decision makers to think in ranges rather than false precision, which is where the real risk lives anyway. Another thing that trips people up involves the treatment of lifting costs. In many models, operators assume a flat dollar-per-barrel operating cost throughout the life of the well. That does not reflect reality. Water cut increases over time, which means more volume to lift per unit of hydrocarbon produced. Gas lift or artificial lift requirements kick in at some point. These costs escalate non-linearly, and a model that assumes flat OPEX will systematically overstate the NPV of mature assets. I typically build a water-cut escalation curve based on field analogs and let the lifting cost track that curve. It usually shifts the break-even price up by two to five dollars per barrel depending on the field profile.
Tax regimes are another area where simplified models cause real errors. The United States has a fluctuating depletion allowance and state-level severance taxes that vary by basin. International projects face royalty structures that scale with price — the Oman model, for example, has a sliding scale that can double the effective tax burden when prices spike above sixty dollars per barrel. I have seen projects look profitable at the base commodity price assumption and then vanish entirely when you stress-test them at higher price scenarios because the tax bite accelerates faster than the revenue increase. If you are building a model from scratch, start with the production profile, not the financial assumptions. The production shape determines everything else. Get the decline curve right using Arps or a more modern segmented exponential model if you are dealing with unconventional reservoirs, and the economics tend to sort themselves out. You can adjust discount rates and cost assumptions afterward. But if your production forecast is wrong, no amount of tweaking the financial parameters will fix it. The biggest bottleneck in this work is data quality. You will routinely receive reserve estimates from geoscience teams that contain assumptions about future technology performance, unproven pipeline capacity, or optimistic water handling constraints. None of those are lies, but they are forward-looking guesses wrapped in numbers that look factual. I always cross-check the key inputs against at least two independent sources before I build the financial model. If the company has produced similar wells in the same play, I pull those actual results. If not, I find analog fields and adjust for depth and temperature differences. It adds time upfront but saves hours of rework later when someone asks why the actual cash flows do not match the forecast.
There is also a quiet skill to structuring the scenario table. Most models have a base case, an upside, and a downside. That is insufficient for any asset where commodity price volatility matters, which is all of them. I add a price-decline intersection scenario where low prices coincide with steeper-than-expected declines, because that is the regime where most projects actually fail. It rarely looks good, and it should not. The point is to know the worst plausible outcome before the board asks about it. You can find templates and sample models scattered across industry websites, but most of them are either too simplistic for real decision-making or tied to specific software that most companies do not use. The practical workaround is to build your own framework based on a few well-constructed example models from reputable sources, then strip out the proprietary assumptions and rebuild around your own cost structures and fiscal terms. This usually takes about two weeks of iteration for someone with basic spreadsheet proficiency and a week or two of actual field data exposure. The discipline rewards people who are honest about uncertainty. The models will never be more accurate than the inputs, and in this industry the inputs are always guesses with error bars. The best economists I have worked with were the ones who could clearly articulate what their model did not know, not the ones who produced the prettiest single-number output. That is where the actual value sits.
