Separate mandatory needs from free cash
A cash balance is not a free decision budget. Contractual payments, seasonal financing and maintenance may come first. Before discussing an acquisition, restore the minimum needs of the existing business. If equipment wears out, refusing replacement can create a temporarily attractive flow at the price of later deterioration.
Compare alternatives at one date, in one currency and after a compatible set of costs. Do not compare pre-tax project profit with after-tax interest savings without stating the difference. A promised acquisition synergy is not as reliable as a contractually specified payment. Different risk needs explanation, not just a ranking of percentages.
Hypothetical production expansion
A company can invest 50 million roubles in a project producing 15 million at the end of each of four years, with no residual value. At an educational rate of 10%, the present value is 47.55 million and NPV is −2.45 million. Total receipts exceed the investment, but after time value the project does not meet the stated return requirement.
If the same project produces 17 million each year, present value is 53.89 million and NPV is +3.89 million. This does not prove every such project should be accepted. Check flow ranges, cost overruns, interdependence and the rate’s basis. All figures are fictional and do not concern the October 2026 market.
A buyback does not automatically create value
Reducing share count can raise EPS without improving operations. A buyback spends money that could earn income or cover obligations. If the company overpays or replaces stable financing with risky financing, higher EPS is not enough justification.
An acquisition needs the same discipline: separate price, target stand-alone flows, integration costs and expected synergy. Do not count brand value twice if its effect is already in future margin. NYU material on moving from operating assets to equity helps separate business value from cash and other non-operating assets.[98](#source-cv-equity-bridge)
Incentives and testing promises
Ask what management is rewarded for. Revenue growth, EPS and economic return on investment can point decisions in different directions. This is not proof of abuse; it is a reason to inspect incentives and the history of major decisions. Compare the original project budget with the outcome and promised launch dates with actual dates.
For a series of projects, keep a journal that does not erase failures. A good latest project does not cancel an expensive earlier one, and a large acquired business need not be efficient. Show organic results separately from changes in group composition, or scale growth can hide weak returns.
The result is a map of alternatives, constraints and success conditions. It must allow the conclusion that data are insufficient. Returning cash, reducing debt and expanding are not best in every circumstance. This guide gives no preferred dividend policy, target debt level or instruction to buy a company with a buyback; it evaluates decision logic, not the corporate-action label.
What to do with a partly completed project
After investment begins, people may continue only because much has already been spent. For a new decision, separate sunk costs from future payments and receipts that can still change. This does not automatically mean stopping: completion may be the best remaining option. But the original budget must not replace comparison of available actions.
Check the cost of doing nothing. Declining expansion may preserve flexibility or may lose required access or product quality. Name these consequences in the model rather than adding abstract “strategic value” after every negative calculation.
A good post-analysis note explains which actions remain available, which constraints now exist and which data management will use to assess the result. That makes responsibility measurable without promising a mistake-free choice in advance.
NPV versus IRR: scale can change the ranking
NPV is the sum of future net flows discounted at the selected rate, including the negative initial investment. IRR solves NPV=0, but its percentage does not measure absolute value added. MIT treats NPV, internal rate and payback as different project-evaluation criteria.[3](#source-gv-mit-projects) Compare them under explicit constraints, risk and timing.
Project A requires 10 million today and returns 13 million in one year. B requires 100 million and returns 120 million at the same time. With equal risk and a 10% educational rate, IRRs are 30% and 20%, while NPVs are 1.82 and 9.09 million. If mutually exclusive and the budget permits either, a maximum-NPV objective prefers B despite its lower IRR. The extra 90 million from A to B produces an extra 107 in one year: incremental IRR≈18.89%, NPV≈7.27 million.
With a 10-million budget, B is unavailable; ranking does not create financing. Nor may ten copies of A replace B without evidence. For independent divisible projects with a simple initial cost, profitability index PI=PV of future receipts/initial investment is useful. Here PI is about 1.182 and 1.091; it shows relative result but does not erase B’s larger NPV. With indivisibility, dependence or later negative flows, PI alone is insufficient.
Ordinary and discounted payback
Another educational project is −100 today and +40 at the end of each of the next three years, with no residual value. Cumulative receipts are 40, 80 and 120. On the literal calendar, the initial 100 is recovered at the end of year three. The often-published 2.5 years assumes even receipts inside year three: 2+(100−80)/40. That is a separate assumption, not the exact date of year-end payments.
At 10%, discounted receipts are about 36.36, 33.06 and 30.05 million, total 99.47. There is no discounted payback within three years and NPV≈−0.53 million. Do not record it as “zero years” or extend nonexistent receipts beyond the project. At zero rate the two paybacks coincide; with r≤−1 the annual discount formula is invalid in this domain. Payback ignores receipts after the recovery point and does not replace full-project NPV.
What changes with modified internal rate
MIRR separately sets a financing rate f for negative flows and a reinvestment rate z for positive flows. For n equal periods, MIRR=(FV of positive flows at z / |PV of negative flows at f|)^(1/n)−1. This is an explicit educational construction, not the unique root of the original IRR. For −100,+60,+60, z=10%, f=8% and n=2, FV=60×1.10+60=126, PV of costs=100 and MIRR≈12.25%. Here f does not change the result because all cost is at the initial date; later costs would differ.
Without positive receipts, negative investment, or with zero duration, this MIRR construction does not apply. Do not replace its error with zero return. Financing and reinvestment parameters need separate support; MIRR does not remove scale differences or guarantee reinvestment at the stated rate. PI and MIRR glossary definitions require separate entries; this connecting section is not presented as their publication.