Capital budgeting

In short

Capital budgeting is the process of evaluating and selecting long-term investment proposals — a new plant, a machine replacement, a product launch, an expansion — whose costs and benefits spread over many years. Because such outlays are large, largely irreversible and decided on estimated future cash flows, a wrong choice locks up capital and damages shareholder value permanently. The core work is estimating incremental after-tax cash flows and discounting them at the firm's cost of capital, then ranking proposals by NPV, IRR, PI, payback or ARR. It is a guaranteed scoring topic in B.Com and M.Com financial management papers, in CA/CS/CMA and UGC-NET syllabi, and appears in almost every finance placement round at Indian banks, NBFCs, Big Four valuation and transaction-advisory teams, and corporate FP&A roles.

Explained

What capital budgeting is and why it matters

• Definition: the planning process used to decide whether a firm's long-term investments (fixed assets, projects) are worth funding, by comparing the present value of expected cash inflows with the outlay.
• Also called capital expenditure decision or investment decision — one of the three core finance functions along with financing and dividend decisions.
• Features: large funds committed, long time horizon, high risk and irreversibility, effect on future cost structure and competitive position, difficulty of forecasting, and strategic (not merely arithmetic) content.
• Objective: maximise shareholder wealth — accept a project only if it earns more than the cost of the capital used to fund it.
• Process steps: (1) identify proposals, (2) screen against strategy, (3) estimate incremental cash flows, (4) evaluate by appraisal techniques, (5) select within the capital budget, (6) authorise and implement, (7) post-completion audit comparing actual against projected.

Types of capital budgeting decisions

• Expansion / growth: adding capacity for existing products, or entering new products and markets (diversification).
• Replacement and modernisation: swapping a worn or obsolete asset for a more efficient one; benefit appears as *cost savings*, not new revenue.
• Mutually exclusive: accepting one rules out the others (diesel genset vs solar rooftop) — rank and pick the best.
• Independent: evaluated on its own merits; all acceptable ones can be taken if funds allow.
• Contingent / complementary: one is worthwhile only if another is taken (a plant and its effluent treatment unit).
• Mandatory / statutory: pollution control, fire safety, CSR-linked assets — evaluated on *least cost*, since there is no accept/reject choice.
• Divestment / abandonment: whether to shut or sell a line — accept if sale value exceeds the PV of continuing cash flows.

Estimating cash flows — the rules that are actually examined

• Use cash flows, not accounting profits: add back depreciation and other non-cash charges.
• Use incremental (differential) flows: only what changes because of the project — "with project" minus "without project".
• Use after-tax flows throughout; the discount rate (cost of capital) is also after-tax, so interest/finance cost is never deducted again in the cash flows (that is double counting).
• Ignore sunk costs — money already spent (a completed feasibility study) is irrelevant.
• Include opportunity cost — the market rental or sale value forgone on land or a building already owned.
• Include working capital: the initial increase is an outflow at t=0 and is recovered in the terminal year; it is not depreciated.
• Include salvage value net of tax, plus any tax saved or paid on the gain/loss on sale relative to written-down value.
• Allocated overheads that would be incurred anyway are excluded; only incremental overheads count.
• Depreciation matters only for tax — it is a non-cash charge but creates a tax shield = depreciation × tax rate.
• Opening format: Initial outlay = cost of asset + installation + transport + incremental working capital − sale proceeds of old asset (net of tax).
• Operating CFAT = (Revenue − Cash operating cost − Depreciation) × (1 − t) + Depreciation, i.e. EBIT(1−t) + Depreciation.
• Terminal year CFAT = operating CFAT + after-tax salvage + working capital released.
• Inflation: discount nominal (money) cash flows at the nominal rate, or real flows at the real rate — never mix. Fisher relation: (1+nominal) = (1+real)(1+inflation).

Traditional (non-discounted) techniques

• Payback period (PBP): years taken for cumulative cash inflows to recover the initial outlay. Accept if below a management-set cut-off.
  • *Merits:* simple, understood by non-finance managers, emphasises liquidity, crude risk screen in volatile markets.
  • *Demerits:* ignores time value of money, ignores all cash flows after the payback point, no objective cut-off, measures recovery not profitability.
• Payback reciprocal (1/PBP) approximates IRR only for long-lived projects with even annual flows.
• Accounting (average) rate of return, ARR: average annual profit after tax ÷ average (or initial) investment, in per cent.
  • *Merits:* uses familiar accounting data, covers whole life, simple to compute.
  • *Demerits:* uses profits not cash flows, ignores time value, two different denominators give different answers — always state which is used.
• Discounted payback period: payback computed on *discounted* flows; fixes the time-value flaw but still ignores post-payback flows.

Discounted cash flow (DCF) techniques

• Net Present Value (NPV): PV of all cash inflows minus PV of outflows at the cost of capital k. Accept if NPV > 0; between mutually exclusive projects pick the highest NPV.
  • *Merits:* considers time value and all flows, uses the correct accept criterion (wealth addition in rupees), additive (NPV of A+B = NPV of A + NPV of B), handles changing discount rates.
  • *Demerits:* needs an estimate of k, expresses result in absolute rupees so large and small projects are not directly comparable, assumes intermediate flows are reinvested at k.
• Internal Rate of Return (IRR): the discount rate at which NPV = 0, i.e. the project's own compound rate of return. Accept if IRR > k (hurdle rate). Found by trial and error plus linear interpolation, or by the financial function IRR.
  • *Merits:* a percentage figure managers like, no need to know k to compute it, considers time value and all flows.
  • *Demerits:* assumes reinvestment at IRR (usually unrealistic); multiple or no IRR when the sign of cash flows changes more than once; can rank mutually exclusive projects wrongly where scale or timing differs; not additive.
• Modified IRR (MIRR): compounds inflows to the terminal year at the cost of capital, then finds the rate equating that terminal value to the outlay — removes the reinvestment-rate and multiple-IRR problems.
• Profitability Index (PI) / Benefit–Cost ratio: PV of inflows ÷ initial outlay. Accept if PI > 1; the right ranking tool under capital rationing because it measures present value created per rupee invested.
• Net Present Value Index = NPV ÷ initial outlay = PI − 1.
• Terminal Value method: compounds each inflow to the project's end at the reinvestment rate and compares its PV with the outlay — logically equivalent to MIRR.
• NPV vs IRR conflict arises from differences in (a) scale of outlay, (b) timing/pattern of flows, (c) project life. The crossover rate is the discount rate where the two NPV profiles intersect. NPV prevails because it measures absolute wealth and uses the market-determined reinvestment rate.
• Unequal lives: compare using the Equivalent Annual Annuity / Equivalent Annual Cost (EAC) = NPV ÷ annuity factor, or the replacement-chain (LCM of lives) method.

Risk analysis and capital rationing

• Risk-adjusted discount rate (RADR): add a risk premium to k for riskier projects — simple but the premium is subjective and compounds risk over time.
• Certainty-equivalent (CE) approach: multiply each expected flow by a CE coefficient (0–1) and discount at the risk-free rate; theoretically superior to RADR.
• Sensitivity analysis: change one variable at a time (price, volume, cost, life, k) to see how NPV responds; identifies the critical variable and the break-even value at which NPV = 0.
• Scenario analysis: change a consistent set of variables together (optimistic / base / pessimistic).
• Probability / expected NPV: ENPV = Σ(probability × NPV); risk measured by standard deviation and coefficient of variation (σ ÷ mean) for comparing projects of different size.
• Decision tree analysis: maps sequential decisions and chance outcomes with probabilities; used for phased or pilot-then-expand projects; roll back from the right to compute expected values.
• Simulation (Monte Carlo): repeated random sampling of input distributions gives a probability distribution of NPV.
• Real options: value of flexibility — option to expand, abandon, defer or switch — which a plain NPV understates.
• Capital rationing: the capital budget is capped (hard rationing by the market, soft rationing by internal policy). Procedure: rank divisible projects by PI and take them until funds run out; for indivisible projects, examine feasible combinations and choose the combination with the highest total NPV within the limit.

Indian practice, standards and statutory context

• The hurdle rate is normally the firm's WACC, often with a divisional or project risk adjustment; public-sector appraisal typically uses a social discount rate (historically around 12% in Indian project appraisal practice — confirm the current rate, as government guidance is revised).
• Depreciation differs for two purposes: books follow Schedule II of the Companies Act, 2013 (useful-life based, straight line or WDV); tax follows Section 32 of the Income-tax Act, 1961 with block-of-assets WDV rates. Cash-flow calculations must use tax depreciation, because only that drives the tax shield. Additional depreciation under Sec 32(1)(iia) has applied to manufacturing additions subject to conditions — verify the current position, as these incentives have been progressively withdrawn alongside the lower concessional corporate tax rates.
• Corporate tax: Indian companies may opt for the concessional regime under Sec 115BAA (22% plus surcharge and cess, about 25.17% effective) forgoing most incentives, or the manufacturing rate under 115BAB; the older regime carried 30% / 25% slabs. Exam questions simply state the rate — use what the question gives.
• MAT under Sec 115JB can apply under the old regime and alters actual tax outflow; unabsorbed depreciation and carry-forward losses shift the timing of the tax shield in early project years.
• Ind AS 36 (Impairment of Assets) uses the same DCF logic to compute "value in use"; Ind AS 16 governs capitalisation of the initial outlay — borrowing cost capitalisation under Ind AS 23.
• Large capex by listed companies requires board approval and disclosure under SEBI LODR; project finance appraisal by banks follows RBI norms, where DSCR, IRR and sensitivity cases are mandatory in the appraisal note.
• Government/PPP projects are appraised with economic (social) cost-benefit analysis, VGF analysis and ERR alongside financial IRR, via PPPAC/NITI Aayog processes.

Key points

• Capital budgeting — appraising long-term investments whose cash flows run beyond one year.
• Incremental cash flow — only flows that change because of the project; the single most important estimation rule.
• Sunk cost — already incurred, therefore irrelevant to the decision.
• Opportunity cost — value forgone on an owned resource used by the project; must be charged.
• Depreciation tax shield — depreciation × tax rate; the only cash effect of a non-cash charge.
• CFAT — EBIT(1−t) + depreciation; add working-capital recovery and net salvage in the last year.
• Payback period — outlay ÷ annual inflow (even flows); ignores time value and post-payback flows.
• ARR — average PAT ÷ average investment; profit-based, ignores time value.
• NPV — PV of inflows − PV of outflows; accept if positive; measures rupee addition to wealth.
• IRR — rate where NPV = 0; accept if above the cost of capital; assumes reinvestment at IRR.
• PI — PV of inflows ÷ outlay; accept if > 1; best ranking measure under capital rationing.
• MIRR — reinvests inflows at k, giving a single, realistic rate of return.
• NPV–IRR conflict — caused by scale, timing or life differences; NPV is the theoretically correct tie-breaker.
• EAC / equivalent annual annuity — converts NPV into a yearly figure so unequal-life projects are comparable.
• Capital rationing — fixed budget; rank divisible projects by PI, indivisible ones by best feasible combination NPV.
• Sensitivity vs scenario — one variable at a time vs a consistent bundle of variables.
• RADR vs certainty equivalent — raise the discount rate vs shrink the cash flows and discount at the risk-free rate.
• Post-completion audit — comparing actuals with the appraisal; curbs over-optimistic forecasting.

Formulas and examples

Example 1 — Cash flow estimation and NPV. A machine costs ₹10,00,000, life 5 years, straight-line depreciation to zero salvage, additional working capital ₹1,00,000 recovered at the end. Annual cash revenue ₹6,00,000, cash operating cost ₹2,80,000, tax 30%, k = 12%.

• Depreciation = 10,00,000 / 5 = ₹2,00,000 p.a.
• EBIT = 6,00,000 − 2,80,000 − 2,00,000 = ₹1,20,000; PAT = 1,20,000 × 0.70 = ₹84,000.
• CFAT (years 1–4) = 84,000 + 2,00,000 = ₹2,84,000.
• Year 5 CFAT = 2,84,000 + 1,00,000 (WC released) = ₹3,84,000.
• PVA factor (5 yrs, 12%) = [1 − 1.12⁻⁵]/0.12 = (1 − 0.56743)/0.12 = 3.6048.
• PV of 5-year annuity of 2,84,000 = 2,84,000 × 3.6048 = ₹10,23,763.
• PV of extra ₹1,00,000 in year 5 = 1,00,000 × 0.56743 = ₹56,743.
• Initial outlay = 10,00,000 + 1,00,000 = ₹11,00,000.
• NPV = 10,23,763 + 56,743 − 11,00,000 = −₹19,494 → marginally negative, reject at 12%. PI = 10,80,506 / 11,00,000 = 0.982.
• Note the lesson: the working capital outlay at t=0 is recovered only in year 5, and that timing difference alone turns this project down.

Example 2 — Payback, ARR, NPV and IRR on uneven flows. Project outlay ₹5,00,000; CFAT: Y1 ₹1,50,000, Y2 ₹2,00,000, Y3 ₹2,00,000, Y4 ₹1,00,000. Life 4 years, nil salvage, straight-line depreciation, k = 10%.

• Cumulative CFAT: 1,50,000 → 3,50,000 → 5,50,000. Payback falls in year 3: 2 + (1,50,000 / 2,00,000) = 2.75 years.
• Total CFAT = ₹6,50,000; total depreciation = ₹5,00,000 → total PAT = ₹1,50,000 → average PAT = ₹37,500. Average investment = (5,00,000 + 0)/2 = ₹2,50,000 → ARR = 15.0%.
• Discount factors at 10%: 0.9091, 0.8264, 0.7513, 0.6830.
• PV = 1,36,365 + 1,65,280 + 1,50,260 + 68,300 = ₹5,20,205 → NPV = +₹20,205, PI = 1.040. Accept.
• IRR trial at 13%: factors 0.8850, 0.7831, 0.6931, 0.6133 → PV = 1,32,750 + 1,56,620 + 1,38,620 + 61,330 = ₹4,89,320 → NPV = −₹10,680.
• Interpolation: IRR = 10 + [20,205 / (20,205 + 10,680)] × 3 = 10 + 0.654 × 3 = ≈ 11.96% (true IRR ≈ 11.93%; because the NPV curve is convex, linear interpolation slightly overstates IRR). Since IRR > 10%, accept — consistent with NPV.

Example 3 — Replacement decision with tax on salvage. An old machine has book WDV ₹1,00,000, remaining life 4 years, current sale value ₹1,20,000. A new machine costs ₹6,00,000, life 4 years, no salvage, and cuts annual cash operating costs by ₹2,50,000. Both depreciated straight line. Tax 30%, k = 10%.

• Sale of old: gain = 1,20,000 − 1,00,000 = ₹20,000; tax = ₹6,000 → net proceeds = ₹1,14,000.
• Net initial outlay = 6,00,000 − 1,14,000 = ₹4,86,000.
• Incremental depreciation = (6,00,000/4) − (1,00,000/4) = 1,50,000 − 25,000 = ₹1,25,000 p.a.
• Incremental CFAT = savings(1 − t) + t × incremental depreciation = 2,50,000 × 0.70 + 0.30 × 1,25,000 = 1,75,000 + 37,500 = ₹2,12,500 p.a.
• PVA (4 yrs, 10%) = 3.1699 → PV = 2,12,500 × 3.1699 = ₹6,73,604.
• NPV = 6,73,604 − 4,86,000 = +₹1,87,604 → replace. PI = 1.386.

Example 4 — Unequal lives (EAC). Machine A: outlay ₹3,00,000, CFAT ₹1,20,000 for 3 years. Machine B: outlay ₹4,50,000, CFAT ₹1,30,000 for 5 years. k = 10%.

• A: PVA(3, 10%) = 2.4869 → PV = ₹2,98,428 → NPV = −₹1,572.
• B: PVA(5, 10%) = 3.7908 → PV = ₹4,92,804 → NPV = +₹42,804.
• EAC-style annuity: A = −1,572/2.4869 = −₹632 p.a.; B = 42,804/3.7908 = +₹11,291 p.a. → B is better on both the raw NPV and the per-year basis, and only the annuity figure makes the 3-year and 5-year projects legitimately comparable.

Example 5 — Capital rationing with indivisible projects. Budget ₹10,00,000.

• PI order Q, P, R, S: take Q (3,00,000) + P (4,00,000) = 7,00,000, balance 3,00,000 — R does not fit, so add S → total outlay ₹9,00,000, total NPV ₹2,65,000.
• Combination check: Q + R = ₹9,00,000 → NPV ₹2,55,000; P + R = ₹10,00,000 → NPV ₹2,70,000; P + Q + S = ₹9,00,000 → NPV ₹2,65,000.
• Best feasible combination is P + R with NPV ₹2,70,000 — proof that strict PI ranking can be sub-optimal when projects are indivisible.

Where it is used

• Corporate finance / FP&A teams at Tata Steel, Reliance, L&T, Infosys and Maruti Suzuki: capex notes with NPV, IRR, payback and sensitivity cases placed before the board's investment committee.
• Project finance and credit appraisal at SBI, IDBI, PFC, REC, SIDBI and NBFCs: DSCR, project IRR versus equity IRR, and stress cases in the appraisal memorandum under RBI norms.
• Infrastructure and PPP — highways (NHAI/HAM and BOT models), metro rail, solar and wind parks: financial IRR against economic IRR, viability gap funding, concession-period analysis.
• Valuation, transaction advisory and due diligence at the Big Four and investment banks: DCF for enterprise value, and impairment "value in use" testing under Ind AS 36.
• Manufacturing and operations roles: machine replacement, automation, energy-efficiency and captive-solar proposals judged on payback and EAC.
• Job roles that test it directly: management trainee (finance), credit analyst, equity research associate, project finance analyst, business analyst, CA articles and audit of capex.

Interview questions

• Q: Why is NPV theoretically superior to IRR? A: NPV measures the absolute rupee addition to shareholder wealth, assumes reinvestment at the market cost of capital, is additive across projects, and always gives a single unambiguous answer.
• Q: Why do you add depreciation back to profit when computing cash flow? A: Depreciation is a non-cash charge deducted only to compute taxable profit; after tax is calculated, it is added back because no cash left the firm.
• Q: If depreciation is non-cash, why does it matter at all? A: It reduces taxable income, so it saves tax equal to depreciation × tax rate — the depreciation tax shield, which is a real cash benefit.
• Q: Should interest on the project loan be deducted in the cash flows? A: No — the financing cost is already captured in the discount rate (WACC); deducting it again double counts it. It is deducted only in a separate equity-IRR (FCFE) calculation discounted at the cost of equity.
• Q: How do you treat land the company already owns? A: Charge its opportunity cost — the market value or rental forgone — not its historical book cost, because using it for the project sacrifices that alternative.
• Q: When can a project have more than one IRR? A: When cash flows change sign more than once (a non-conventional pattern such as a mid-life overhaul outflow); then use NPV or MIRR.
• Q: Two mutually exclusive projects — one has higher NPV, the other higher IRR. Which do you pick? A: The higher-NPV project, since it adds more wealth; the conflict comes from differences in scale, timing or life, and NPV uses the correct reinvestment assumption.
• Q: How would you compare a 3-year and a 7-year machine? A: Convert each NPV into an equivalent annual annuity (or use the replacement chain over the LCM of lives) so both are expressed on the same yearly basis.
• Q: What is capital rationing and how do you allocate funds? A: A ceiling on capital available; rank divisible projects by profitability index and fund down the list, but for indivisible projects test feasible combinations and choose the one with the highest total NPV within the budget.
• Q: Payback is theoretically weak, so why do Indian firms still use it? A: It is simple, communicates liquidity and recovery speed, and acts as a crude risk filter where forecasts beyond a few years are unreliable — it is used as a screen alongside NPV, not instead of it.
• Q: How do you handle inflation in capital budgeting? A: Be consistent — either inflate cash flows and discount at the nominal rate, or use real flows with the real rate; mixing the two understates or overstates NPV.
• Q: What is the difference between sensitivity and scenario analysis? A: Sensitivity changes one variable at a time to find the critical one and its break-even value; scenario analysis changes a consistent set of variables together into best, base and worst cases.
• Q: How does working capital enter the calculation? A: The incremental investment is an outflow when it is made (usually t=0), not depreciated, and is shown as an inflow in the terminal year when it is released.

Common mistakes

• Deducting interest or loan repayment from cash flows while also discounting at WACC — double counting the financing cost.
• Including sunk costs (past feasibility or market-research spend) and ignoring opportunity costs of owned assets.
• Forgetting the working capital outflow at t=0 or its recovery in the final year, and wrongly depreciating it.
• Using book (Companies Act Schedule II) depreciation instead of tax depreciation for the tax shield, or ignoring tax on the gain/loss when the old asset is sold.
• Ranking mutually exclusive projects by IRR or by payback instead of NPV, and ranking under capital rationing by NPV instead of PI.
• Using the wrong present-value table — annuity factor where single-year discount factors are needed, or applying an annuity factor to visibly uneven flows.
• Averaging ARR on the initial investment in one part of an answer and on average investment in another, without stating the base.
• Treating a single-point NPV as certainty — no sensitivity, scenario or break-even analysis attached to a large capex recommendation.

Related topics

Time value of money · Cost of capital and capital structure · Working capital management · Mergers, acquisitions and valuation basics · Credit analysis and lending · Sources of finance: equity, debt and hybrids