The Warehouse That Is Already Paid For
A manufacturer is deciding whether to launch a new product line. Its finance VP recommends approval, citing a positive NPV. The analyst's notes give these facts:
- Projected new sales EUR 8.0m per year; variable costs 65% of sales.
- Incremental fixed cash costs EUR 1.2m per year.
- 15% of sales will come from customers switching from an existing product whose contribution margin is 40% on EUR 12.0m of annual sales.
- The project uses an idle company-owned warehouse that could be leased to a third party for EUR 0.6m per year net.
- EUR 4.0m of R&D has already been spent developing the product.
- Equipment costs EUR 3.0m, depreciated straight-line over 5 years to zero, no salvage.
- Net working capital of EUR 0.5m is required and recovered at the end of year 5.
- Tax rate 25%; WACC 10%.
The VP's analysis charges the R&D against the project and ignores the warehouse and the switching effect.
1. Rebuild the incremental annual operating cash flow and the NPV, and state whether the project should proceed.
Read the worked answer
Incremental EBITDA
- New contribution: 8.0 x 35% = 2.80
- Cannibalised contribution: 12.0 x 15% x 40% = (0.72)
- Incremental fixed cash costs: (1.20)
- Opportunity cost of warehouse: (0.60) EBITDA = +0.28
Operating cash flow
- Depreciation: 3.0 / 5 = 0.60
- EBIT = 0.28 - 0.60 = (0.32)
- Tax = -0.32 x 25% = +0.08 (a shield)
- OCF = EBIT + depreciation - tax = -0.32 + 0.60 + 0.08 = EUR 0.36m per year
NPV (annuity factor at 10%, 5 years = 3.7908; discount factor year 5 = 0.6209)
- Outlay: equipment 3.0 + NWC 0.5 = (3.50)
- PV of OCF: 0.36 x 3.7908 = 1.3647
- PV of NWC recovery: 0.5 x 0.6209 = 0.3105 NPV = -3.50 + 1.3647 + 0.3105 = EUR -1.82m
Decision: reject. The project destroys value at a 10% hurdle. The VP's NPV is positive only because of the two errors below.
2. Address the VP's two treatments specifically: the R&D charge and the excluded warehouse and switching effects.
Read the worked answer
R&D is sunk. The EUR 4.0m has been spent regardless of the decision. Including it cannot change the incremental cash flows and only misallocates a past cost. Correct NPV is EUR 4.0m higher than an analysis that charges it - but the correct project NPV is still negative, so removing the R&D charge does not rescue the project. The general rule: sunk costs are excluded; opportunity costs are included; allocated fixed overheads that do not change are excluded.
Warehouse opportunity cost is real. The building has an alternative use worth EUR 0.6m per year net. Using it for the project forgoes that cash. The fact that the warehouse is already owned, fully depreciated and currently empty is irrelevant - the relevant question is the best alternative use, not the historical cost.
Cannibalisation is real. 15% of the new line's EUR 8.0m of sales displaces a product earning a 40% contribution margin, so EUR 0.72m of contribution disappears. Excluding it double-counts the benefit; a company that launches every 'profitable' line on this basis shrinks its own existing business.
Sensitivity. If the warehouse genuinely had no alternative use, EBITDA would be 0.88, OCF = 0.28 x 0.75 + 0.60 = 0.81, and NPV = -3.50 + 0.81 x 3.7908 + 0.3105 = -0.12m - still negative, but marginal enough that the cannibalisation rate (15%) and the fixed cost estimate become the deciding assumptions and should be stress-tested before a final no.
3. What single assumption would you test first, and why?
Read the worked answer
The 15% cannibalisation rate, because it is the largest single discretionary number in the build-up (EUR 0.72m of the EUR 0.28m EBITDA) and it is the one the project sponsor is most likely to have understated.
Mechanics of the test:
- At 0% switching, EBITDA rises to 1.00 and OCF to 0.28→ (EBIT 0.40) x 0.75 + 0.60 = 0.90; NPV = -3.50 + 0.90 x 3.7908 + 0.3105 = +0.12m. The sign flips.
- The breakeven switching rate sits between 15% and 0% - so the entire decision rests on a number the VP has not modelled at all.
How to pin it down: customer-level overlap data between the two products, historical switching rates from prior launches, and sales-force interviews rather than a planning assumption.
Secondary tests: the EUR 0.6m warehouse rent (verifiable with a broker), and whether any of the EUR 1.2m fixed cost is an allocation of existing overhead that would be incurred anyway - if so, it is excluded and NPV rises further.
Recommendation to the board: reject as submitted, but commission the cannibalisation estimate before treating the answer as final.