Where the parameters come from

0. What this tool is not, against the PIV 2026

References are to the Exposure Draft — Principi Italiani di Valutazione 2026, published by Fondazione OIV in December 2025. They are a draft, not standards in force: the consultation closed on 23 March 2026 and at the date of writing OIV still lists the text under “Documenti in bozza”. Nothing here is a claim of compliance with principles that are not yet in force.

This tool does not produce a valuation compliant with the PIV, and it does not produce a parere valutativo either. The second point is not a choice available to it: the ED records that IVS 2024 excluded limited scope valuations from IVS-compliant valuations, and that the PIV 2025 consequently excluded pareri valutativi from the types of engagement consistent with the PIV. The category no longer exists, so the tool cannot fall back on it.

What is missing is the whole of Part Two. A valuation under the PIV requires an engagement (II.2), agreed in writing, which must state the configuration of value to be determined; a named Esperto di valutazione who carries it out under the professional conduct requirements of II.1 and answers for it; an information base whose analysis must be consistent with the approaches chosen, the configuration of value sought and the purpose of the estimate (I.5.9), and whose objective limitations must be declared; a valuation report (II.4); and working papers (II.5). This tool has none of the five. It has an anonymous user, a form, and seven numbers.

Consequence, stated plainly: the output is an order-of-magnitude estimate produced by a machine. It is a starting point for a conversation with a valuer, not a substitute for one, and it cannot be attached to a proceeding that requires a valuation.

0b. Configuration and level of value

The ED requires that the definition of the configuration of value always be reported in the valuation (I.6.22), and that every valuation also make explicit the valuation perspective adopted (I.7.1). What this tool computes is:

• Configuration: intrinsic value — the value the fundamentals support, not an exchange price observed on a market.
• Perspective: the entity as it stands, stand-alone, with no synergies accessible to a specific acquirer.
• Procedure: asset side. The operating invested capital is valued and financial liabilities are then subtracted, as III.1.55 describes.
• Level of value: 100% of the capital, the level of value referred to the whole business in the sense of III.3.5. It is not the value of a shareholding.
• Assumption: going concern.
• Before premiums and discounts. No control premium, no discount for lack of control and no discount for lack of marketability is applied. III.4.3 makes the reason explicit: premiums and discounts have no meaning unless the base of value they are calculated from is defined. Whoever applies them takes this figure as the base and states which level of value they are moving it to.

1. Two blocks, two sources

The parameters are divided into two groups, with different sources for a precise reason.

Risk block — unlevered beta, target leverage, debt spread, equity risk premium and country risk premium. They can be measured only where prices exist, so the source must be the universe of listed companies.
Source: Aswath Damodaran, NYU Stern, Europe dataset, updated 9 January 2026. Country premiums dated 5 January 2026. Open-access data.
Risk-free rate: ECB, euro-area AAA yield curve, ten-year maturity, observation dated 3 September 2026. The AAA yield is used instead of the Italian government yield because Italy risk is already included in the country premium; including it in the base rate as well would count it twice.
Tax rate: 27.9%, comprising 24% IRES and 3.9% IRAP.

Plan block — margin, working capital, capital intensity, depreciation rate, steady-state return on capital and growth. These have no external source at all, and that is the point: they are read from the accounts of the business being valued, or asserted by whoever is valuing it. A sector median cannot describe a single company’s balance sheet, and presenting one as if it could was the defect this replaces.

EV/EBITDA has been removed from the tool. There is no reliable public evidence on transaction multiples for unlisted businesses, and listed-company multiples carry a liquidity premium that systematically overstates them. The refusal is methodological, not a gap in the data: the tool no longer promises a multiples comparison.

2. The plan parameters come from the company

This section replaces the previous derivation from Mediobanca aggregates and the size correction that rested on it.

Margin, working capital, capital intensity and the depreciation rate are read from the accounts of the business being valued:

EBITDA margin = EBITDA / revenue

Net working capital on revenue = net working capital / revenue

Capital intensity = net fixed assets / revenue

Depreciation rate = depreciation / net fixed assets

The capital intensity is also the capex required per euro of new revenue: if the business needs those fixed assets to produce that revenue today, it needs them in the same proportion to produce more. Previously these were two independent parameters, taken from two different columns of the same panel, with nothing guaranteeing that they agreed with each other.

Growth carries no default. Revenue growth in year one and in year five open at zero. The first figure the user sees is their own business at flat revenue; any growth is something they have asserted, not something the tool assumed on their behalf.

The steady-state return on capital is derived, not tabulated. It is year-five NOPAT over invested capital at the close of year five, the capital on which the following year would earn. The user may override it; the tool states which of the two applies, and warns when the derived value has been capped at the slider maximum of 25%.

3. The specific risk premium

The betas and leverage ratios come from listed companies. Applying them unchanged to an unlisted SME prices that business as though its shares traded daily. A premium is therefore added to the cost of equity, by revenue band:

Between the anchors the premium is interpolated linearly in log₁₀ of revenue, and is flat at 4.0% below the first anchor and at 1.5% above the last. The scale is logarithmic because the economic distance between two businesses a factor of five apart is the same at any level of turnover.

The anchors sit at the centre of each band, not on its boundaries. Anchoring on the boundaries would give the bottom of the medium band the small-company premium: a €10m business would have paid 4.0% instead of the 2.5% its band assigns. The geometric centre of the medium band is √(10 × 50) = √500 million, and the other two anchors sit a factor of five away from it.

The values are not rounded, and the reason is not tidiness. From that construction an exact property follows: €10m is the geometric mean of the first two anchors and €50m of the last two, so on both band boundaries the premium is exactly the arithmetic mean of the premiums of the two anchors around it — 3.25% at €10m, the mean of 4.0% and 2.5%; 2.00% at €50m, the mean of 2.5% and 1.5%. A business on the boundary pays the average of the two bands it separates, not the value of either. Rounding the centre to €25m broke that property and left a discrepancy of just over a tenth of a point — small enough to be absorbed by a generous tolerance, which is exactly how a defect survives a test.

There are no bands. A banded premium jumped by 1.5 points between €9.9m and €10.1m of revenue, which is about 12% of the value of a €10m business for a difference in turnover that means nothing economically. The labels small, medium and large survive on screen only as a reading of where the business sits on the curve, next to the premium actually applied.

Ke = rf + β relevered × (ERP + CRP) + specific risk premium

The premium is additive and is not scaled by beta: it is not market risk, so it is not multiplied by sensitivity to the market. Zero is the listed company, not the middle band. Setting the middle band to zero would give the large band a negative premium, that is a cost of capital below a comparable listed company, which is not defensible.

Calibration of the three premiums

The parameter measures the distance between the business being valued and the listed comparable that beta, leverage and spread come from. Zero is the listed company: no unlisted business reaches it, and the scale between the anchors measures how much more heavily non-listing weighs as size falls.

The conceptual anchor is III.2.4, not a statistical series. What the premium is meant to capture is set out there: the risk drivers that weigh more heavily as size falls. The three values are the professional judgement of what those drivers are worth on the Italian unlisted business, expressed on a continuous curve. They are a declared choice, of the same kind as the cap on perpetual growth, and they are stated as such.

The US decile series enter only as an order-of-magnitude check, and nothing more. In the Valuation Handbook — U.S. Guide to Cost of Capital, 2017 edition, data to 31 December 2016, the lower deciles carry premiums in the range of roughly 3% to 6% — micro-cap grouping 3.67%, the tenth decile alone 5.59% — against roughly 1% to 2% for the middle and upper deciles. That tells you a premium of four points at the bottom of the curve and one and a half at the top is the right size of number rather than off by a factor. It does not calibrate the three values, and it could not: those series are cut by market capitalisation, they are a different market, and they mix in a liquidity component treated separately there and deliberately included here.

Italian valuation practice on unlisted SMEs points the same way. It has no single citable source, and is offered for what it is: an observation of professional practice, not evidence.

The value at the upper anchor, 1.5%, is the residual non-listing component: at that level of turnover the business reaches institutional finance and a competitive market for disposals, so what remains is only the gap against a security traded on a regulated market. The central anchor, 2.5%, is the intermediate value: the gaps between the three anchors are 1.5 and 1.0 points over intervals of equal logarithmic width, that is a regular and slightly convex decline. It corresponds roughly to a 20–25% discount on value.

The series cited treat the size premium as distinct from a marketability discount, which they apply separately to equity. Here the two sit in the same field, which is why it must not be added to a DLOM on the result: that would count the same thing twice.

None of the three values has a public measurement behind it. There is no observable series of the non-listing premium for the Italian market: they are a methodological choice of whoever is valuing, declared as such, on the same footing as the cap on perpetual growth. The construction of the curve that interpolates them is exact; the three numbers are not. Anyone using the tool can replace them.

There is no free public measurement of the size premium for unlisted Italian companies. These values are a signed methodological choice anchored to the literature on size premia in the lower deciles, not a datum. Do not combine this field with a marketability or minority discount applied to the result: the two capture the same thing in part, and adding them counts it twice.

4. Update cycle

The Damodaran datasets are published in January. The risk block is regenerated once a year through the ingestion scripts, and every parameter displays the date of its own source. Dates are never generated automatically: when a source date is missing, the field remains empty and the tool says so.