Altman Z-Score for Belgian SMEs: a practical guide with worked examples
Altman's Z-Score is the academic benchmark for predicting business failure. Discover the Z'' version adapted for unlisted Belgian SMEs — its formula, thresholds and a fully worked example based on annual accounts filed with the NBB.
In brief
Altman's Z-Score sorts a company into one of three zones: safe, grey or distressed. The Z'' version (1995) is specifically calibrated for unlisted, non-manufacturing SMEs — the profile of the vast majority of Belgian companies. With four ratios pulled from annual accounts filed at the NBB, you can decide in under a minute whether a counterparty deserves a supplier credit line, an invoicing cap or enhanced audit scrutiny.
Why an academic score today?
Edward Altman's model was published in 1968 for listed US industrial firms. It was revised twice, and the Z''-Score version (1995) is the one that applies to private, non-manufacturing companies — almost every Belgian SME. Why is it still the reference fifty years later?
- It is peer-reviewed and cited in dozens of academic papers
- It was empirically validated on thousands of real bankruptcies
- It is explainable: four ratios, a linear combination, a threshold. No black box
- It is citable in audit work, credit committees, or in front of a regulator (FSMA, CFI)
Exactly what you need when you have to justify a credit refusal or back up an enhanced due diligence file. For the legal KYC context in Belgium, see our article on the 9/18/2017 anti-money-laundering law.
The Z''-Score formula
Z'' is a weighted combination of four ratios:
1
Z'' = 6.56·X1 + 3.26·X2 + 6.72·X3 + 1.05·X4
Where:
| Variable | Definition | Source in NBB annual accounts |
|---|---|---|
| X1 | Working capital / Total assets | (Permanent capital − Fixed assets) / Total balance sheet |
| X2 | Retained earnings / Total assets | Equity (rubric 10/15) as a proxy |
| X3 | EBIT / Total assets | Operating result (rubric 9901) / Total balance sheet |
| X4 | Equity / Total liabilities | Equity / Debts (rubric 17/49) |
The largest coefficient (6.72 on X3) shows that operating profitability carries the heaviest weight — consistent with a credit manager's intuition: a business that loses money on its core activity is structurally at risk, whatever its balance sheet looks like.
The three zones
| Z'' | Zone | Reading |
|---|---|---|
| > 2.60 | Safe | Low one-year failure probability |
| 1.10 – 2.60 | Grey | Uncertainty, heightened vigilance |
| ≤ 1.10 | Distressed | Significantly elevated failure probability |
The grey zone is deliberately wide: Altman prefers a cautious verdict over a false positive that would refuse credit to a healthy customer.
Fully worked example
Take a Brussels-based services BV whose 2024 annual accounts (filed at the NBB Central Balance Sheet Office) show:
| Rubric | Amount (€) |
|---|---|
| Total balance sheet (20/58) | 1,200,000 |
| Equity (10/15) | 380,000 |
| Debts (17/49) | 820,000 |
| Fixed assets (20/28) | 450,000 |
| Long-term debts (17) | 280,000 |
| Operating result (9901) | 95,000 |
Computation:
- Permanent capital = 380,000 + 280,000 = 660,000 €
- Working capital = 660,000 − 450,000 = 210,000 €
- X1 = 210,000 / 1,200,000 = 0.175
- X2 = 380,000 / 1,200,000 = 0.317
- X3 = 95,000 / 1,200,000 = 0.079
- X4 = 380,000 / 820,000 = 0.463
1 2 3
Z'' = 6.56 × 0.175 + 3.26 × 0.317 + 6.72 × 0.079 + 1.05 × 0.463
= 1.148 + 1.033 + 0.531 + 0.486
= 3.20
Verdict: safe zone. This BV can be treated as a reliable trading partner over a 12-month horizon. Operating profitability (X3) is sound and the balance sheet is balanced.
Limitations and caveats
Automation via the Company Belgium API
Rather than manually retrieving annual accounts from the NBB website and computing ratios in Excel, the Company Belgium API returns the pre-computed Z''-Score for the most recent fiscal year available, along with the X1–X4 components and the zone assignment:
1
GET /api/companies/{enterpriseNumber}/financial-intelligence
The response contains an academicScores block with the Altman score, the Conan-Holder score, and a consensus between the two models with a confidence level (high, medium, low) — exactly what an auditable deliverable should look like.
For access details, see the BCE Company Belgium API documentation.
Going further
- Cross the Z'' with the Conan-Holder score to raise confidence in the verdict
- Learn to read an NBB balance sheet to understand where the rubrics come from: full schema vs abridged schema
- Integrate Z'' into an automated KYC verification workflow for regulated professions — or directly into the dashboard of an accountant automating their portfolio
- Understand the difference between EBITDA, EBIT and operating result — key to computing X3 correctly
Frequently asked questions
Which version of Altman's Z-Score should I apply to a Belgian SME?
For almost all Belgian companies (SRL, SC, small SA), use the Z''-Score version published by Altman in 1995. It is designed for unlisted, non-manufacturing businesses. The original 1968 version and the 1983 Z'-Score apply only to listed industrial firms and will give a biased verdict on a Belgian services SME.
Where can I find the data to compute the Z''-Score of a Belgian company?
In annual accounts filed at the NBB Central Balance Sheet Office. For each closed fiscal year, retrieve rubrics 20/58 (total assets), 10/15 (equity), 17 (long-term debts), 17/49 (total debts), 20/28 (fixed assets) and 9901 (operating result). The Company Belgium API exposes this data and computes Z'' automatically on the most recent fiscal year.
Is the Z''-Score reliable for a company less than 3 years old?
Only partially. Ratio X2 (retained earnings / total assets) assumes an accumulated track record of profits. For a very young company X2 will be mechanically low or negative if the first years are loss-making, which penalises the score. Best practice is to complement Z'' with a qualitative review of the business plan and monthly tracking of operational KPIs.
What if the Z''-Score classifies a company in the grey zone?
The grey zone (1.10 to 2.60) signals statistical uncertainty: neither clearly healthy nor clearly distressed. Best practice is to cross-check with a second model (e.g. Conan-Holder), verify the 3- to 5-year trend, and add qualitative indicators: director seniority, recent events in the Belgian Official Gazette, possible presence on sanctions or PEP lists.
Comments
Related articles

Belgian chart of accounts 2026 (PCMN/MAR): the essential guide
The Belgian chart of accounts (PCMN/MAR) remains in 2026 the backbone of every Belgian company's bookkeeping. This article walks through the 7 classes, the « micro / small » size thresholds applicable to financial years closed in 2026, classic coding pitfalls, and how a modern invoicing module can generate PCMN-compliant entries without manual input.

Enterprise number vs establishment unit: the difference in the BCE
The enterprise number (BCE/KBO) identifies a legal entity, the establishment unit identifies a physical activity point. Confusing the two wastes minutes in API integration and hours in commercial analysis. Here is exactly what differentiates them and how to use them correctly.

The Conan-Holder model: predicting SME failure within 3 years
The Conan-Holder score is the French-speaking equivalent of Altman's Z-Score, specifically calibrated on French and Belgian SMEs. Discover its formula, its five ratios, its interpretation bands and a worked example to assess the probability of failure of a Belgian company over three years.
