How to Calculate the Altman Z-Score

The five ratios and a worked computation — from balance sheet line items to a distress zone verdict.

The the Altman Z-score is one of the few bankruptcy-prediction models that has held up over decades of out-of-sample testing. Developed by Edward Altman in 1968, it condenses a company's financial position into a single number by weighting five accounting and market ratios. Understanding the altman z score formula — not just the final number but the mechanics behind each input — is essential for anyone using it as a distress signal or as a feature in a broader model. This post walks through each ratio, the exact balance sheet and income statement line items it requires, a fully worked example, and the practical gotchas that cause misapplication.

The five inputs and where to find them

Each variable in the Z-score maps to specific financial statement line items. Getting the mapping wrong is more common than it should be.

• X1 — Working capital / total assets. Working capital is current assets minus current liabilities. Both come from the balance sheet. This ratio measures short-term liquidity relative to the asset base. A shrinking X1 is an early sign that near-term obligations are crowding out operating capacity.
• X2 — Retained earnings / total assets. Retained earnings is the cumulative net income kept in the business since inception, found in the equity section of the balance sheet. This ratio captures leverage age and profitability history. Young companies with large accumulated deficits will have a low or negative X2 regardless of current performance.
• X3 — EBIT / total assets. Earnings before interest and taxes comes from the income statement; divide by total assets from the balance sheet. This is a clean measure of asset productivity before financing and tax effects are layered in. Use operating income as a proxy only if interest income is not embedded in it.
• X4 — Market value of equity / total liabilities. Market value of equity is shares outstanding multiplied by current share price — not book equity. Total liabilities is the sum of current and non-current liabilities from the balance sheet. This is the ratio most often misapplied; using book equity systematically understates X4 for profitable firms and overstates it for distressed ones.
• X5 — Sales / total assets. Net revenue from the income statement divided by total assets from the balance sheet. This asset-turnover ratio captures how efficiently the company generates revenue from its asset base. Capital-intensive industries tend to have structurally lower X5 values, a known limitation of the original model.

The formula with coefficients

Altman estimated the following coefficients using multiple discriminant analysis on a matched sample of bankrupt and non-bankrupt manufacturers:

Z = 1.2(X1) + 1.4(X2) + 3.3(X3) + 0.6(X4) + 1.0(X5)

The zone thresholds for the original public-firm model:

• Z > 2.99: Safe zone — low distress probability.
• 1.81 < Z < 2.99: Grey zone — elevated uncertainty; monitor closely.
• Z < 1.81: Distress zone — historically high probability of financial stress within two years.

The coefficient on X3 (3.3) is the largest, meaning asset profitability is the single most influential component. X4's coefficient (0.6) is the smallest among the five, but because market cap can be a large number relative to liabilities, it still contributes meaningfully in practice.

Worked example — Hypothetical company

The following figures are entirely fabricated for illustration. Assume a mid-size manufacturer, Hypothetical Industrial Corp., with the balance sheet and income statement snapshot below as of its most recent fiscal year-end.

Selected financials (all figures in $M, hypothetical):

• Current assets: $180
• Current liabilities: $120
• Total assets: $500
• Total liabilities: $300
• Retained earnings: $60
• EBIT: $55
• Net revenue: $620
• Shares outstanding: 40 million at $9.50 per share → market cap = $380

Computing each ratio:

• X1 = (180 − 120) / 500 = 60 / 500 = 0.120
• X2 = 60 / 500 = 0.120
• X3 = 55 / 500 = 0.110
• X4 = 380 / 300 = 1.267
• X5 = 620 / 500 = 1.240

Variable	Ratio value	Coefficient	Contribution
X1 — Working capital / assets	0.120	1.2	0.144
X2 — Retained earnings / assets	0.120	1.4	0.168
X3 — EBIT / assets	0.110	3.3	0.363
X4 — Market cap / liabilities	1.267	0.6	0.760
X5 — Sales / assets	1.240	1.0	1.240
Z-Score			2.675

Z = 0.144 + 0.168 + 0.363 + 0.760 + 1.240 = 2.675. This falls in the grey zone (1.81–2.99). The company is not in acute distress, but the score is close enough to the distress boundary that a deterioration in EBIT margin or a share-price decline would push it into the red. Notice that X5 (asset turnover) and X4 (market cap coverage) are carrying the score — weak retained earnings and thin liquidity are partially offset by a still-functioning revenue engine and a market cap that covers liabilities.

The private-firm Z′-score

The original model requires a market cap, which private companies do not have. Altman published an adapted model — the Z′-score — specifically for private firms. The key change: X4 uses book value of equity instead of market cap in the numerator. The coefficients are also re-estimated:

Z′ = 0.717(X1) + 0.847(X2) + 3.107(X3) + 0.420(X4′) + 0.998(X5)

Thresholds shift accordingly: above 2.9 is safe, below 1.23 is distress, and the grey zone sits between. When screening a universe that mixes public and private issuers — or when using book-value proxies for speed — the Z′-score is the appropriate formulation. Applying the public-firm coefficients with book equity in X4 is a common and silent error.

Data gotchas and look-ahead risk

Several implementation errors are common enough to call out explicitly:

• X4 must use market cap, not book equity. For public firms, plugging in shareholders’ equity from the balance sheet instead of market cap produces a mechanically different model. Results are not comparable to published Z-score benchmarks.
• Point-in-time financials. When building a historical screen or a corporate default events dataset linkage, use the financials available as of the date you are evaluating — not the restated or as-reported figures from a later filing. Using any data that was not yet public at the evaluation date introduces look-ahead bias and inflates backtest performance.
• Fiscal year-end vs. trailing twelve months. X5 in particular can swing materially if a company changed its fiscal year or made a large acquisition mid-period. Trailing-twelve-month revenue against year-end assets can distort asset turnover.
• Industry limits. Altman's original sample was U.S. manufacturers. Financial firms, REITs, and utilities all have capital structures that make the ratios — especially X1 and X4 — structurally incomparable. Applying the Z-score to a bank is not meaningful without significant adaptation.
• Currency and consolidation. For multinationals reporting in a foreign currency, ensure all five inputs are from the same consolidated statement. Mixing segment-level figures with group-level assets produces a ratio with no interpretive value.

Putting the Z-score to work

The Z-score is most useful as a screening tool, not a standalone verdict. A score in the grey zone warrants deeper analysis — cash flow coverage, covenant headroom, near-term debt maturities — rather than a binary pass or fail. To build a distress screener with meaningful signal, the Z-score is a natural first-pass filter: anything below 1.81 goes into a watchlist for fundamental review.

Linking Z-score trajectories to a corporate default events dataset allows you to validate the model's predictive window on your own universe and time period — and to calibrate how much weight to give a grey-zone reading versus a score that just crossed below the distress threshold for the first time. Trajectory matters as much as the absolute level: a score that has fallen from 3.4 to 2.3 over three quarters is a very different signal from one that has been stable at 2.3 for two years.