ROE vs ROCE vs ROIC: What Each One Actually Measures

All three measure "how much profit does this company generate relative to the money invested in it" — they just disagree on which money counts. ROE only counts shareholders' money. ROCE counts shareholders' money plus long-term debt (all "capital employed" in the business). ROIC counts only the capital actually deployed into operations, deliberately excluding idle cash and non-operating assets. The gap between them, more than any single number, is usually where the real story is.

**ROE (Return on Equity)** = Net profit ÷ Shareholders' equity. Purely a shareholder's-eye view: for every rupee of equity capital in the business, how much profit came back. It says nothing about how that equity is supported — a company can pump ROE up simply by taking on more debt, since debt doesn't appear in the denominator at all.

**ROCE (Return on Capital Employed)** = EBIT ÷ (Total assets − Current liabilities), or equivalently EBIT ÷ (Equity + long-term debt). It measures returns on *all* the long-term capital funding the business, debt and equity together, using pre-interest, pre-tax profit — which makes it capital-structure-neutral in a way ROE isn't.

**ROIC (Return on Invested Capital)** = NOPAT (Net Operating Profit After Tax) ÷ Invested capital, where invested capital typically excludes non-operating cash and investments. It's the most surgical of the three — trying to isolate the return the actual operating business generates, stripped of financing effects and idle balance-sheet clutter.

A company with ₹200 crore equity, ₹150 crore long-term debt, ₹300 crore net profit driven by a period of high leverage, ₹380 crore EBIT, and ₹300 crore NOPAT, holding ₹80 crore of idle cash outside the operating business:

• **ROE** = 300 ÷ 200 = **150%** — an eye-catching number, but heavily amplified by the debt sitting below it.
• **ROCE** = 380 ÷ (200 + 150) = 380 ÷ 350 = **108.6%** — still strong, but visibly more moderate once debt is counted as capital that also had to earn a return, not just a lever for equity.
• **ROIC** = 300 ÷ (350 − 80) = 300 ÷ 270 = **111%** — similar to ROCE here since idle cash was relatively small, but on a company sitting on a large cash pile, ROIC would separate cleanly from ROCE where ROCE would not.

The pattern to notice: ROE is the most easily inflated by leverage, ROCE and ROIC converge when there's little idle cash, and ROIC diverges from both when a company is sitting on a lot of cash that isn't actually working in the business.

Use **ROE** when your question is specifically "how well is this working for shareholders" — but always alongside a debt figure, since ROE alone can't tell you whether a high number reflects genuine efficiency or just leverage. Use **ROCE** as the more balanced default for comparing operational efficiency across companies with different amounts of debt — it's the ratio screener.in and most Indian equity research foreground for exactly this reason. Use **ROIC** when comparing companies with meaningfully different cash balances, or when you specifically want to isolate operating performance from financing and cash-management decisions.

• Comparing ROE across two companies without checking their debt-to-equity — a highly leveraged company can post a much higher ROE than a conservatively financed peer with an objectively better business.
• Treating a high ROCE as automatically good without checking whether it's driven by genuinely efficient operations or by a shrinking capital base (a company divesting assets can see ROCE rise even as the underlying business stagnates).
• Using ROIC and comparing it directly to a company's cost of capital without knowing what that cost of capital actually is — ROIC only becomes meaningful once it's compared to WACC.
• Assuming all three should roughly agree — a large gap between ROE and ROCE is itself a signal (usually about leverage) worth investigating, not an error to reconcile away.

Return on assets, debt-to-equity, EBIT, working capital