Three ways to measure debt

Leverage is measured three ways in ordinary use, and the three are not versions of one another. One compares debt with book equity, one with earnings, and one compares the interest bill with the profit available to pay it. Before any of them means anything, somebody has decided what counts as debt.

The word debt has four choices inside it

Gross or net of cash. Net debt subtracts cash on the view the company could repay from the balance sheet today, though the cash may be trapped in a subsidiary or required for operations. Use net debt to ask whether borrowings could be repaid now, and gross debt for anything to do with a covenant, since a lender nets your cash only if the agreement says so.

Operating lease liabilities in or out. Under ASC 842 they sit on the balance sheet, so a total debt figure may already include them; adding them again is double counting.

Pension deficits. An underfunded obligation is a claim on future cash, measured with a discount rate the company selects within limits, and it rarely appears in any leverage ratio at all.

Preferred stock. A fixed, cumulative, redeemable instrument behaves like debt. A perpetual, non-cumulative one behaves more like equity.

Debt to equity, and the denominator's weak spot

Total debt divided by shareholders equity. Graham's chapter 14 test in The Intelligent Investor was shaped differently: for industrial companies he asked that long-term debt not exceed net current assets, a test against working capital rather than book equity. The debt-to-equity threshold of 1.0 often attributed to him is a different test on a different denominator, and a company can pass one and fail the other.

Shareholders equity is an accounting figure: what was raised plus what was retained. Years of buybacks leave a small equity balance and a high ratio in good health, and a goodwill write-down under ASC 350 pushes the ratio up in the year the impairment test failed, with no cash moving and no lender paid.

A utility at 1.43 and a technology company at 0.19 are not more and less prudent. They finance different assets against revenue of different volatility, and a regulated revenue stream supports borrowing a product cycle does not. Where equity is negative the ratio is not meaningful; where debt is exactly zero, the ratio is a real 0.0, not a missing figure.

Net debt to earnings, and interest coverage

Net debt to earnings before interest, tax, depreciation and amortisation states roughly how many years of pre-tax, pre-depreciation earnings would repay the borrowings. It is what lenders write into loan agreements. That is why it matters more than its analytical merit deserves: earnings before depreciation ignores the capital spending needed to keep earning, so true repayment capacity is lower than the ratio suggests, most of all at the asset-heavy companies that borrow the most.

Interest coverage, operating income divided by interest expense, uses an accounting profit while interest is paid in cash, and it is computed on last year's rate rather than next year's. A company with no interest expense has no coverage ratio at all. A very large number in its place would misread a debt-free balance sheet as an outstanding credit, so ValueScreener withholds the ratio in that case and reports debt to equity of 0.0 as the real measurement.

What none of the three sees

All three are levels on one day. None distinguishes a company with nothing due for six years from one holding identical ratios and a bond maturing next year. For a bank none of them means anything at all: deposits are the raw material of the business, not borrowing, and equity is regulated against risk-weighted assets.

Check yourself