Which Dividends Are Not Covered by Cash? Free-Cash-Flow Coverage Screening in Python
August 5, 2026
What’s the question?
The conventional test of dividend safety is the payout ratio: dividends divided by net income. Anything under 60% reads as comfortable, and a screen built on that threshold will pass most of the large-cap dividend universe.
Net income is an accrual figure. It charges depreciation on assets bought years ago, and it says nothing about what a company spends on new assets this year. Dividends are settled in cash. Free-cash-flow coverage measures the same obligation against the cash actually available: operating cash flow minus capital expenditure, divided by dividends paid. A coverage ratio above 1.0 means the company funded its distribution out of what the business produced after reinvestment. Below 1.0 means the difference came from borrowing, share issuance, or the cash balance.
The practical question is which dividend payers pass the earnings test and fail the cash test, and how far the two rankings diverge across a large universe.
The approach
The universe is the current S&P 500. Each company contributes its most recent annual filing, with fiscal periods ending between April 2025 and June 2026. Built from SEC EDGAR public filings and market data.
- Pull annual net income, operating cash flow, capital expenditure, and the common stock dividend
- Keep companies that paid a common dividend and reported positive net income
- Exclude Financials and Real Estate, where the capital-expenditure line is not comparable: banks and insurers have no meaningful capex against operating cash flow, and REIT capital spending mixes maintenance with portfolio growth
- Compute the payout ratio as dividends over net income, and coverage as free cash flow over dividends
- Measure the Spearman rank correlation between the two, then isolate the companies where they disagree
That leaves 253 dividend payers.
Code
import xfinlink as xfl
from scipy.stats import spearmanr
xfl.set_api_key("YOUR_API_KEY") # free at https://xfinlink.com/signup
fields = ["period_end", "net_income", "operating_cash_flow",
"capital_expenditures", "dividends_paid_common"]
tickers = xfl.index("sp500")["ticker"].dropna().unique().tolist()
raw = xfl.fundamentals(tickers, period_type="annual", start="2024-06-01",
fields=fields, max_rows=10000)
latest = raw.sort_values("period_end").groupby("ticker").tail(1)
df = latest[
(latest["dividends_paid_common"] > 0)
& (latest["net_income"] > 0)
& latest["operating_cash_flow"].notna()
& latest["capital_expenditures"].notna()
& ~latest["gics_sector"].isin(["Financials", "Real Estate"])
].copy()
df["fcf"] = df["operating_cash_flow"] - df["capital_expenditures"]
df["payout"] = df["dividends_paid_common"] / df["net_income"]
df["coverage"] = df["fcf"] / df["dividends_paid_common"]
rho, pval = spearmanr(df["payout"], df["coverage"])
blind_spot = df[(df["payout"] < 0.60) & (df["coverage"] < 1.0)]
print(f"Spearman: {rho:.3f} uncovered: {(df['coverage'] < 1).sum()} of {len(df)}")
print(blind_spot.sort_values("coverage")[["ticker", "gics_sector", "payout", "coverage"]])Full script with formatting and visualisation: dividend-free-cash-flow-coverage-python.py
Output
=== Dividend Coverage: Cash vs Earnings ===
Universe: 253 S&P 500 dividend payers (latest annual filing, ex-Financials, ex-Real Estate)
Fiscal periods end 2025-04-25 to 2026-06-30
Spearman rank correlation, payout vs coverage: -0.756 (p=4.0e-48)
Dividend not covered by free cash flow: 43 of 253 (17.0%)
Median coverage: 2.57x Median payout: 39.8%
--- Payout ratio under 60%, free cash flow below the dividend (18 names) ---
Ticker Sector NetInc OCF Capex FCF Div Payout Cover
PCG Utilities 2,703 8,716 11,787 -3,071 277 10.2% -11.09x
CNP Utilities 1,052 2,486 4,870 -2,384 574 54.6% -4.15x
ORCL Information Technology 17,087 31,977 55,663 -23,686 5,725 33.5% -4.14x
AES Utilities 910 4,306 5,929 -1,623 501 55.1% -3.24x
ATO Utilities 1,199 2,049 3,561 -1,512 554 46.2% -2.73x
MOS Materials 541 825 1,359 -535 282 52.2% -1.89x
AWK Utilities 1,111 2,059 3,126 -1,067 648 58.3% -1.65x
EXC Utilities 2,768 6,254 8,529 -2,275 1,617 58.4% -1.41x
AEE Utilities 1,461 3,353 4,128 -775 768 52.6% -1.01x
NI Utilities 930 2,362 2,782 -420 533 57.3% -0.79x
AEP Utilities 3,696 6,944 8,453 -1,509 2,008 54.3% -0.75x
EIX Utilities 4,459 5,800 6,515 -715 1,274 28.6% -0.56x
NUE Materials 1,744 3,234 3,422 -188 511 29.3% -0.37x
PEG Utilities 2,111 3,298 3,272 26 1,258 59.6% 0.02x
LEN Consumer Discretionary 2,078 217 189 28 521 25.1% 0.05x
TRGP Energy 1,923 3,917 3,333 584 815 42.4% 0.72x
HUM Health Care 1,188 921 546 375 426 35.9% 0.88x
MMM Industrials 3,250 2,306 910 1,396 1,562 48.1% 0.89x
Sector mix of that group:
Utilities 11
Materials 2
Information Technology 1
Consumer Discretionary 1
Energy 1
Health Care 1
Industrials 1
--- Payout ratio above 100%, cash covers the dividend more than 1.5x (5 names) ---
Ticker Sector NetInc FCF Div Payout Cover
TPR Consumer Discretionary 183 1,094 299 163.4% 3.65x
CVS Health Care 1,768 7,807 3,409 192.8% 2.29x
TSN Consumer Staples 474 1,177 697 147.0% 1.69x
HSY Consumer Staples 883 1,823 1,085 122.9% 1.68x
ABBV Health Care 4,226 17,816 11,819 279.7% 1.51xWhat this tells us
The rank correlation of −0.756 means the two measures mostly agree: companies with a high payout ratio tend to have low cash coverage. The agreement is not complete. Of 253 payers, 43 did not cover the dividend from free cash flow, and 18 of those reported a payout ratio below 60%. A screen with a 60% cutoff would have passed every one of them.
Eleven of the 18 are utilities, which points at structure rather than distress. Regulated utilities spend far more on plant than they charge to depreciation, because the rate base is deliberately growing. American Electric Power earned 3,696 and distributed 2,008, a payout of 54.3%; operating cash flow of 6,944 against capital expenditure of 8,453 left free cash flow of −1,509. The dividend was funded from debt and equity issuance, and the regulator permits a return on the enlarged asset base. That is a working model, but its safety depends on capital-market access and rate-case outcomes, neither of which the payout ratio observes.
Oracle is the largest shortfall in dollar terms. A payout ratio of 33.5% looks untroubled until capital expenditure of 55,663 is set against operating cash flow of 31,977, which produces free cash flow of −23,686 and coverage near −4.1x.
3M shows a different mechanism. Capital expenditure of 910 is modest, and the gap comes from operating cash flow of 2,306 sitting well below net income of 3,250. Payout reads 48.1%; coverage is 0.89x.
The reverse error also appears. Five companies posted a payout ratio above 100% while covering the dividend more than 1.5 times in cash. CVS Health earned 1,768, distributed 3,409, and still generated 7,807 of free cash flow, for coverage of 2.29x. Non-cash charges such as impairments and intangible amortisation depress reported earnings without touching the cash that pays the dividend.
So what?
Run both ratios and treat the gap between them as the signal. Where cash coverage sits far below the earnings payout, the dividend depends on external funding, and the relevant risk is a credit and capital-markets risk rather than an operating one. Where cash coverage sits far above the earnings payout, a payout ratio above 100% is an accounting artefact and not a warning.
For capital-intensive sectors, the earnings payout ratio has almost no discriminating power: every utility in this sample passes it. Coverage separates them. Setting a minimum coverage ratio alongside a maximum payout ratio, and reviewing the disagreements by hand, catches both failure modes that a single-ratio screen produces.
Built with xfinlink — free financial data API for Python. pip install -U xfinlink
pip install -U xfinlink