What Buybacks Really Deliver: Four Companies, Three Different Tests

The signal

The useful signal is the combination of net ownership gain and post-investment cash capacity. Gross repurchase dollars alone establish neither.

The denominator can do real work

CPKC’s third-quarter earnings attributable to controlling shareholders were C$917 million, against C$920 million a year earlier. Diluted weighted-average shares fell from 911.4 million to 877.6 million. The denominator effect was therefore about 3.85%, while the earnings numerator slipped about 0.33%. Multiplying those changes produces approximately 3.51% per-share growth using the disclosed rounded inputs. The published EPS figures, C$1.04 and C$1.01, themselves imply about 3.0%; rounding explains why these two presentations do not match exactly. [1]

That is a genuine increase in the earnings claim represented by a share. It is also different from an increase in the company’s total profit. An investor should be able to hold both observations at once. Repurchases can improve the denominator without resolving pressures on margins, taxes or financing costs. The right question is whether the price paid and the remaining balance sheet justify the ownership benefit.

The strongest counterexample is a growing numerator

McDonald’s third-quarter net income rose from US$2.278 billion to US$2.486 billion, approximately 9.13%. Its diluted weighted-average shares declined from 715.9 million to 708.3 million, adding a denominator effect of approximately 1.07%. The resulting bridge is about 10.30%, close to the change implied by reported diluted EPS of US$3.18 and US$3.51. Most of the improvement came from earnings, not share-count arithmetic. [4]

The company’s release separately reported operating-income growth and differentiated reported from adjusted results. That distinction matters because a repurchase explanation should not absorb changes caused by restructuring charges, trading performance or tax. [3]

This case defeats the broad claim that rising EPS at a repurchasing company is necessarily cosmetic. It also suggests a useful reading order: understand the business’s earnings movement first, then measure how the share denominator amplified or diluted that movement.

GM shows why the numerator must be checked

GM’s release reported US$3.335 billion of third-quarter net income attributable to stockholders, but its diluted EPS reconciliation used US$3.305 billion attributable to common stockholders. The comparable prior-year common-shareholder numerator was US$1.297 billion. Investors who simply divide the headline net-income figures by diluted shares will mix definitions. The same release reported 849 million period-end shares, versus 904 million at year-end 2025. An unchanged holding would represent about 6.48% more of the outstanding share base, subject to consistent share-class treatment. [7]

GM’s stock-information page records US$5.3 billion of program repurchases across the first three quarters. Its new US$20 billion authorization is a different object: a future permission, not cash already distributed. [8]

A smaller share base can coexist with employee awards and other issuance. The difference between gross shares purchased and net shares outstanding is therefore useful, but it cannot automatically be labelled employee dilution; transaction timing and other corporate actions also matter.

Cash coverage produces a different comparison

For the first nine months, McDonald’s generated US$8.254 billion of operating cash and spent US$2.401 billion on capital expenditure. The simple remainder was US$5.853 billion. Cash repurchases of US$1.770 billion plus dividends of US$3.954 billion totalled US$5.724 billion, leaving US$129 million on this deliberately narrow bridge. Acquisitions, financing repayments and other investing flows still need separate consideration. [4]

CPKC’s nine-month operating cash was C$4.252 billion. Property additions of C$2.035 billion plus C$18 million of Meridian Speedway additions left C$2.199 billion before other investing flows. Cash repurchases of C$3.388 billion and dividends of C$644 million totalled C$4.032 billion: a C$1.833 billion distribution gap on that definition. Its filing also discloses borrowing and other cash movements. [2]

These are individual cash reconciliations, not a league table of financial health. The currencies differ, business models differ, and a company can rationally distribute accumulated cash or use financing. The gap identifies a funding question; it does not identify which dollar paid for which repurchase.

PepsiCo makes the free-cash-flow definition visible

PepsiCo’s 36-week filing reports US$7.950 billion of operating cash and US$2.182 billion of capital spending. That leaves US$5.768 billion under a CFO-minus-capex definition. Cash dividends of US$5.935 billion and repurchases of US$739 million totalled US$6.674 billion, implying a US$906 million distribution gap. Its reported free cash flow is US$5.857 billion because it adds US$89 million of property-sale proceeds; using that measure reduces the gap to US$817 million. [6]

The accompanying statements also distinguish ordinary repurchases from withholding-tax payments when employee stock units convert. Those transactions belong in a complete equity-and-cash reconciliation; the accounting compensation expense is not a substitute for either cash amount. [5]

Both cash-flow presentations can be useful. Trouble begins when an analyst silently switches between them across companies. State whether asset sales are included, preserve the issuer’s reporting period, and show financing and acquisition flows outside the recurring operating bridge.

Ownership and EPS require different share counts

Weighted-average diluted shares measure the denominator used across an earnings period. Period-end outstanding shares measure the ownership base at a point in time. A late-quarter repurchase can reduce the latter immediately while having only a modest effect on the quarter’s average. Treating those two numbers as interchangeable can make a program appear more or less effective than it was.

The ownership calculation assumes an investor keeps the same number and class of shares. It must be adjusted for stock splits and cannot compare incompatible voting or economic rights. The EPS calculation requires the earnings numerator actually attributable to the common shareholders on the same diluted basis. Preferred dividends, participating securities and convertible instruments may require adjustments.

This is why our calculator produces separate outputs rather than an all-purpose buyback score. A company can pass the ownership test and fail the cash-coverage test for one period; another can fund distributions comfortably while buying shares at an unattractive valuation. Those are different investment judgments.

A transparent stress test, not a valuation verdict

The illustrative calculator starts with common-shareholder earnings rising from 100 to 110 and diluted shares falling from 100 to 95. The earnings increase is 10%; the denominator effect is approximately 5.26%; combined EPS growth is approximately 15.79%. If period-end outstanding shares also fall from 100 to 95, an unchanged holding gains approximately 5.26% in proportional ownership. These are hypothetical units, not a forecast for any company in the study.

For the separate cash bridge, operating cash of 150 less capex of 40 leaves 110. Repurchases of 70 plus dividends of 20 use 90, leaving a cushion of 20. Reduce operating cash to 120 and keep the other inputs fixed: the bridge turns into a gap of 10. This does not predict borrowing or a dividend cut. It shows the amount that would need another source of funds, less spending or a smaller distribution.

Keep investment assumptions explicit. Cutting capex to preserve a payout might improve today’s cash bridge while weakening tomorrow’s business. Increasing capex can do the opposite. A calculator cannot settle that trade-off without evidence about the returns on the investment.

What would change the conclusion?

The evidence does not support a blanket accusation that buybacks manufacture growth. McDonald’s provides a clear earnings-growth countercase, and GM’s net share reduction represents a real change in the ownership base. Nor does it support treating every authorization as shareholder value already delivered. Authorizations, execution, dilution, cash funding and purchase price occupy different stages of the analysis.

For an existing holding, build the bridge over several years, not just the strongest quarter. Ask whether recurring earnings grow, whether net shares fall after issuance, and whether distributions fit the cash remaining after investment and contractual obligations. Then examine what the company paid relative to the future cash flows it was buying on shareholders’ behalf.

The practical decision is where to spend analytical attention. A large authorization with little net ownership gain warrants dilution reconciliation. Rising EPS with flat earnings warrants a denominator bridge. A persistent cash gap warrants a balance-sheet and investment review. None of those observations, by itself, is a buy or sell signal.

The evidence, side by side

Buyback Ownership and Funding Bridge

Illustrative assumptions only. No live prices, forecasts or valuation recommendation.

Require e 0>0, e 1>0 and all share counts>0 for EPS growth; cash quantities finite, with negative CFO allowed as an explicit cash-burn case. No missing input is treated as zero.

Scenarios to watch

Observed countercase

McDonald’s combines positive earnings growth with a smaller share denominator; do not attribute the full EPS gain to buybacks.

Illustrative cash slowdown

CFO falls from 150 to 120 with capex 40 and distributions 90 unchanged: the distribution gap changes from−20 to+10. No probability is assigned.

Method and limitations

Four purposively selected companies, not a representative market panel. We read eight issuer disclosures and used the latest periods available by 13 November 2026. EPS comparisons use matching quarterly common-shareholder numerators and diluted weighted-average denominators. Cash bridges retain each issuer’s currency and period; CPKC includes Meridian property additions. Calculations use published rounded inputs. We do not aggregate cash amounts across currencies or annualize PepsiCo’s 36-week period.

No estimate of intrinsic value, causal effect on stock prices or market-wide prevalence is made.

Cash bridges exclude acquisitions, disposals, debt service and lease financing from the central equation and require them to be examined separately.

The study distinguishes cash repurchases from accrued amounts, announced authorizations and employee award accounting. Differences in company definitions remain material.

Primary sources and evidence register

Source dates and reporting periods differ. Several documents from one issuer are one source family, not independent confirmations. The evidence download contains the observations used in this article.

Frequently asked questions

Does a buyback automatically increase my ownership?

Only if outstanding shares fall after other issuance and you retain the same holding. Gross repurchase amounts alone cannot establish that result. Compare consistent, split-adjusted period-end shares.

Does higher EPS prove the business earned more?

No. EPS can rise because common-shareholder earnings grow, the diluted denominator falls, or both. Use the numerator and denominator actually reported in the EPS reconciliation.

Does a distribution gap prove debt funded the buyback?

No. Cash is fungible. The company may use accumulated cash, disposals, borrowing or other sources. The gap identifies a question for the complete financing statement.

Why not rank the four companies?

Their currencies, fiscal periods, investment needs and business models differ. The cases demonstrate mechanisms; they do not form a comparable valuation or financial-strength ranking.