Great Business, Wrong Price? What Historical Returns Reveal

Operating success does not guarantee a rewarding stock return. A company can grow earnings while its shareholders lose money, or grow earnings at a similar pace while its shares climb. The price paid at the start, the earnings ultimately delivered and the valuation investors are willing to assign at the end all matter. But even that arithmetic does not explain why expectations changed.

Paycom, Regeneron and Adobe offer three historical illustrations. Each increased trailing reported earnings over the selected two-year period. Their shareholders had strikingly different outcomes. These cases were chosen after their returns were known; they are not a representative sample or a test of whether expensive shares usually underperform.

What the comparison measures

A price-to-earnings ratio divides the share price by earnings per share. Because price equals EPS multiplied by P/E, the change in share price can be decomposed exactly into an earnings factor and a multiple factor. For Paycom, EPS nearly doubled, yet the P/E fell by roughly three-quarters. The combination produced a share-price loss of about 52%; three cash dividends trimmed the simple shareholder loss to 51.7%. Adobe's EPS growth was smaller, but its P/E edged up rather than collapsing.

This is an accounting identity, not a finding that the initial multiple caused a later return. Share prices respond to forecasts, discount rates, competition, product performance and surprises that the two endpoint ratios cannot isolate. The comparison is most useful when the business evidence is read alongside the valuation bridge rather than replaced by it.

For perspective, SPY, an S&P 500 ETF, returned about 0.8% in Paycom's window and 18.3% in Regeneron's and Adobe's windows. These are same-date simple returns including cash distributions without reinvestment, not risk-adjusted comparisons. [9]

Paycom: growth meets a reset in expectations

Paycom's trailing revenue rose from about $991 million at the beginning of its window to about $1.63 billion at the end, while its operating margin increased from 22.3% to 27.6%. Reconstructed trailing GAAP diluted EPS rose from $2.95 to $5.84. Those are substantial improvements in reported operating results. Yet the share price fell from $404.94 to $194.45. A starting P/E of 137.3x left the return highly exposed to a change in what investors expected from future growth. By the end, the P/E was 33.3x. [1]

One visible reset came after the market closed on October 31, 2023. Paycom's full-year revenue guidance range moved from $1.715 billion-$1.717 billion in its earlier release to $1.679 billion-$1.684 billion, a $34.5 million reduction in the midpoint. The shares closed at $244.97 on October 31 and $150.69 on November 1, a 38.5% close-to-close fall. The announcement preceded that reaction, but the price move cannot be assigned exclusively to the guidance change: investors also reassessed what the release implied about later growth. Moreover, Paycom's shares had already fallen about 39.5% from the selected January 2022 start to October 31, 2023. The event is therefore not an explanation for the entire two-year loss. [2]

Paycom paid three dividends totaling $1.125 per share during 2023. Adding those distributions to the exit price, without reinvesting them, changes the measured return from a 52.0% price loss to a 51.7% simple shareholder loss. The distinction is small here, but essential to a consistent comparison. [3]

Regeneron: earnings need a closer reading

Regeneron's trailing revenue increased from about $5.24 billion to $7.24 billion over its selected window. Its free cash flow, defined here as operating cash flow less cash purchases of property and equipment, also rose. Nevertheless, its operating margin slipped from 32.3% to 29.6%, and its share price declined from $462.24 to $274.13. Trailing EPS rose 80.6%, but the P/E fell from 46.4x to 15.2x. That leaves a 40.7% share-price and shareholder loss: no cash dividends were recorded in the window. [4]

The EPS figure needs qualification. The reconstructed trailing effective tax rate fell from about 30.7% at the start to 0.65% at the end. This is a rate derived from the particular trailing periods, not Regeneron's reported first-half 2019 rate, which was 15.1%. Pretax income grew by about 25.1% across the same reconstructed periods, far less than diluted EPS. The earlier tax rate, a 2018 intercompany-asset tax benefit and other period-specific items complicate any claim that underlying operations alone drove the earnings increase. Regeneron also recorded a $400 million 2019 research-and-development charge associated with its Alnylam collaboration; that payment was separate from a $400 million equity purchase. These items remain in the reported GAAP figures rather than being stripped out to improve the narrative. [5]

There is a larger warning about selection. Move both observation dates forward by one year, to October 2018 and October 2020, and Regeneron's simple return changes from -40.7% to +41.0%. That does not invalidate the original observation, but it does make a sweeping conclusion from that one window untenable. The company is in this article partly because it demonstrates how a persuasive-looking historical example can change when the calendar changes. [6]

Adobe: the counterexample

Adobe had a different outcome. From October 2017 to October 2019, its trailing revenue rose from about $6.90 billion to $10.64 billion and reconstructed trailing diluted EPS increased from $3.18 to $5.62. The share price went from $147.94 to $266.66, an 80.2% gain. Its P/E moved slightly higher, from 46.5x to 47.4x. The multiple did not need to contract merely because it began at a level that might look demanding in isolation. [7]

Adobe's free-cash-flow margin remained high, although it eased from 37.5% to 35.6%. Reported revenue growth is not the same as organic growth: acquisitions, including Marketo, affect the comparison. Taxes also shaped reported earnings. A reconstructed trailing effective tax rate fell from 18.4% to 5.7%, and Adobe discussed effects of U.S. tax reform and its corporate tax structure. It would be no more sound to credit Adobe's entire return to operational excellence than to blame Paycom's entire loss on its opening P/E. [8]

Adobe also illustrates why the information cutoff matters. Its fiscal third-quarter results had been published before both October 2 observation dates. The trailing EPS calculations therefore use those third-quarter filings, not older second-quarter figures. Using statements that were already stale on the selected dates would distort both the earnings comparison and the implied P/E. [7]

Stock-based compensation complicates the cash-flow picture. Paycom's trailing cash-flow SBC addback was 11.1% of revenue at entry and 7.4% at exit; Adobe's was 6.1% and 7.1%. SBC is noncash in the period and is already added back within operating cash flow. It should not be added to free cash flow again, yet continuing share grants can dilute owners. Free cash flow alone does not settle that trade-off. [1,7]

What the cases can and cannot tell us

The figures use financial reports available by each historical observation date. Trailing GAAP diluted EPS is reconstructed as the latest filed full-year per-share amount, less the comparable earlier year-to-date amount, plus the latest year-to-date amount. The inputs are issuer-reported EPS rounded to cents, so the resulting growth rates and P/E ratios are approximate at the displayed precision. The prices are FMP historical daily closes, split-adjusted but not dividend-adjusted; cash distributions with ex-dates inside each holding window are added separately, without reinvestment. Split histories were checked independently, although the original FMP split-response records were not preserved. [1,4,7,9]

This is independent financial research, not a replay of historical Qualtix ratings and not a validated backtest. The three examples were selected retrospectively and cannot establish how frequently growth overcomes valuation compression, or the return a new investor should expect. GAAP earnings include tax, dilution and nonrecurring effects; free cash flow is not a substitute for evaluating acquisition spending or return on invested capital. The exact endpoint dates also matter, as Regeneron's reversal shows. No conclusion here concerns whether any of these shares is attractive today.

The practical lesson is narrower than "avoid high P/E stocks." Ask what earnings are made of, what expectations a starting price appears to require and what must remain true for those expectations to hold. Then check how much of the eventual return came from delivered earnings, changing valuation and cash paid to shareholders. Three histories cannot supply a universal rule, but they can make the question more precise.

Sources and calculation notes

• 1. Paycom FY2020 10-K, Q3 2021 10-Q, FY2022 10-K and Q3 2023 10-Q. Trailing EPS, revenue, margins and cash-flow SBC addbacks are reconstructed from filed annual and year-to-date statements available at each observation date.
• 2. Paycom August 2023 outlook and October 31, 2023 release. Compare the two full-year revenue guidance ranges; prices follow the method in note 9.
• 3. Paycom dividend history. Three 2023 payments of $0.375 per share fall inside the selected window.
• 4. Regeneron FY2016 10-K, Q2 2017 10-Q, FY2018 10-K and Q2 2019 10-Q. Trailing figures use the same full-year-minus-earlier-YTD-plus-latest-YTD method.
• 5. Regeneron FY2018 10-K and Q2 2019 10-Q in note 4. TTM tax rates are reconstructed, not issuer-reported half-year rates; the filings discuss the intercompany asset transaction and Alnylam collaboration.
• 6. Regeneron one-year sensitivity: apply the price and distribution convention in note 9 to October 2, 2018-October 2, 2020, shifting both original endpoints by one year. This is one selected-window check, not a robustness test across all dates.
• 7. Adobe FY2016 10-K, Q3 2017 10-Q, FY2018 10-K and Q3 2019 10-Q. The third-quarter filings precede both October 2 dates; trailing figures and cash-flow SBC addbacks use filed annual and nine-month statements.
• 8. Adobe FY2018 10-K and Q3 2019 10-Q in note 7. Reconstructed TTM tax rates differ from issuer-reported quarter or nine-month rates.
• 9. Daily closing prices: FMP historical EOD full endpoint and adjustment convention. Simple return equals (exit split-adjusted close plus ex-date cash distributions)/entry split-adjusted close minus one. PAYC and REGN split histories show no splits in these windows; SPY issuer distributions supply its cash payments. The original FMP API request receipts and split responses were not archived.