Pay and Capital Allocation

Herbalife has repurchased about $6.5 billion of its own stock since 2007. The last $2.0 billion of it, spent between 2020 and 2022, bought 42.5 million shares now worth $518 million, and roughly two-thirds of that money went to a single departing holder. The incentive plan that governs what management does with the next surplus pays on two measures — local-currency net sales and adjusted EBITDA — and on neither cash flow nor debt.

The record of the last surplus

The buyback programme is the reason the balance sheet looks the way it does. The FY2025 Form 10-K states plainly that the company's debt "has not resulted from the need to fund our normal operations, but instead has resulted primarily from our share repurchase programs," and puts cumulative repurchases since 2007 at approximately $6.5 billion [1].

The most recent tranche is the one a buyer today still owns. In August 2020 Herbalife completed a modified Dutch auction tender offer, repurchasing about 15.4 million shares at $48.75 for $750.0 million; open-market purchases took the 2020 total to 18.4 million shares for $892.1 million at an average of $48.53 [2]. In January 2021 the company bought approximately 12.5 million shares directly from Carl C. Icahn and his affiliates for approximately $600.0 million, or $48.05 a share; with open-market buying, 2021 totalled 20.4 million shares for $982.7 million. A further 3.7 million shares followed in 2022 at an average of $35.73, for $131.8 million [3].

Sources: repurchase volumes and prices per the FY2021 [4] and FY2022 [5] Forms 10-K; value today derived at the 27 July 2026 close of $12.18.

Forty-two and a half million shares for $2,006.6 million is an average of $47.21. At $12.18 the same shares would cost $517.7 million. The $1,489 million difference is larger than Herbalife's entire market value today, and it is the arithmetic behind the shareholders' deficit and the debt stack described in Business and Balance Sheet and Debt and Covenants.

Two-thirds of that money went to one seller. The $600.0 million January 2021 block is named to Icahn in the filing itself. The August 2020 tender is not attributed in the 10-K, but Icahn's Section 16 record reports the sale of 14,722,025 shares at $48.75 on 12 August 2020 — the tender price, and 96% of the shares the company bought in that offer. Taken together, approximately $1,318 million of the $2,007 million, or 66%, retired the position of a single activist holder rather than shrinking the share count against the open market.

A commitment that lapsed

In February 2022, with the balance sheet already carrying the 2020-2021 spending, the company built repurchases into guidance for the first time: "Our 2022 guidance includes the assumption of $50 million in share repurchase per quarter, which reflects the minimum buyback amount we anticipate completing on a quarterly basis" [6]. Asked whether that was a commitment, the finance chief said the intent was to give investors "a sign of commitment to our share repurchase program beyond the words that we have historically provided" [7].

The full-year outcome was $131.8 million — less than two of the four promised quarters [8]. In 2023 the company repurchased nothing [9], and the $1.5 billion authorisation expired in February 2024 with approximately $985.5 million unused [10]. The pattern is consistent rather than random: the company bought heavily at $48, tapered at $36, and stopped entirely below $10. That record sits alongside the restored $400 million restricted-payments basket the April 2026 amendment created [11].

What the board buys now

Capital allocation since 2025 has moved from the share count to acquisitions. Four transactions have closed in fifteen months, all small at the base and all carrying earn-outs that are several times the base.

Sources: Pruvit and Pro2col asset purchases, milestone payments and the $5m-$25m and $46m contingent ceilings per the FY2025 Form 10-K, Note 2 [12]; Link BioSciences per the same note [13]; Bioniq per the Q1 2026 Form 10-Q [14].

The Pruvit and Pro2col assets were bought together for $19 million on 17 April 2025, of which $14.4 million was allocated to Pro2col's software. Pruvit carries a milestone of $5 million to $25 million payable in the second quarter of 2027; Pro2col carries subscriber-milestone payments capped at $46 million in aggregate and running to 2035, of which $5 million was already triggered in 2025 [15]. Link BioSciences cost $6.5 million for a 51% interest in the acquiring vehicle [16]. Bioniq, closed on 30 April 2026, is $55 million of base consideration payable over five years plus up to $95 million of sales-milestone payments [17].

Base cash of about $85.5 million is genuinely small against $252.9 million of 2025 free cash flow, which is management's own framing: the finance chief told the February 2026 call the programme is "not a huge use of cash" and that the company is "still looking to do small acquisitions but still get our total debt down to $1.4 billion by the end of 2028" [18]. The $166 million of contingent ceilings is the part that is not small: if the milestones are hit, the pay-away lands in 2027 and beyond, inside the window in which roughly $700 million of debt must be retired to reach the $1.4 billion target set out in Debt and Covenants.

One transaction supplies an outside mark. In February 2026 Cristiano Ronaldo's holding entity took an initial 5% of HBL Pro2col Software in exchange for services and sponsorship rights, then exercised an option for a further 5% at $7.5 million, reaching 10% fully diluted, with options over another 10% expiring in January 2028 [19]. Taken literally, $7.5 million for 5% values the platform at $150 million, against $14.4 million of software assets recognised nine months earlier. The caveat is real: this is a fixed-price option granted alongside a sponsorship, not an arm's-length clearing price, and the first tranche was paid in services. It is a directional signal, not a valuation.

The two numbers management is paid on

For 2025 the Compensation Committee set the annual incentive on two equally weighted metrics: targeted local-currency net sales and adjusted EBITDA [20]. The performance stock units reintroduced in 2025 run on the same pair, again split half and half, measured over three years to December 2027 [21].

So the short-term plan and the long-term plan measure the same two things. Nothing in either references cash flow, leverage, debt reduction, return on capital, or share-price performance relative to a peer group. That matters here more than it would elsewhere: the finance chief has stated his own ranking — "My number one priority is still to get our gross debt down to $1.4 billion by 2028, which would get our net debt below $1 billion" [22] — and that priority appears in no incentive metric the proxy discloses. The stock appreciation rights, which are struck at the grant-date price and pay only on appreciation, are the sole element that moves with the share price, and they are a form rather than a metric.

What that omission cost is on the record. The company's own leverage reconciliation carries both ends of it [23]. Between 2021 and 2025 Herbalife generated $1,143.8 million of free cash flow and cut Credit Agreement total debt from $2,845.8 million to $2,050.0 million, and over the same four years its equity fell from $4,433 million to $1,335 million, because adjusted EBITDA dropped 24.7% and the enterprise multiple went from 7.6 times to 4.6 times. Of the four terms in that sentence, the incentive plan pays on one — adjusted EBITDA. The forward version of the same trade, and what is different about it now, is worked through in Valuation Arithmetic.

The local-currency construction also decides how the growth this report has been examining is credited. Net sales are measured with exchange-rate movement removed, and adjusted EBITDA is "further adjusted to include for bonus purposes, the impact of changes in currency exchange rates" [24]. Volume and price count identically, and India's post-GST surge — the subject of India Concentration — counts in full.

The bar, and where it was set

The 2025 adjusted EBITDA target was $620.0 million against $645.8 million achieved on the same basis in 2024 [25]. Actual came in at $634.9 million, 102.4% of target, which paid 124.0% on that half. Local-currency net sales of $4,920.2 million missed the $4,967.8 million target and paid 95.0%. The blended payout was 109.5% [26].

The slope is steep. Payout runs from 50% at 85% of the EBITDA target to a cap at 110% [27], which implies roughly ten points of payout for each 1% of EBITDA above plan. A 2.4% beat against a target set 4.0% below the prior year's result produced a 24-point premium on the larger half of the bonus.

The Committee addresses the target directly, and its answer is the strongest fact against reading the bar as low: the target "was below 2024 Adjusted EBITDA," but "reflected anticipated foreign currency headwinds," and "on a constant currency basis, the target would have been approximately $690 million, representing approximately 9% year-over-year growth" [28]. That is a coherent defence, and it is not verifiable from the outside, because the currency assumptions embedded in the plan are not disclosed and neither are the multi-year PSU targets, which the proxy says simply "correspond to the Company's long-range forecast" [29].

What can be checked is that the incentive plan runs on a fourth definition of EBITDA, alongside the three the loan documents and the earnings deck already use.

Sources: reported adjusted EBITDA per the June 2026 corporate overview, Historical Financial Summary [30]; the incentive-plan measure is the Company-Selected Measure in the proxy's pay-versus-performance table, presented as adjusted for bonus purposes [31] [32].

The gap is not always in management's favour: the bonus measure ran $29 million to $31 million above reported adjusted EBITDA in 2021, 2022 and 2023, $11 million above in 2024, and $22.7 million below in 2025. The gap runs in both directions across the five years. What it establishes is that the payout number differs from the number the market prices, and only the proxy discloses it. The multiple-denominator problem set out in Debt and Covenants extends into the pay plan.

Pay versus shareholder return

The pay-versus-performance table required by Item 402(v) puts the two series side by side.

Source: 2026 Proxy Statement, pay-versus-performance table [33]; the $100 investment is measured from 31 December 2020 and the peer index is the sixteen-company TSR group named in footnote 6, Pay v. Performance — p.71") [34]. 2022 and 2025 combine the two individuals who held the chief-executive role in those years.

Compensation actually paid to the two people who held the chief-executive role in 2025 was $27.2 million — $14.0 million to Michael Johnson as Executive Chairman and former CEO, $13.1 million to Stephan Gratziani — more than double any earlier year in the table [35]. Compensation actually paid marks unvested awards to year-end, and the shares roughly doubled during 2025 from a February low. The measure did what it is designed to do. The cumulative column is the one that does not move: $100 invested at the end of 2020 was worth $26.83 at the end of 2025, against $71.50 in the peer group.

Total summary-table compensation for the six named officers was $22.9 million in 2025, about 9% of the year's free cash flow, and the disclosed CEO pay ratio is 151:1 against a median employee at $50,032 [36] [37]. The benchmark set is larger than the company: the sixteen-name compensation peer group had median revenue of $6.0 billion and median market capitalisation of $4.1 billion, placing Herbalife at the 45th percentile on revenue and the 32nd on market value [38].

Shareholders registered that. Say-on-pay carried with approximately 51.5% of votes cast at the 2025 annual general meeting [39]. The Committee's stated response was that "a key theme we heard from shareholders was a desire for an increased emphasis on performance-based long-term incentives," and it reintroduced PSUs in 2025 after pausing them in 2023 and 2024 [40].

The next agreement moves the other way. An employment agreement with Michael Johnson dated 18 February 2026 runs from 1 May 2026 to 30 April 2027 at a salary of $700,000, with a target bonus of 100% of salary and long-term incentives of $5,600,000 [41]. Those awards are 25% restricted stock units and 75% stock appreciation rights, with no performance stock units, and all of them vest on the date of the 2027 annual general meeting — roughly twelve months. The agreement also carries a $500,000 annual personal-use jet allowance [42]. The defence is that three-quarters of the value is in SARs struck at market, which pay nothing unless the shares rise; the counterpoint is that the 2025 award to the same executive was half PSUs, and a year after telling shareholders it had heard them on performance-based long-term incentives, the board wrote a package containing none, in an instrument that vests in a year.

Ownership, and who has been buying

The register is concentrated in funds rather than insiders. At the 9 March 2026 record date, Vanguard held 12.08%, Nantahala Capital Management 8.40%, The Baupost Group 8.33%, Route One Investment Company 7.28%, BlackRock 7.16% and Renaissance Technologies 5.43%; all eleven directors and six named officers together held 5.12% [43]. Icahn, whose affiliated entities beneficially owned 23.85% of the outstanding shares as recently as the 2020 record date, is gone [44].

The Section 16 record since January 2021 shows 48 open-market purchases by insiders totalling $5.8 million, against $19.6 million of open-market sales excluding the Icahn transactions. Nineteen of those 48 purchases, worth $0.9 million, were made by one director, Juan Miguel Mendoza, who is also a top independent distributor. Since January 2024 only two purchases were made by executive officers: Johnson bought $498,300 at $8.07 in February 2024, and the Chief Legal Officer bought $25,800. Two officers sold in May 2026 after the refinancing and the share-price recovery — the Chief Commercial Officer $1.9 million and the Chief Operating Officer $0.6 million.

The proxy's related-party section carries three items worth a professional investor's attention. Gratziani was paid approximately $753,994 in 2025 "in consideration for suspending his distributorship operations." The sister and brother-in-law of director Juan Miguel Mendoza earned approximately $1,263,466 under the Marketing Plan. And the company reimbursed Johnson $475,457 for chartered air transportation from BLADE Urban Air Mobility, where his son-in-law serves as a senior executive and aircraft broker; the Audit Committee reviewed and approved it [45]. Two of the eleven directors are classified as non-independent specifically because they earn distributor income [46], which is a structural feature of this model rather than an anomaly — but it means the board that sets the Marketing Plan payout includes people paid by it.

What would move the read

The favourable reading is that the destructive capital allocation was executed by a different management team at a different price, that the board has since chosen the cheapest available use of cash — retiring debt — and that acquisitions have been kept to a scale the balance sheet can absorb. The record supports the first two claims; the third depends on earn-outs that have not yet come due.

Three things would change it. The first is the 2026 say-on-pay result and whether the Committee's next design reinstates performance conditions on the Executive Chairman's award. The second is a repurchase authorisation: the restricted-payments basket now permits roughly $400 million [47], and reinstating buybacks before the $1.4 billion gross-debt target is met would say which priority is real. The third is the contingent-payment schedule — Pruvit's milestone falls due in the second quarter of 2027 and Bioniq's payments run to 2031, so the 2026 and 2027 cash flow statements will show whether the acquisition programme stayed at $85 million or became a $250 million one.