Insperity, Inc. (NYSE: NSP) — A Founder Buying the Wreck of His Own Underwriting Cycle

Independent fundamental research. This article contains no buy/sell recommendation and no price target (the sole, labeled exception is the author’s Take, below).

Report date: 2026-07-11 · Price reference: $45.51 (2026-07-10) · Market cap: ~$1.7B · Enterprise value: ~$1.9B (corporate, ex-client-float) · Shares out: ~37.7M

⚡ The Author’s Take

The author’s own subjective opinion and general information, not investment advice. The analysis below carries no position and no price target.

HOLD — the most ownable of the distressed pure-play PEOs, but only at the founder’s price, not this one. Accumulate-on-weakness in the high-$20s to low-$30s (where the CEO was buying); don’t chase the +133% bounce to $45. Conviction: medium. Tag: “The founder bought the wreck — below here, so should you.”

Insperity and TriNet are the same accident from two different cars. Both are pure-play PEOs whose earnings are really a health-insurance underwriting spread, and both got run over by the 2024–2025 medical-cost trend (large claims, GLP-1s) that outran their annually-repriced client billings. But Insperity is the more violent wreck: structurally thinner margins (peak operating margin ~4.5% vs. TriNet’s 10.8%), a two-stage collapse — 2024 self-inflicted by a big, expensed Workday/HRScale investment, then 2025 by the benefit-cost miss — that pushed it to an outright GAAP loss and an 83%-peak-to-trough drawdown to $19.54. Where TriNet levered its balance sheet to buy back stock at the top and vaporized its equity, Insperity is the higher-quality allocator by a wide margin: one-share-one-vote, essentially zero M&A, organically built, and — the single most compelling fact on the name — a 37-year founder-CEO (Paul Sarvadi) who put ~$15.4 million of his own money into the stock, code-P open-market buys, into the crash (~$23 in March, adding up to $34 in June). That is not a man who thinks his franchise is impaired.

So why only HOLD, and why not here? Three reasons. First, the “recovery” is thinner than it looks: Q1-2026’s gross profit per worksite employee of $332 is down 2% year-on-year and Q1 is always the seasonal high — full-year 2026 likely lands in the ~$250s, still far below the $285 peak. Second, the ~5.6% dividend is not covered by earnings ($90M paid vs. $39M adjusted net income and a GAAP loss); management defended it by cutting the buyback to nothing, drawing $369M on the revolver, and loosening its leverage covenant — a real yellow flag, and a dividend cut is the most likely bear catalyst. Third, corrected for that drawn revolver, NSP trades at ~9.5x mid-2026 EV/EBITDA — in line with TriNet, not cheaper — so the deep-value optics (0.25x sales) are a gross-up artifact. The bull case is the founder’s conviction plus a genuine growth option (HRScale, the Workday-built mid-market platform TriNet has no analog for); the bear case is a stretched dividend on a seasonally-flattered, still-declining margin. At $45 those cancel to a HOLD. In the high-$20s–low-$30s — where Sarvadi was actually buying — the founder-conviction, organic-quality, and HRScale-optionality story is worth owning. Bull-flip trigger: full-year gross profit per WSEE inflecting back toward $270+ with the dividend held. Bear-flip trigger: a dividend cut, or gross profit per WSEE stuck in the $240s into 2027 (proving the trough is structural, not cyclical).

📈 Stock Price Action — Five-Year Event Map

Insperity is a deeper, longer round-trip than its PEO peers. The stock set its all-time-high close of $115.97 back in July 2019 and — unlike TriNet, which made new highs in 2024 — never exceeded it, re-peaking only near ~$107 in April 2023 before a two-and-a-half-year slide. It bottomed at a five-year low of $19.54 on 11 March 2026 — an ~83% drawdown from the 2019 high — then rebounded violently, roughly +133%, to $45.51 (10 Jul 2026). It remains ~61% below its all-time high, inside a 52-week range of $19.54–$57.05. The whole arc traces the health-benefit-cost cycle — amplified, in Insperity’s case, by a self-funded multi-year technology investment (HRScale/Workday) that hit the P&L at the worst possible moment.

Cycle narrative. (1) Insperity’s 2019 high was a peak it never revisited — a structurally thin-margin PEO whose share price re-rated with WSE growth but stalled well below the old high even in the 2021–2023 boom. (2) Through 2024 the benefit-cost trend began outrunning Insperity’s pricing while the company simultaneously ramped self-funded investment in HRScale, a double squeeze on already-thin margins. (3) The April-2025 drop confirmed the blowout on the Q1 2025 print. (4) Through late 2025 benefit costs peaked (+9% per covered employee), HRScale spend maxed out, and gross profit per worksite employee bottomed — pushing full-year 2025 to a GAAP operating loss. (5) The February-2026 capitulation to ~$19.54 came on the Q4-2025 report: a full-year loss, soft FY26 guidance, and open questions about whether a ~5%-yielding dividend was safe on negative earnings. (6) The rebound began with the Q1-2026 beat — benefit cost growth halved to +5%, a renegotiated UnitedHealthcare contract (pooling limit cut from $1M to $500K), the first HRScale clients going live — and, notably, ~$15M of open-market buying by the founder-CEO. The price move is FACT; the attributed cause is INTERPRETATION, cross-referenced to earnings dates, 8-K guidance, and benefit-cost disclosures.

1. Executive Summary

Insperity is a pure-play professional employer organization (PEO): via a co-employment model it becomes the administrative employer for small and medium-sized businesses, bundling payroll, health and benefits, workers’ compensation, and HR compliance so a 40-person firm can offer enterprise-grade benefits. It serves ~303,000 average paid worksite employees (WSEEs), reports ~$6.8B of “revenue,” and is run by its co-founder, Paul Sarvadi, CEO since 1989.

The economic reality is far smaller than the headline. Of ~$45.6B in gross billings that flow through Insperity, ~$38.8B is worksite-employee payroll (excluded from revenue) and ~$5.9B is direct pass-through cost (payroll taxes, benefits, workers’ comp) — leaving a gross profit of just ~$900M, about 2% of gross billings. That thin spread is the whole business, and it is dominated by one variable: the gap between what Insperity charges for health benefits and what those benefits cost. Insperity self-insures health (through a UnitedHealthcare arrangement), so it bears the medical-cost trend directly, with pricing that resets only annually while claims accrue continuously.

That model just failed twice in a row. Gross profit fell 14% in 2025 to $900M — even as reported revenue rose 4% — and operating income swung from a $250M peak (2022) to a −$10M loss (2025), with a GAAP net loss of $7M and EPS of −$0.19. The collapse was two-stage and worth separating: 2024 was self-inflicted — gross profit actually grew, but operating income fell 47% because Insperity ramped a large, mostly-expensed investment in HRScale, its new Workday-built mid-market platform, and staffed up for it; 2025 was the underwriting miss — benefit cost per covered employee rose +9.2% against +3% pricing, taking gross profit per WSEE from $285 to $242. Volume compounded the pain: average WSEEs are flat-to-shrinking (−1% in Q1 2026), so all revenue “growth” is price.

The stock fell 83% peak-to-trough to $19.54 (February 2026), then rebounded ~133% to $45.51 as Q1 2026 showed the benefit-cost trend halving (+5.2%) and management’s “margin recovery in year one” plan tracking. Three things now define the debate. The bull anchor is unusually strong: founder-CEO Sarvadi bought ~$15.4M of stock, open-market, into the crash — an emphatic conviction signal, backed by a genuinely higher-quality capital-allocation record than peers (one-share-one-vote, zero M&A, no debt-funded top-of-cycle buyback), plus HRScale as a real growth option. The bear anchors are equally concrete: the Q1 gross-profit-per-WSEE “recovery” is seasonally flattered and still down year-on-year; the ~5.6% dividend is not covered by earnings and is being defended with a drawn revolver and a loosened leverage covenant; and corrected for that debt, the stock is not actually cheap versus TriNet.

What the market is pricing. At ~9.5x mid-2026 EV/EBITDA and ~21x depressed adjusted EPS, the stock discounts a partial recovery — gross profit per WSEE grinding back toward 2024 levels and WSEEs stabilizing — but not a full return to peak economics or a growth re-acceleration. That is roughly fair, which is why the memo body reaches no verdict. The judgment that matters: whether the founder’s conviction and the HRScale option outweigh a stretched dividend and a still-unproven margin recovery. At the founder’s entry prices (high-$20s–low-$30s), the risk/reward is asymmetric; at $45, it is balanced.

2. Business Overview

What Insperity does. Insperity is one of the oldest and largest pure-play PEOs in the United States, founded in 1986 (as Administaff) and led by co-founder Paul Sarvadi ever since. Through co-employment, Insperity becomes the administrative employer-of-record for its SMB clients: it runs payroll and payroll-tax administration, sponsors and administers the health and welfare benefit plans, provides workers’-compensation coverage, and delivers HR compliance, employment-law guidance, and an HR technology platform. The client directs the actual work; Insperity handles the employer machinery and absorbs much of the administrative and regulatory burden. Its flagship bundled offering is branded Workforce Optimization / HR360; it also sells a lighter HRCore package and, newly, the mid-market HRScale solution.

The revenue gross-up — read gross profit, not “revenue.” Insperity’s income statement is a classic PEO gross-up, and it is even more misleading than TriNet’s. Walking down the funnel for FY2025:

Insperity keeps roughly 2% of the ~$45.6B that flows through it. The reported “$6.8B revenue” overstates the economic scale by an order of magnitude, and — critically — it rose 4% in 2025 while the real top line (gross profit) fell 14%. Any analysis anchored on the revenue line will miss the entire story. This memo treats gross profit, and its per-unit form gross profit per worksite employee per month, as the true revenue and margin gauges.

Volume — the KPI. PEO economics scale with worksite employees (WSEEs). Average paid WSEEs grew explosively post-COVID (250,745 in 2021 → 312,102 in 2023, +17.7% then +5.8%) but have since stalled: 307,261 (2024, −2%), 310,089 (2025, +1%), and 303,049 in Q1 2026 (−1% year-on-year). Client retention runs ~83% (a ~9–12% annual attrition rate); new-client-sales contribution has decelerated to ~+1%. Because Insperity’s operating leverage depends on spreading a large fixed corporate cost base over more WSEEs, a flat-to-shrinking base is a structural drag — and it is being spread over the same base into which management poured HRScale investment.

Customer mix. Insperity targets “best-in-class” SMBs across professional services, technology, financial services, healthcare, and other white-collar verticals — the same broadly pro-cyclical, white-collar-hiring-sensitive cohort as TriNet, though Insperity is somewhat more diversified across mainstream SMBs and less concentrated in venture-backed tech. Verdict: a mission-critical, recurring HR-outsourcing service whose true economic scale (~$900M gross profit) is a thin ~2% spread on a huge billings flow — a spread dominated by health-benefit underwriting and currently shrinking, with volume flat-to-down.

3. Industry Dynamics

The PEO industry structure is covered in depth in the author’s companion TriNet analysis and applies equally here, so this section emphasizes what is specific to Insperity’s position.

A structurally supported end-market wrapped around a cyclical underwriting pool. The demand side is durable: millions of US SMBs face rising regulatory complexity (multi-state payroll, pay-transparency, leave, privacy, AI-employment law), and outsourcing that burden to a PEO is a genuine value proposition — ADP and Paychex both grow their PEO worksite-employee counts high-single-digit even now. But the profit pool is a health-insurance underwriting spread, and that is cyclical. Insperity self-insures (through a UnitedHealthcare ASO-style arrangement, booking IBNR reserves quarterly), so when the medical-cost trend spikes — as it did in 2024–2025 on outpatient inflation, specialty pharmacy, GLP-1 utilization, and large-claim frequency — the spread collapses. This is the Marathon capital cycle in miniature: benign post-COVID years supported aggressive volume growth and thin pricing; the whole industry underpriced into a rising trend; the correction (repricing, client shedding, benefit-plan redesign) is now underway.

Insperity’s structural handicap: thinner margins than TriNet, and far thinner than the diversified scale players. The single most important industry fact for NSP is that its economics are structurally thinner than its closest peer’s. Insperity’s peak operating margin (~4.5% in 2022) is less than half TriNet’s peak (~10.8%), because Insperity retains a smaller spread per worksite employee and carries a heavier fixed corporate cost base (a high-touch service model with a large field organization of “Business Performance Advisors”). The same medical-cost shock that halved TriNet’s operating margin pushed Insperity’s into negative territory — there was simply less cushion. Against the scaled, diversified incumbents the gap is starker still: ADP TotalSource (~740k WSEEs) and Paychex run their PEO arms as ~24% of ~40%-operating-margin, float-rich, diversified machines that barely registered the 2024–25 trend. Insperity, like TriNet, has the underwriting risk of a health insurer without the scale, diversification, or capital buffer of the largest ones.

Competitive intensity and switching costs. The space is crowded — ADP, Paychex, TriNet, Justworks, Rippling, Gusto, and Deel all compete for SMB HR/benefits — and switching costs, while real (payroll/benefits migration friction, co-employment lock-in), are lower than in core enterprise payroll. Insperity’s ~83% retention and elevated attrition through the repricing confirm that clients facing a benefit-cost-driven price increase will shop. Regulation (state PEO licensing, IRS CPEO certification, ERISA/ACA) is a modest barrier to entry and a source of stickiness but does nothing to dampen the underwriting cyclicality. Verdict: a structurally supported industry with a cyclical, underwriting-driven profit pool — and Insperity is a sub-scale, thin-margin participant in it, more exposed to the cycle than any of its larger peers.

4. Competitive Position

Name the moat — and its narrowness. Insperity’s advantage is a combination of modest customer switching costs (co-employment lock-in, the friction of migrating payroll/benefits/compliance mid-relationship) and a high-touch service reputation built over ~40 years — its Business Performance Advisor model and long tenure give it a brand and service depth at the premium end of the SMB market. In Greenwald’s taxonomy this is a weak-form customer-captivity advantage with a service/brand overlay; it is not a scale-cost moat (ADP and Paychex have far more scale), not a network effect, and not a technology moat (which is precisely the gap HRScale is meant to fill).

The share-stability test — failed, like the whole industry. A durable moat should hold share and price. Insperity’s retention held around 83% but attrition ran at the high end of its 9–12% range through the repricing, and volume went flat-to-negative — clients were willing to shop when benefit costs pushed prices up. That said, one nuance cuts for Insperity: management deliberately shed lower-profitability clients as part of the margin-recovery plan, so some of the volume loss is a chosen mix improvement rather than pure competitive defeat. The distinction matters — TriNet’s attrition looks more like indiscriminate franchise shrinkage; Insperity’s is partly a deliberate cull — but the net result is still a shrinking worksite-employee base.

The “success penalty” and HRScale — the one genuinely differentiated bet. Insperity’s most interesting competitive move is HRScale, a mid-market (150–5,000-employee) solution combining its HR services with Workday’s client-facing HCM technology. It targets a real structural leak: Insperity’s historical “success penalty,” whereby clients it helped grow eventually outgrew its technology and defected to enterprise HCM platforms. HRScale is designed to retain those maturing clients and win larger new logos where a services-plus-Workday bundle is a “hand-in-glove fit.” If it works, it is a durable differentiator that neither TriNet nor the software-natives can easily replicate (the Workday partnership is exclusive-ish and hard to source). This is the single feature that could widen Insperity’s moat — and it is entirely unproven (first clients went live Q1 2026; no revenue disclosed).

Verdict: a crowded market, a weak-to-moderate moat, and one credible widening option. Insperity is a sub-scale, thin-margin PEO whose service/brand advantage is real but narrow, whose share proved contestable when it raised price, and whose best hope for durable differentiation (HRScale) is a promising but unproven mid-market bet. Better-positioned than a commodity payroll processor; materially weaker than the diversified scale incumbents.

5. Growth History and Forward Opportunities

History — a post-COVID surge that stalled. Insperity compounded WSEEs impressively out of the pandemic (+17.7% in 2022, +5.8% in 2023) as SMB hiring boomed, driving gross profit to ~$1.05B and operating income to a $250M peak (2022). But growth stalled hard thereafter: WSEEs shrank 2% in 2024, grew just 1% in 2025, and are down 1% in Q1 2026. The quality of the 2021–2023 growth was also flattered by a benign claims environment (low benefit costs) and rising float income — both of which reversed. Underneath, all recent “growth” is pricing (+3% revenue per WSEE), not units.

The forward story — a two-year plan, growth deferred to 2027. Management’s framework, laid out by Sarvadi on the Q1 2026 call, is explicit: Year 1 (2026) is margin recovery; Year 2 (2027) is regaining growth. The margin-recovery levers are the priority; growth is deliberately deferred. The growth levers themselves are: (1) HRScale — the Workday-built mid-market catalyst, with ~6,000 WSEEs of signed commitments to onboard within six months and a pipeline “building,” targeting the 150–5,000-employee segment where average deal size is far larger (so fewer salespeople are needed per unit of growth); (2) sales-force rebuild — modest additions to Business Performance Advisors/Consultants, ramping into 2027; (3) AI-enabled service — HR360 AI agents to lift productivity and retention. Management frames HRScale as addressing the “success penalty” and expects it to “play a significant role in regaining growth momentum as we move into 2027.”

How to weigh it. The growth case is real but back-end-loaded and unproven. Even management guides to stabilization (WSEEs −1% to −2.3%), not growth, in 2026, with HRScale revenue not yet visible in the financials and the macro backdrop deteriorating (Insperity’s own client survey showed 54% of SMBs expecting a negative economic impact, up from 42% in January). The AI question cuts both ways, as at TriNet — management argues AI is spurring new-business formation (~500k applications/month, +12% year-on-year) as fast as it displaces headcount, but a sustained white-collar contraction would hit the WSEE base. Verdict: currently negative (shrinking) growth, with a credible but unproven path to re-acceleration in 2027 riding almost entirely on HRScale. The burden of proof sits with HRScale client-count and revenue disclosures over the next few quarters — which do not yet exist.

6. Financial Quality

The two-stage collapse. Insperity’s earnings decline is best understood as two separate events, because they have different implications:

Stage one (2024): self-inflicted. Gross profit actually grew to $1,052M, but operating income fell 47% (to $117M) because operating expenses jumped ~$117M — the new HRScale/Workday investment (~$29M of P&L expense in 2024, from zero) plus a ~$60M (+13%) increase in salaries and higher SBC as Insperity staffed up for the mid-market push. This was a chosen investment, not a market shock. Stage two (2025): the underwriting miss. Benefit cost per covered employee rose +9.2% against +3% pricing; gross profit per WSEE fell 15% to $242; gross profit dropped $152M — and with opex still elevated, operating income went negative. A ~$40M adverse swing in prior-period claims run-off (from a +$29M benefit in 2024 to an −$11M cost in 2025) added to the damage.

Margins improve with the cycle, not with scale. Because the profit is a ~2% underwriting spread on a huge billings flow, small moves in the benefit-cost-versus-pricing gap swing the P&L violently — the operating leverage that lifted margins on the way up worked brutally in reverse. The Q1 2026 print showed the rate of decline slowing (benefit cost +5.2% vs. +9.2%), aided by a renegotiated UnitedHealthcare contract that cut the large-claim pooling limit from $1M to $500K per claimant for 2026 (ceding tail risk at a higher premium — a de-risking, not a tailwind) plus a favorable client-mix shift from shedding unprofitable accounts. But — a critical skeptical flag — Q1 gross profit per WSEE of $332 is down 2% year-on-year, and Q1 is always the seasonal peak (the payroll-tax component of gross profit front-loads into H1). Full-year 2025 was $242 precisely because the back half decays; a $332 Q1 2026 is consistent with a full-year figure in the ~$250s, not a return to the $285 peak. The recovery is real but partial and seasonally flattered.

Returns and the balance sheet. With book equity a thin ~$46M stub (from cumulative buybacks against $850M of treasury stock — but with positive retained earnings of $638M, unlike TriNet’s accumulated deficit) and a 2025 operating loss, ROE and ROIC are not meaningful at the trough. The important correction to the naive read: Insperity is not sitting on net cash. The ~$660M of cash and securities on the balance sheet is mostly client float; the company’s own adjusted cash fell to just $57M at year-end 2025, and it had drawn $369M on its revolver. On a corporate basis Insperity is modestly net debt, not net cash, and its liquidity cushion has visibly thinned through the down-cycle.

Cash flow — a QoE trap in both directions. Reported operating cash flow swung from +$520M (2024) to −$278M (2025), and reported FCF from +$482M to −$309M. Neither figure is real — both are dominated by client-fund timing (payroll-deductions-payable swung +$335M in 2024 and −$357M in 2025). Normalizing out the float, underlying cash generation was modestly positive in 2025 (roughly the −$7M net loss plus ~$45M D&A and ~$61M SBC, less ~$31M capex) — but still less than the $90M dividend plus buyback plus HRScale spend, which is why cash drew down and the revolver was tapped. Verdict: economics improve with the medical-cost cycle, not with scale; the reported top line and reported cash flow both mislead; and the balance-sheet cushion is thinner and more drawn than the headline cash suggests. This is a genuinely lower-margin, higher-operating-leverage version of the PEO model at a cyclical (and partly self-inflicted) trough.

7. Capital Allocation

Insperity’s capital-allocation record is materially higher-quality than TriNet’s — and its single most important recent act is a genuine bullish signal — but it carries real, specific blemishes.

The founder is buying — the headline. Into the 2026 collapse, founder-CEO Paul Sarvadi bought ~534,987 shares of open-market stock (all code “P”) for ~$15.4M: ~202,000 shares at ~$23 in March, ~100,000 at ~$28–30 in May, and 233,000 at $34.05 in June — buying not just at the bottom but adding on the way up. No other insider bought; there was no large discretionary selling. For a 37-year founder-CEO to deploy ~$15M of personal capital into his own stock at the depths of an 80%+ de-rating is about as strong a conviction signal as exists in the Form 4 record — and it stands in stark contrast to TriNet, where the CEO bought nothing after a comparable collapse. (Balancing caveat: Sarvadi also pledges shares as loan collateral — ~250,000, ~17% of his holdings, with board authorization to go to 35% — so his personal posture is not purely one-directional; but pledging is not selling, and the code-P buys are unambiguous.)

Organic, disciplined, one-share-one-vote. Insperity is a one-share-one-vote company (no dual-class wedge — Sarvadi’s control rests on his 3.86% economic stake and board structure, not super-voting shares), it has done essentially zero M&A in five years (100% organic growth), and — decisively — it did not lever its balance sheet to buy back stock at the top of the cycle, the exact error TriNet made. Its buyback cadence was mildly procyclical (heaviest in higher-price years: $203M in 2019, $131M in 2023) and it throttled repurchases to just $19M in 2025 as cash flow deteriorated — imperfect timing, but it cut the buyback rather than borrow to fund it. One quality caveat: with SBC of ~$50–61M/year, much of the gross buyback merely mops up dilution — net share count fell only ~8% over seven years despite ~$650M+ of repurchases.

The dividend — the yellow flag. Insperity pays a $0.60/quarter regular dividend (~$2.40/year, ~$90M, ~5.6% yield) that it has raised ~20%/year since 2018 and clearly treats as sacred. In 2025 that dividend was not covered: $90M paid against $39M of adjusted net income and a GAAP loss. Rather than cut it, management cut the buyback to near zero, drew the revolver, and — the clearest tell — amended its credit facility in December 2025 to raise the maximum leverage covenant from 3.00x to 3.75x and to exclude dividends from the interest-coverage covenant. Defending an uncovered payout by loosening covenants is a real capital-discipline concern; a dividend cut is a live risk if the margin recovery stalls, and would likely be the sharpest bear catalyst.

HRScale — the big mid-cycle bet. The Workday/HRScale investment (announced February 2024) is the largest capital-allocation call in Insperity’s recent history — mostly expensed (the primary reason adjusted EBITDA roughly halved to $131M), funded through a cyclical trough, and not yet revenue-generating. The bull read: a founder betting the company’s growth re-acceleration on a genuinely differentiated product. The bear read: a large, cash-absorbing project layered on top of a down-cycle and an uncovered dividend.

Incentives and governance. Compensation is reasonably aligned: the 2025 annual bonus keyed off adjusted EBITDAIC, WSEE growth, and Workday execution — and the formula bit (only the Workday goal paid; the EBITDA and growth goals missed, zeroing most of the bonus and the 2025 LTIP EBITDA tranche). A special June-2025 retention PSU for Sarvadi and the COO — 5-year, pure stock-price hurdles starting at $105 (~+70% from grant) — is well-structured but its magnitude and mid-crisis timing drew a ~3-point drop in say-on-pay support (to 94%). Governance blemishes: a staggered/classified board (an entrenchment device, partly explaining the absence of activist pressure), the founder share pledging the board keeps expanding, and ongoing equity-plan dilution. Verdict: an above-average, founder-aligned, organic allocator — the anti-TriNet on balance-sheet discipline — but with a stretched, covenant-defended dividend, a staggered board, and a large unproven mid-cycle bet. The founder’s ~$15M of buying is the single most bankable fact in the whole memo.

8. Changes and Headwinds — Last Two Years

The damage (2024–2026): (1) the two-stage margin collapse — 2024 self-inflicted by HRScale/staffing opex, 2025 by the +9.2% benefit-cost blowout — to a GAAP operating loss; (2) volume stall — WSEEs flat-to-negative, −1% in Q1 2026; (3) gross profit per WSEE from $285 to $242; (4) a dividend rendered uncovered, defended by a drawn revolver and loosened covenants; (5) an 83% peak-to-trough stock collapse to $19.54.

The self-help response (2025–2026): (1) a 3-year plan (margin recovery in 2026, growth in 2027); (2) the renegotiated UnitedHealthcare contract (pooling limit $1M→$500K) capping large-claim tail risk; (3) benefit-plan redesign, strategic pricing, and deliberate shedding of unprofitable clients (favorable mix); (4) a ~$9M Q1-2026 restructuring (~4% of non-sales headcount) to fit the smaller base; (5) the HRScale launch (first clients live Q1 2026); and (6) the founder’s ~$15M of open-market buying.

Net. The changes are a coherent, credibly-executed turnaround under a founder with real skin in the game — but the recovery is early, seasonally-flattered, and 2027-weighted, and the dividend overhang is unresolved. Verdict: the changes strengthen the thesis directionally, but the two biggest questions — whether gross profit per WSEE recovers durably and whether the dividend survives — remain open, and the market has already re-rated the stock +133% off the low in anticipation.

9. Risk Analysis (Risk Matrix)

Risk of a total loss is low: Insperity generates modest underlying cash, carries manageable (if increased) leverage, reinsures large claims and workers’-comp tail risk, and is backed by a founder buying stock. The dominant risks are a dividend cut and a stalled margin recovery — either would validate the bear view that this is a structurally thin-margin, cyclically-impaired PEO whose payout was unsustainable.

10. Valuation Discussion (Embedded Expectations)

Where it trades. At $45.51: ~$1.7B market cap; on a corporate basis (adding the ~$369M drawn revolver, netting the ~$57M own adjusted cash) EV ≈ $1.9B. There is no P/E (GAAP loss). P/S is 0.25x (AZI 7th percentile) — but that is a gross-up artifact on a 2%-margin billings flow and is not a meaningful cheapness signal. The honest forward multiples: EV/adjusted-EBITDA ≈ 9.5x on the FY2026 midpoint ($200M) and ~21x depressed adjusted EPS (FY2026 midpoint ~$2.10). Corrected for the drawn revolver, NSP is not cheaper than TriNet (~9.4x EV/EBITDA) — the deep-value optics dissolve once the balance sheet is read properly.

Embedded-expectations read. At ~9.5x mid-cycle-guidance EBITDA on a trough margin, the market prices a partial recovery: gross profit per WSEE grinding back toward 2024 levels and WSEEs stabilizing, with HRScale contributing modestly — but not a full return to 2022 peak economics ($286 GP/WSEE, $250M operating income) or a growth re-acceleration. That is a defensible, roughly-fair expectation.

Scenario analysis (normalized earnings power):

What the market is underwriting correctly vs. incorrectly. The base case looks fairly priced. The stock is cheap only if you believe the bull (margin fully recovers and HRScale re-accelerates growth) and expensive if you believe the bear (the trough is structural and the dividend is cut). The wide dispersion (~$22 to ~$85) reflects genuine uncertainty about a thin-margin underwriter mid-repair. Note the tell embedded in the founder’s behavior: Sarvadi bought aggressively at $23–34 and stopped adding as the stock reached the mid-$30s — his revealed-preference “value zone” sits well below today’s $45. Cross-checks: EV/EBITDA ~9.5x is far below ADP/PAYX (~18–20x) — an earned discount for a sub-scale, thin-margin, cyclical pure-play; it is roughly in line with TriNet, the direct comp. The single number that adjudicates the valuation is gross profit per WSEE per month recovering toward the $270s–$280s. No price target and no recommendation.

11. Variant Perception

Consensus view. The sell-side treats NSP as a cyclical PEO past the worst of its benefit-cost blowout, with margin recovery tracking but volume still declining, a stretched dividend, and a full recovery unproven — a “show-me” turnaround where the +133% bounce has already discounted the near-term good news.

The strongest bull case. Insperity is a founder-conviction, organic-quality PEO at a cyclical (and partly self-inflicted) trough, with the healthiest capital-allocation DNA in the group and a genuine growth option nobody else has. The benefit-cost trend has halved (+5.2% vs. +9.2%), the UnitedHealthcare renegotiation caps the large-claim tail that did the damage, deliberate client-mix improvement is lifting profitability per account, HRScale is live and addressing the “success penalty” with a differentiated Workday bundle, and the founder — 37 years in the chair — just bought ~$15M of stock. As gross profit per WSEE recovers toward the $275+ range and HRScale converts in 2027, normalized earnings power reverts toward $4–5 and the stock re-rates. You are buying a durable niche franchise from a conviction-buying owner at a fraction of its former value.

The strongest bear case. Insperity is a structurally thin-margin PEO (~2% spread) whose “recovery” is seasonal optics on a still-declining unit margin, propping up a dividend it cannot afford. Q1 gross profit per WSEE was down year-on-year; full-year 2026 will land in the $250s, not the $285 peak. The dividend ($90M) exceeds sustainable earnings ($39M adjusted NI), is being funded by a drawn revolver and a loosened covenant, and a cut is coming if margins don’t fully mend — a de-rating catalyst. The HRScale bet is large, expensed, and unproven; the WSEE base is shrinking; the balance-sheet cushion is thinning; and corrected for debt the stock is no cheaper than higher-margin TriNet. The founder’s buying is admirable but he bought at $23–34, not $45 — the easy money has been made.

The 3–5 assumptions that matter most, and what would falsify each:

1. Does full-year gross profit per WSEE recover toward $270+? Bull needs it; bear wins if it stalls in the $240s–$250s. Falsifier: the FY2026 figure (Q1 seasonal-adjusted).
2. Is the dividend safe? Bull: recovery covers it by 2027; bear: a cut is the catalyst. Falsifier: a dividend cut, or two more quarters of coverage < 1x.
3. Does HRScale produce real revenue? Bull: a 2027 growth engine; bear: an expensed bet that never scales. Falsifier: HRScale client-count/revenue disclosures over 2026–2027.
4. Was the WSEE loss a deliberate profitable cull or franchise impairment? Bull: chosen mix improvement; bear: share loss. Falsifier: WSEE trajectory + profitability-per-account trend.
5. Will the founder’s conviction be vindicated? Bull: $15M at the lows marks the bottom; bear: even an owner can be early/wrong. Falsifier: the margin and dividend outcomes above.

The factor-positioning read. NSP screens as an extreme deep-value, deeply-negative-momentum small-cap (Value +0.55, Momentum −0.56, SmallSize +0.49; beta 0.64) — even more stretched on the value/momentum axes than TriNet — with a poor risk-adjusted record (y3 −23.7%/yr, max drawdown −66%) and a violent +133% recent bounce. This is a textbook falling-knife-that-bounced. The variant-perception opportunity, if any, is that consensus is anchored on the falling-knife framing and a founder’s aggressive open-market buying — a signal a pure quant screen cannot see — may mark a genuine cyclical bottom if the margin data cooperates. Absent that confirmation, the factor read says “cheap, hated, and thinly-covered” — which is what both a value trap and a bottoming turnaround look like from the outside.

12. Fact vs. Interpretation Table

13. Open Questions

1. Where does full-year 2026 gross profit per WSEE actually settle once the Q1 seasonal peak washes out — the $250s (bear) or back toward $270+ (bull)?
2. Is the dividend safe through 2026–2027, or will management finally cut it if coverage stays below 1x?
3. What is HRScale’s actual revenue and client trajectory? No revenue is yet disclosed; the entire growth case depends on it.
4. How much of the WSEE decline is deliberate profitable-client culling vs. genuine competitive loss?
5. How much cumulative capital has the Workday/HRScale program consumed, and when does it turn cash-accretive?
6. Will the $500K UnitedHealthcare pooling election actually reduce earnings volatility, or simply re-tax the spread via higher premiums?
7. Is the founder’s ~$15M of buying a bottom signal or an early call — and does his share pledging complicate the read?

14. What Must Be True

For the bull case (the turnaround compounds):

• Gross profit per WSEE recovers durably toward $270–285 as the UnitedHealthcare pooling change, plan redesign, and favorable client mix take hold across the full year (not just the seasonal Q1).
• WSEEs stabilize and return to growth in 2027, with HRScale producing real, disclosed revenue and addressing the success penalty.
• The dividend is held and becomes covered again by 2027 as margins recover.
• Falsification test: Full-year 2026 gross profit per WSEE inflects toward $270+ with the dividend maintained, and HRScale reports a growing client count. If, by mid-2027, GP/WSEE is still stuck in the $240s–$250s or the dividend has been cut, the bull case is broken.

For the bear case (value trap / structural impairment):

• The Q1 margin “recovery” proves seasonal, and full-year gross profit per WSEE stalls in the $240s–$250s.
• The dividend is cut as coverage stays below 1x, triggering a de-rating and validating the “unsustainable payout” thesis.
• HRScale never scales into a meaningful, profitable revenue line; the WSEE base keeps shrinking.
• Falsification test: Full-year gross profit per WSEE recovers toward $270+ AND the dividend is maintained through 2027. If both occur, the “structurally impaired / uncovered dividend” thesis is wrong and the franchise has genuinely reset.

As with TriNet, both cases share the same falsification test, pointed in opposite directions — the gross-profit-per-WSEE trajectory and the dividend decision over the next several quarters will adjudicate the debate. The distinguishing feature here is that the person with the most information — the founder who has run the company for 37 years — has already voted with ~$15M of his own capital.

15. Source Appendix

See the separate Appendix B — Source Appendix (stitched into the combined report). Primary sources: Insperity FY2021–FY2025 10-Ks and Q1 2026 10-Q (SEC EDGAR, CIK 0001000753); 2024–2026 DEF 14A proxies; recent 8-Ks (FY25 results, the December-2025 credit-facility amendment, the February-2026 realignment, the 2026 annual-meeting vote); the Form 4 insider corpus; the Q1 2026 earnings-call transcript (ROIC.ai); financial data cross-checked via ROIC.ai; price and factor data via AZI Trading and FactorsToday; and public TriNet, ADP, and Paychex filings for PEO-model framing and peer comparison.

APPENDIX A — Standard Diligence Questionnaire

Insperity, Inc. (NYSE: NSP) · Report date 2026-07-11 · Price reference $45.51

Supplemental to the analysis. Answers grounded in the primary sources; Fact/Interpretation labels applied where it matters.

General

What thoughtful questions have other investors asked about this company? Whether the ~5.6% dividend is safe on negative earnings (the central debate); whether the Q1-2026 gross-profit-per-WSEE improvement is a real recovery or seasonal optics; whether the benefit-cost/GLP-1 trend is structurally repriceable; whether HRScale (the Workday mid-market bet) will ever produce revenue; how much of the WSEE decline is deliberate profitable-client culling vs. competitive loss; and what to make of founder-CEO Sarvadi’s ~$15.4M of open-market buying into the crash. The most sophisticated framing treats NSP as a thin-margin (~2%) insurance-underwriting spread business mid-repair, with a founder-conviction overlay.

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? A cyclical low — an outright GAAP operating loss in 2025 (−$10M operating, −$0.19 EPS) vs. a $250M operating peak in 2022. The 2022–2023 peak was flattered by benign post-COVID claims, WSEE surge, and float income — all reversed.

Driven by external environment or internal actions? Both, in two stages. 2024’s operating-income halving was self-inflicted (HRScale/Workday investment + staffing). 2025’s swing to a loss was the external benefit-cost blowout (+9.2% per covered employee vs +3% pricing). The swing variable — the benefit-cost-vs-pricing gap — mixes an exogenous medical-cost trend with Insperity’s own pricing/underwriting discipline.

How stable are revenues? Reported revenue (~$6.8B) is deceptively stable — it rose 4% in 2025. The real economic top line (gross profit) is volatile: it fell 14% to $900M. NSP keeps only ~2% of ~$45.6B gross billings, so small cost/price moves swing earnings violently.

Outlook for products/services; how big is the market? Large, under-penetrated US PEO/HR-outsourcing market, structurally supported by rising regulatory complexity — but NSP is currently shrinking within it (a share/execution issue). Growing, domestic, mid-single-digit market; HRScale targets the larger mid-market (150–5,000 employees).

Business Quality & Competitive Moat

Is the industry getting more or less competitive? More — ADP, Paychex, TriNet, and software-natives (Rippling, Gusto, Justworks, Deel) all compete for SMB HR/benefits.

How profitable is the business (ROIC, ROE)? At a trough: operating margin −0.15%, a net loss. ROE/ROIC not meaningful at the trough (thin ~$46M book equity). Peak operating margin was only ~4.5% (2022) — structurally thinner than TriNet (~10.8%).

How profitable is the industry — barriers to entry? Moderate barriers (state PEO licensing/CPEO, compliance content, benefit-risk pooling). Switching costs real but modest (proven by attrition). The scaled incumbents (ADP/PAYX) are structurally advantaged; NSP is sub-scale and thin-margin.

Can the business be easily understood? Yes, once you strip to gross profit and gross profit per WSEE — an HR-administration and benefits-underwriting spread on a large payroll flow. The benefit-cost ratio is the one number that matters.

Can it be undermined by foreign low-cost labor? No — domestic, regulated, US-employment-law service. AI is the relevant disruption vector; management argues (plausibly) that human judgment/liability transfer at high-stakes moments resists software-only substitution. HRScale is the technology-competitiveness answer.

Do brands matter? Nature of competition? Switching costs? Brand/service reputation matters at the premium SMB end (Insperity’s ~40-year, high-touch Business Performance Advisor model). Competition is on benefit-plan competitiveness, price, and service. Switching costs are real but moderate.

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? The ~40-year brand, service organization, and client relationships are largely unrecognized (NSP has minimal goodwill/intangibles — ~$13M — because it is organically built, near-zero M&A). The HRScale software is partly capitalized (~$34M unamortized).

Off-balance-sheet liabilities? Health-plan underwriting exposure (self-insured via UnitedHealthcare, IBNR reserves; large claims pooled above $500K from 2026) and workers’-comp claim exposure (captive-insurance sub; WC reserves). Stop-loss/pooling reinsured. No unusual off-balance-sheet financing.

How conservative is the accounting? Reasonable, but note: benefit-cost IBNR estimates swing results (prior-period run-off swung ~$40M adverse in 2025); WC reserve run-off has flattered results ($29–32M/yr benefit — reversal risk); and HRScale dev costs began being capitalized in 2025, flattering the P&L expense line.

How CapEx-hungry? Light — capex ~$31–40M/yr (mostly capitalized software). Capital-light operations; the real “capital” is the implicit insurance risk capital and the field service organization (expensed).

Capital Allocation & Management

How much FCF, and how is it used? Underlying cash generation (ex client-float timing) was modestly positive in 2025 (~+$68M pre-working-capital) but less than dividend + buyback + HRScale spend — hence a revolver draw. Reported FCF (−$309M) is a client-float timing artifact, not real burn. Priorities: fund HRScale, defend the dividend, opportunistic buyback (first cut).

Significant acquisitions recently? None — 100% organic, ~zero M&A in five years (a quality positive).

Buying back shares? Modestly and mildly procyclically ($203M/2019, $131M/2023, throttled to $19M/2025). Did NOT lever up to buy at the top (contrast TriNet). Heavy SBC leakage (~$50–61M/yr) means net share count fell only ~8% over seven years.

Issuing large amounts of new shares to insiders? SBC ~$61M/yr (moderate); a +1.62M-share plan increase approved 2026. Net share count still falling.

Compensation policy / motivations of management? Bonus on adjusted EBITDAIC + WSEE growth + Workday execution (the formula bit — most of 2025 bonus zeroed); LTIP 75% adjusted EBITDA + 25% relative TSR (2025 EBITDA tranche paid 0%); a special 5-year, $105-hurdle retention PSU for the founder/COO. One-share-one-vote. Say-on-pay ~94–97%. Founder-CEO of 37 years, 3.86% owner, buying stock — strongly aligned, with the pledging caveat.

Valuation & Market Data

Is the stock an ADR, MLP, or K-1 issuer? No — ordinary US common stock, NYSE-listed, one-share-one-vote, 1099 dividend.

Dividend policy? $0.60/qtr regular (~$2.40/yr, ~5.6% yield), raised ~20%/yr since 2018, treated as sacred; one special ($2.00) in Q4 2021. Currently uncovered by earnings — the key risk.

How profitable is the business? At a trough — operating loss in 2025; ~2% gross-profit margin on gross billings; peak operating margin only ~4.5%. Entirely benefit-cost-ratio dependent.

Is net income diverging from cash from operations? Reported OCF is dominated by client-fund timing (swung −$357M in 2025), so it diverges wildly from net income in both directions — a PEO-specific QoE trap. Normalize out the float to judge true cash generation.

Risks & Downside

What factors would cause the stock to decline? A dividend cut; a second benefit-cost miss; gross profit per WSEE stuck in the $240s; HRScale failing to scale; a white-collar/SMB employment recession; a WC reserve reversal.

Risk of a catastrophic loss? Low-to-moderate — a sustained underwriting spiral on a thinning liquidity cushion is the tail; mitigated by the $500K pooling cap, WC reinsurance, and a conviction-buying founder.

Chance of a total loss? Low — generates modest underlying cash, manageable (if increased) leverage, positive retained earnings, and a defended (if stretched) dividend.

Recent News & Events

Has the business environment changed recently? Yes — the benefit-cost trend halved (+5.2% Q1’26 vs +9.2% 2025), a renegotiated UnitedHealthcare contract (pooling $1M→$500K) took effect for 2026, HRScale launched (first clients live Q1 2026), and SMB sentiment deteriorated (54% expect a negative economic impact).

Significant acquisitions? None — growth is organic; the strategic move is the Workday/HRScale partnership (not an acquisition).

Change in accounting policies? Began capitalizing a portion of HRScale development costs in 2025 (ASC 350-40). Amended the credit facility (Dec 2025) to loosen the leverage covenant and exclude dividends from the interest-coverage test.

Recent changes — new markets, facilities, management? HRScale launch (new mid-market segment); a ~$9M Q1-2026 restructuring (~4% of non-sales headcount); AI-agent rollout in HR360; founder-CEO buying ~$15.4M of stock. Management is stable (Sarvadi since 1989; CFO Jim Allison).

APPENDIX B — Source Appendix

Insperity, Inc. (NYSE: NSP) · Report date 2026-07-11 · Price reference $45.51 (2026-07-10)

All facts trace to the primary sources below. Primary sources over secondary; recent over stale.

SEC filings (primary — EDGAR, CIK 0001000753)

Earnings call / management

Quantitative data providers

Peer / industry context (public)

Analytical frameworks

• Competition Demystified (Greenwald & Kahn) — moat taxonomy (weak customer captivity + service/brand overlay), share-stability test (failed under repricing), the HRScale “moat-widening” bet.
• Capital Returns (Marathon) — the PEO health-underwriting capital cycle: benign years → thin pricing/volume growth → medical-cost spike → repricing/client-shedding correction.

Notes on data reliability

• Revenue is a gross-up. The ~$6.8B “revenue” is a thin (~2%) markup on ~$45.6B gross billings; the memo uses gross profit and gross profit per WSEE per month as the real top-line and margin gauges.
• “Net cash” is a mirage. ROIC’s balance-sheet netting implied net cash; the filing shows only ~$57M of own adjusted cash against a $369M drawn revolver — NSP is modestly net-debt on a corporate basis. AZI P/B ($1.78 book, 43rd pctile) is garbled — ignored; equity is a thin stub anyway.
• Reported FCF is a client-float artifact (−$309M in 2025 reflects a −$357M payroll-payables timing swing, not a burn).
• ROIC.ai and AZI are third-party aggregators; every material figure was reconciled to the 10-K/10-Q. Ratios driving a verdict (gross profit per WSEE, benefit-cost trend, dividend coverage) are reproducible from the filing line items.