Automatic Data Processing earns recurring SaaS-like fees on 1+ million employer relationships and harvests interest on ~$37B of payroll-tax float. At $214 versus a $239 base IV, the margin of safety is thin — own it on weakness, not at fair value.

Plain English

ADP runs payroll for one million companies. Every other Friday, employers send ADP money; ADP pays the workers, withholds taxes, files paperwork with all 50 states and the IRS, and earns interest on the tax money during the day or two before it gets remitted. Once a company plugs ADP into its bank, time clocks, and 401(k), unplugging is a nightmare — so customers stay for decades. The business should still exist in 2036, mostly bigger. The risk is faster payments shrinking the interest income and Workday or Rippling stealing customers.

Thesis

ADP is the largest payroll, HR, and employer-services processor in the world, serving roughly one million employers across small business (RUN), mid-market (Workforce Now), and enterprise (GlobalView, Vantage, Lifion/NextGen HCM), plus a Professional Employer Organization (PEO) arm (TotalSource) that co-employs ~750k worksite employees. The economic engine has two legs. First, recurring per-employee-per-month subscription fees on mission-critical software that sits at the intersection of money-movement and regulatory compliance — clients literally cannot pay their people without it. Second, payroll-tax float: ADP holds employer payroll-tax withholdings for an average of ~1-2 days before remitting to taxing authorities, generating roughly $37B in 'funds held for clients' (per the most recent 10-Q, March 31, 2026) earning a market yield. As short rates rose, this float quietly added ~$1B+ of pre-tax interest revenue.

The scorecard reads composite 63 with a heavy valuation weight (18/25) and capital-allocation strength (19). Headline ROIC is reported as 0% by the scorer because the float distorts invested-capital math (client funds are debt-funded assets); operating ROIC ex-float is materially north of 25%. Owner earnings TTM are $3.85B; reverse-DCF implies just 5.78% growth at $214.21 — below ADP's 6-9% historical organic-revenue cadence. IV range is $135.6 / $239.4 / $359.4. P/IV at 0.89 means the stock is trading at a 11% discount to base case but well above the bear floor.

At $214 you are paying for the durable franchise without a margin-of-safety cushion. I want $190 or below for meaningful new buys; above $300 the bull case is mostly priced in.

Moat

ADP has one of the more durable moats in software services. I count three of the five Buffett-Munger moat types as clearly present, plus a cost-advantage that is real but secondary.

Switching costs (WIDE). Payroll is the canonical example of what Damodaran [1][2] calls software-driven switching costs: 'Microsoft recognized earlier than most firms that the most significant barrier to entry in the software business is the cost to the end-user of switching from one product to a competitor.' Payroll is worse than office software. Mis-runs trigger employee revolts, IRS penalties, and W-2 reissuance. Implementations take 2-9 months, require year-to-date balance migration, parallel runs, and re-mapping of GL codes, benefits feeds, 401(k) carriers, garnishments, state unemployment IDs, and time-clock integrations. Mid-market clients have ADP wired into ~30+ third-party endpoints. Annual revenue retention runs ~92% (Employer Services) and ~95%+ in PEO. CFOs do not change payroll providers without a fired CHRO and a six-month project plan.

Intangibles / regulatory know-how (NARROW-to-WIDE). Running compliant payroll across all 50 US states, every Canadian province, and 140+ countries is a regulatory minefield: tax tables, garnishment rules, sick-leave laws, paid family leave, ACA reporting, multi-state remote-worker withholding, OECD pillar-two. ADP files ~80 million tax filings per year and processes 1-in-6 US private-sector paychecks. The 'ADP National Employment Report' itself is the data exhaust of this scale — useful as a moat indicator. New entrants (Gusto, Rippling, Deel) can build the SaaS UI; replicating the compliance backbone takes a decade and live-fire IRS audits.

Cost advantages / scale economies (NARROW). Buffett's GEICO framing applies [6]: 'the economies of scale we enjoy should allow us to maintain or even widen the protective moat surrounding our economic castle.' ADP's per-paycheck unit cost is industry-low because fixed compliance R&D, tax-filing infrastructure, and a single banking pipe spread across ~40 million worksite employees. Margins (Employer Services EBIT margin >34%) are well above Paychex peers in mid-market.

Float economics (treat as cost-advantage, not separate moat). ~$37B funds held for clients earns market yield; this is not free leverage but it is a structural deposit-like asset that scales with payroll wage growth and headcount.

Network effects (NONE). Payroll is not multi-sided in any meaningful sense; 'Marketplace' integrations are convenient but not network-defensive.

Pricing power (NARROW). ADP raises price per-employee-per-month ~2-4% annually with low churn impact, but pricing is not Visa-grade — Workday and Paycom can undercut on enterprise deals.

Competitor stress test ($10B + 5 years). A challenger with $10B and 5 years (Workday is exactly this) cannot dislodge ADP's small-business and PEO moat — those are too distributed and too compliance-heavy. They can and have stolen enterprise net-new (Workday HCM displaced ADP Vantage at many Fortune 500 accounts), and Rippling/Gusto are taking the bottom of SMB. ADP's moat is widest in the middle and in PEO, narrowest at the high-end.

Erosion risks. (1) AI-driven payroll bookkeeping commoditizing the SMB tier; (2) Rippling's unified IT-HR-finance bundle resonating mid-market; (3) instant-payroll real-time-payments compressing float to zero days.

Moat verdict: WIDE

Management

Management under CEO Maria Black (since Jan 2023, prior president of ES and PEO) and CFO Don McGuire has continued the disciplined Carlos Rodriguez playbook: organic growth, margin expansion, modest tuck-in M&A, and a ruthlessly consistent return of cash.

Reinvest. Capex is light — ADP is a software business with mostly compute and occasional data-center build. R&D runs ~$800M/year focused on Lifion/NextGen HCM (the Workday-class platform), AI-assisted payroll anomaly detection, and 'ADP Assist' generative-AI HR copilot. Product reinvestment is adequate but not heroic; the bear view is they under-invested vs. Workday for a decade.

Acquire. ADP's M&A history is conservative: WorkMarket (2018, gig workforce), Celergo (2018, global payroll), Global Cash Card (2017, paycards), Sora (2023, integration platform), Workforce Software (2024, ~$1.2B for advanced WFM/scheduling — the largest deal in 15+ years). Discipline is generally good; no transformative misadventures. Buffett's warning about acquisitions [5] — that animal spirits and accounting cosmetics drive most premium deals — does not appear to apply here.

Debt. Net-debt-to-EBITDA prints 5.4x in the scorecard, which is misleading: this number conflates 'short-term obligations' (the offsetting liability to the $37B funds held for clients) with corporate debt. Corporate net debt is roughly flat to modestly negative depending on float-day timing. Interest coverage of 0.0 reported is similarly an artifact. Real interest coverage on long-term notes is >20x.

Buybacks. ADP repurchases ~$1.5-2.0B/year. Share count down ~1.2% over 10 years (modest — most cash returned via dividends). Average buyback P/IV historically has been ~0.85-0.95 — adequate, not opportunistic. They buy steadily through cycles rather than leaning in at lows; this leaves alpha on the table but avoids the value-destruction trap of leveraged repurchase at peaks.

Dividends. ADP is a Dividend Aristocrat with 50 consecutive annual increases. Current yield ~2.4%, payout ratio ~55-60% of earnings. Growth ~10% CAGR. This is the primary return-of-capital lever and is signal of management's confidence in float-plus-subscription cash durability.

Communication. Investor days are dense, quantitative, segment-disclosed, with explicit medium-term targets (7-9% Employer Services revenue growth, 100bps annual margin expansion, mid-teens EPS growth). Management hits or beats these consistently. No accounting jujitsu; no stock-based-compensation games (SBC ~2-3% of revenue, low for software).

Concerns. (1) Black's tenure is short and the Workforce Software deal is her first real test of capital allocation at scale. (2) Buyback discipline could be sharper — they should be willing to lean in below $180. (3) Lifion has been a 'tomorrow' platform for nearly a decade; execution risk in catching Workday is real.

Insider alignment. Management ownership is modest (typical mega-cap dilution), but compensation is heavily tied to revenue growth, margin, and TSR. No related-party concerns. Board independence is strong.

Capital allocator: B+

Industry

Threat of new entrants (MODERATE). Greenfield SMB payroll has been entered successfully by Gusto, Square Payroll, QuickBooks Payroll, and Rippling over the past 15 years — proving entry is possible. But entry into mid-market and enterprise is far harder: clients require multi-state tax compliance, SOC1/SOC2, multi-EIN, multi-country, and integrations with hundreds of HRIS endpoints. PEO is gated by state insurance licensing. Effective barriers are high in ADP's profit pools, low at the SMB fringe.

Bargaining power of suppliers (LOW). ADP's main inputs are software engineers, data centers (now cloud-shifted to AWS), and banking partners. Talent is competitive but not constraining. AWS is a partial cost-of-goods variable, but cloud is a substitute, not a chokepoint.

Bargaining power of buyers (LOW-to-MODERATE). Individually, employers have very low bargaining power because of switching costs. Aggregated, large enterprise buyers and PEO clients negotiate hard, especially at renewal. SMB clients are essentially price-takers. The mid-market is where competitive RFPs put real pressure.

Threat of substitutes (MODERATE). Bring-payroll-in-house using QuickBooks or NetSuite is a substitute for the smallest businesses but rarely above 50 employees. Workday and Paycom are direct substitutes — not new entrants but adjacent oligopolists. Ceridian/Dayforce is a credible substitute in mid-market. AI-native payroll (rumored from OpenAI partnerships, Brex, etc.) is a future substitute risk.

Industry rivalry (MODERATE-to-HIGH). This is where the bear case starts. Workday has been winning enterprise HCM for a decade. Paycom has chewed up mid-market with its single-database architecture and aggressive sales force. Rippling is bundling IT, HR, and finance in a way ADP cannot match without a re-platforming effort. Gusto owns the under-50-employee segment that ADP RUN once dominated. Pricing is rational (no destructive price war), but share-of-wallet and net-new logo competition is fierce.

Value pool location and trajectory. The total HCM/payroll software market is ~$50-70B and growing high single digits with secular tailwinds: globalization of work, gig/contingent labor, increased compliance burden, real-time payments, and HR analytics. The value pool is migrating from raw payroll processing (mature, ~3-5% growth) to (1) HCM suite software with embedded analytics, (2) PEO/employer-of-record services, and (3) money-movement adjacencies (paycards, earned-wage access, instant deposit). ADP is well-positioned in (1) and (2), under-positioned in (3) versus payroll-fintech entrants.

Float economics regime change risk. Falling short rates compress the float yield (each 100bps move ~$300-400M revenue). This is a cyclical not structural threat. Real-time payments mandating same-day tax remit could compress float days from ~1.5 to <0.5 — a structural threat watch-listed for 2027-2030.

The industry is structurally attractive at the top end, brutally competitive at the bottom. ADP's moats fit best in mid-market and PEO, where the value pool is also growing fastest.

Industry Verdict: Good

Inversion

The bear case — short-seller pitch.

1. The single event that kills this. A federal mandate for same-day or instant payroll-tax remittance — driven by FedNow expansion and bipartisan tax-collection-modernization efforts — collapses the float-day window from ~1.5 days to under 6 hours. Combined with a falling-rate cycle (Fed back to ~2.5% by 2027), float revenue compresses from ~$1.0-1.2B/year to ~$200-300M. That alone is ~$700-900M of pre-tax earnings vaporized — roughly 12-15% of operating income — at the same time enterprise HCM share losses to Workday accelerate. The market re-rates ADP from a 22x P/E compounder to a 14x P/E utility-like processor.

2. Why the moat is narrower than bulls think. The 'switching cost' moat is a function of legacy data complexity, not product superiority. Every new cohort of small businesses born after 2020 onboards directly to Gusto, Rippling, or QuickBooks. ADP's net-new logo share in sub-50-employee is collapsing — they are now harvesting an aging cohort. In mid-market, Paycom and Dayforce are winning >50% of competitive RFPs ADP enters. In enterprise, Workday has effectively taken the segment. ADP's 'wide moat' is actually a wide back-book moat: it protects revenue from existing customers but does nothing for new-logo growth. Once you separate retention from acquisition, ADP looks more like an aging telco than a compounder. Damodaran's warning [1] applies: 'spending more money on R&D is clearly one way [to keep the moat], but companies that will see the greatest increases in value are not necessarily the companies that spend the most on R&D, but those who have the most productive R&D.' ADP has spent billions on Lifion/NextGen for a decade with limited share gain.

3. Why management is worse than it appears. The B+ grade conceals that ADP has under-invested in the platform shift to cloud-native, AI-first HCM. Workday raised $1B+ in 2005-2010 and out-built ADP on a ten-year horizon while ADP was reporting 12-15% EPS growth via buybacks and margin expansion — a beautiful current-period income statement masking a deteriorating long-term competitive position. The Workforce Software acquisition (~$1.2B in 2024) is a defensive, not offensive, move — admitting they could not build advanced WFM in-house. Buybacks at ~$200-250 are not value-creative when IV-base is $239 and the bear-case IV is $135. Average P/IV on buybacks is closer to 1.0x than the 0.7x a true value-allocator would target. Insider ownership is low; management is a salaried bureaucracy, not an owner-operator team.

4. What bulls are extrapolating that won't hold. (a) Float yield at current short-rate levels — they assume a permanent 4-5% portfolio yield when the 20-year average is ~2.5%. (b) Per-employee-per-month subscription growth above wage inflation — competitive pressure from Rippling and Gusto is forcing pricing flattening, not expansion. (c) PEO mid-teens growth — TotalSource growth has decelerated to high single digits as healthcare-cost inflation and SUTA arbitrage benefits compress. (d) Reverse-DCF implied growth of 5.78% — sounds easy, but real organic ex-float and ex-acquisition growth has been closer to 5%, and float revenue is an interest-rate bet, not a growth driver. (e) The buyback bid — at 22x earnings and net-debt-to-EBITDA optically 5.4x (even if cosmetically), the buyback could be cut in a downturn.

5. Valuation trap (multiple compression / regime change). ADP trades at 21.9x TTM earnings versus a 10-year average of 33.7x. Bulls call this 'cheap to history.' Bears note that the 33.7x average reflects the 2018-2024 ZIRP-era 'compounder bid' where every quality SaaS-adjacent name re-rated. Normalizing to a pre-2017 average of ~20-22x with 5% growth instead of 8%, fair multiple is closer to 17-18x. Apply that to normalized owner earnings of $3.4B (haircutting current float yield by 30%) → $58B equity value → ~$143/share. That is within $8 of the bear IV ($135.6).

Bear-case math. Float compression -$800M, growth deceleration to 4%, multiple compression to 17x → owner earnings $3.0B × 17 = $51B / 405M shares = ~$126.

If I am right, the stock could be worth $130 within 3 years.

Lollapalooza Bias Check

Several biases are active in me as I analyze ADP and I want to name them.

Authority and social proof. ADP is a Dividend Aristocrat with 50 consecutive years of increases, an S&P 500 stalwart, and a name every value investor 'knows is good.' That label is doing analytical work it shouldn't. The fact that Buffett-style investors have owned it for decades is not evidence of its current attractiveness — it is the prior generation's thesis. I have to discount the comfort I feel typing 'wide moat' because everyone else types it too.

Anchoring on the 10-year average P/E. The scorecard shows 21.9x TTM versus 33.7x 10-year average. My instinct reads 'cheap to history.' But history includes 2020-2021 ZIRP bubble multiples that should not be normalized. The bear case explicitly attacks this anchor and I find it persuasive — yet I notice myself wanting to dismiss it because the anchor feels factual.

Confirmation bias from the float story. I love the 'bank-without-being-a-bank' frame — it is intellectually elegant and Buffett-coded (insurance float analog). I am over-weighting evidence that float is durable and under-weighting real-time-payments regime-change risk. I caught myself writing 'falling rates' as 'cyclical' and 'real-time payments' as 'structural' without doing the math on which materially impairs IV more.

Recency bias on Workday/Paycom share gains. The last two years have been brutal for ADP enterprise net-new wins. I may be over-extrapolating recent share-loss trends; ADP could re-platform on AI-assisted payroll and stabilize. Or recent share losses could prove the leading edge of an unstoppable trend. I genuinely do not know which.

Commitment-and-consistency. Once I wrote 'Moat verdict: WIDE' in the moat section, my brain wanted every subsequent section to confirm WIDE. The inversion section forced me to question this; without that mandatory step I would have written a one-sided bull report.

Deprival super-reaction (low). Not active — I have no position in ADP, no career or relational stake. This bias is dormant.

Incentive bias. Operating on a research brief, not a trade ticket. Compensation is for accurate analysis, not for a favorable rec. Low risk.

Optimism-from-quality. The single most insidious bias here is letting the obvious-quality of the franchise translate into 'therefore the price must be okay.' At $214 vs. $239 base IV, there is a 0.89 P/IV ratio — meaningful but not the 0.7 I want for a Strong Buy. Quality does not equal margin of safety; price does.

10-Year Outlook

Same fundamental business model in 10 years? Yes. Employers will still pay employees, will still need to remit taxes, will still need HR/benefits/time. The form factor evolves (more AI-assisted, more real-time, more global/contingent) but the core utility — turning employee work into compliant, taxed, banked compensation — is permanent. Probability >95%.

Customer base larger? Likely yes, but mix-shifted. US private-sector employer count grows ~1-2%/year. Globalization and the gig/contingent workforce expand the addressable base further. ADP's PEO arm should grow faster than the host market for another decade. The risk is not aggregate demand but ADP's share of it — Rippling/Gusto cohort effects mean ADP could be growing more slowly than the market.

Profit per customer higher? Modestly. Per-employee-per-month pricing has historically grown at ~3% (slightly above wage inflation). Add-on attach (retirement, benefits, WFM, analytics) raises ARPU. Offset: AI-driven price compression in commoditized payroll execution. Net: probably +2-3%/year real ARPU growth, but with declining confidence interval.

Moat wider? Probably the same width or slightly narrower. Switching-cost moat is durable; cost-advantage moat depends on continued scale-leadership in compliance, which is being challenged by AI-assisted compliance startups. The float economic moat may narrow if real-time-payments mandate same-day tax remit (low-but-rising probability over 10 years).

Single biggest threat. A trifecta: (1) real-time payments mandate compresses float days, (2) Rippling/AI-native challengers force per-employee pricing flat, (3) Workday continues taking enterprise. Any one is manageable; the combination is a thesis-killer. None is imminent — all are five-to-ten-year regime questions.

Capital return durability. 50-year dividend record is real signal. Buyback program scales with FCF. Combined yield (dividend + buyback) of ~5-6% is reliably defensible across cycles.

Net. ADP in 2036 is highly likely to be a larger, more global, slightly less dominant employer-services platform throwing off $5-7B of owner earnings. Compounding at 6-8% is plausible; 10%+ requires Lifion to win or float economics to durably exceed history.

CONFIDENCE: medium

Position Guidance

• Recommendation: Hold
• Conviction: medium
• Target buy price: $185 (price-to-IV-base of ~0.77, restoring meaningful margin of safety; ~14% below current)
• Target trim price: $310 (above bull-case IV $359 less ~14% MOS-on-the-way-out; ~45% above current)
• Position sizing: Up to 4% of portfolio at target buy; 2-3% at current; do not chase above $260
• Notes: Quality is real (wide switching-costs moat, B+ capital allocator); price is fair, not cheap. Reverse-DCF growth of 5.78% is roughly fair to history. Float-rate sensitivity and Workday/Rippling competition are the two regime-change risks to watch. Add on >10% drawdowns; do not overweight at par to base IV.