Part 15 of 15 in the series The Job That Sees It. See every installment →
The Million Dollar Question: ExxonMobil provides its chief executive with a car and a security driver, on the basis of what it calls a credible threat assessment. The company’s 2025 proxy statement itemises what that car cost shareholders in 2024 at $21,295. How much of that figure is the driver’s pay?
Read on for the answer.
Start with the number a reader came for, because this time there is one.
Search chauffeur salary and the federal government will answer. In the US occupational classification the job exists, under a name it shares: SOC 53-3053, Shuttle Drivers and Chauffeurs. The May 2025 Occupational Employment and Wage Statistics estimate, read from the BLS national data file itself, puts it at $17.93 an hour, $37,290 a year, across 248,530 people.
| SOC 53-3053, May 2025 | Hourly | Annual |
|---|---|---|
| 10th percentile | $13.51 | $28,090 |
| 25th percentile | $15.73 | $32,710 |
| Median | $17.93 | $37,290 |
| 75th percentile | $21.97 | $45,700 |
| 90th percentile | $26.47 | $55,070 |
| Mean | $19.17 | $39,880 |
That is the answer, and it is a reliable measurement of something. The question is of what. The ninetieth percentile of this occupation — $55,070 — is roughly what the median American worker earns, since all occupations sit at a $50,980 median. The top tenth of chauffeurs are paid a little better than the middle of everybody.
Two documents explain why, and they are both the survey’s own.
Chauffeur salary: what the federal number measures, and who it leaves out
The first is the occupation’s definition. ONET prints 53-3053 as: “Drive a motor vehicle to transport passengers on a planned or scheduled basis. May collect a fare. Includes nonemergency medical transporters and hearse drivers.*”
Read that again as a job description for the person a reader has in mind. On a planned or scheduled basis. May collect a fare. The code is built around the airport shuttle and the dialysis van. Its own sample job titles are Airport Shuttle Driver, Chauffeur, Driver, Limo Driver, Motor Coach Driver, Shuttle Bus Driver, Shuttle Driver, Special Needs Bus Driver, Transportation Driver, Van Driver. One of those ten is the job in question. The largest employing industry, per the Occupational Outlook Handbook, is Other Transit and Ground Passenger Transportation at 25% of employment, then nursing and residential care facilities at 8% and automobile dealers at 7%. The best-paying industry is Taxi and Limousine Service, at a $43,230 median.
The second document is the survey’s scope, and it is decisive. The Technical Notes for the May 2025 OEWS estimates state that the survey covers “wage and salary workers in nonfarm establishments,” that it “does not include the self-employed, owners and partners in unincorporated firms, household workers, or unpaid family workers,” and, separately, that “Private households (NAICS 814) are also excluded.”
A chauffeur on a family’s payroll is a household worker employed by a private household. On the face of those two exclusions he is outside the frame before anyone asks what he earns. The $37,290 is then the median pay of the workforce that is inside it — transit and ground passenger operators, nursing facilities, car dealerships, taxi and limousine companies — which is a real and useful number and is not the number the search was looking for.
Two more cautions from the same notes bite harder here than in most occupations. The annual figures are not measured: “the annual rate for an occupation is calculated by multiplying the hourly wage rate by a typical work year of 2,080 hours” — a convention, for a job whose defining feature is that its hours are not typical. And “May 2025” is a rolling window assembled from six semiannual panels reaching back to November 2022.
The government’s own employment counts also disagree, which is worth stating rather than smoothing: the OEWS survey file reports 248,530 people in 53-3053, while the Employment Projections table in the Handbook reports 251,500 for the same occupation and year. Different programmes, different scopes — the projections include workers the wage survey excludes — and each figure is used below only for what its own source measures.
How much does a chauffeur cost: the same arrangement, disclosed seven ways
The other half of the search traffic asks the cost question, which is not the salary question. Here the labour statistics have nothing to offer, because the employers who buy this service in its purest form are households, and households publish nothing.
With one exception. When a public company provides a driver to an executive, the arrangement becomes reportable compensation and the company has to put a dollar figure on it. Those figures are filed under the SEC’s perquisite-disclosure rules and computed by the company itself, and read side by side they are astonishing — not for what they reveal about drivers but for what they reveal about disclosure.
| Filer | Fiscal year | Disclosed amount | What the label covers |
|---|---|---|---|
| MGM Resorts | 2023 | $1,136 | Car and driver, personal use, at a marginal hourly rate |
| Visa | 2023 | $6,928 | Car and driver, non-business share only |
| ExxonMobil | 2024 | $21,295 | The car only — driver pay expressly excluded |
| Goldman Sachs | 2018 | $61,978 | A column headed “Car” — commuting and non-business share, driver compensation included |
| Walgreens Boots Alliance | 2024 | $106,738 | “Professional security driver/agent services” |
| ManTech | 2019 | $109,119 | Employee time, salary plus overhead, allocated |
| CACI International | 2025 | $208,342 | “Personal security driver,” third-party billings, personal-use share |
Two orders of magnitude, for arrangements that from the back seat look identical. The spread is not a fact about chauffeurs. It is a fact about accounting, and the filings say so themselves.
MGM’s footnote is the most candid: “The incremental cost for the car and driver was calculated using the 2023 IRS mileage rate, the personal use time and average chauffeur hourly rate. The variable cost rate is used rather than the standard business rate as the Company uses the car and driver for Company business … and would incur the fixed costs of operating the vehicle and employing the driver.” The company already employs chauffeurs; the chief executive’s personal use is billed at the margin. Hence $1,136.
Visa’s is the most instructive paradox. Its method explicitly “includes the cost of the driver’s salary and benefits for the proportion of time the driver was used for non-business trips” — a genuine salary allocation — and it produced $6,928. The filing does not say what the business/personal split was, and the car component is fuel rather than lease. A company car and driver can appear in a proxy statement as under seven thousand dollars.
CACI sits at the other end, and for a reason that has nothing to do with generosity: it is a defence and intelligence contractor, the arrangement is a security detail rather than a comfort, and the figure is the personal-use slice, allocated by mileage, of a third-party contract whose full value is not disclosed. Its method note says the costs are “actual billings for services and fees charged directly to the Company by third party firms providing security personnel” and that “No security personnel under this provision are employees of the Company.”
The one filer reporting something close to an economic cost is ManTech, which for years disclosed the fully loaded internal cost of employee time spent on non-corporate matters for its executive chairman, “primarily as a driver”: “the cost is determined by using the employee’s salary and overhead costs for the year to calculate an hourly cost and allocating that cost based on the percentage of time spent on these matters.” That method yielded $109,119 for 2019, $116,721 for 2018 and $138,117 for 2016 — and even that is only the non-corporate share, with some unquantified part of it other personal staff work.
One honest sentence for the cost question, then: a dedicated private driver is a full-time salaried position plus the price of covering the hours the principal is awake, and no filer in the public record discloses that total — only slices of it, computed several different ways.
The answer is A: the driver’s pay is not in the number
The answer is A — none of it.
ExxonMobil’s 2025 proxy statement lists for its chief executive, for 2024, “$44,703 for residential security,” “$31,906 for security costs related to personal travel,” and “$21,295 for the cost of the car provided for security reasons.” The security driver is named in the disclosure — the company says the services provided include “security services and personnel during travel, car and security driver” — and the method note explains what happens to his pay: “For security personnel employed by the Company, the cost is the actual incremental cost of expenses incurred by the security personnel. Total salary, wages, and benefits are not allocated because the Company already incurs these costs for business purposes.“
So the disclosed cost of that car and security driver contains the car, the maintenance and the fuel, and — on the filing’s own stated method — zero dollars of driver. Nobody is hiding anything: incremental-cost reporting is what the disclosure rules ask for, and Exxon explains its method in plain English in the same paragraph. But it means the number cannot be read as the cost of the arrangement.
A related convention appears at Cisco, which excludes its in-house security personnel’s compensation on the same reasoning — “Compensation is not included because Cisco already incurs that cost for business purposes” — but whose $284,581 figure for fiscal 2025 does include the driver, because that driver is contracted: “The incremental cost for the authorized car service and personal security driver is the actual cost.” Two filers, the same sentence about in-house compensation, and opposite treatments of the driver, depending on whether he is on the payroll or on an invoice.
Pfizer adds a tax layer. Its reading of the rules is that “the cost of the personal use of the car and fuel is imputed as income to the CEO,” while “Tax regulations provide the cost of the driver is not reportable as income to the CEO as a result of the recommendations contained in an independent, third-party security study.” On the company’s own account, then, the car is a taxable perk and the driver, backed by a threat assessment, is not income at all — a position that rests on conditions in the regulations this piece did not examine.
The American licences actually called chauffeur’s licences were not written for chauffeurs
The job has a name in American law, and the name has drifted away from the work.
Michigan still defines the term and still issues the credential. MCL 257.6 makes a “chauffeur” a person who operates a motor vehicle as a motor carrier or carrier of passengers, or “who is employed for the principal purpose of operating a motor vehicle with a GVWR of 10,000 pounds or more,” or who operates a bus or school bus. MCL 257.301(4) requires that anyone driving as a chauffeur hold a chauffeur’s licence.
Then read the exclusions in MCL 257.6(3). “Chauffeur” does not include a farmer, a firefighter driving an ambulance, a person driving a motor home for personal pleasure, a transportation network company driver, a limousine driver, or a taxicab driver — the last two added by 2016 PA 348, effective 21 March 2017. The test for the weight-based route, in subsection (2), turns on employment that “customarily involves transporting for gain or hire any merchandise for display, sale, or delivery” — merchandise, not passengers.
The Michigan chauffeur’s licence is a freight, bus and motor-carrier credential. The man in the dark suit holding the door does not need one, and if he drives a limousine the statute says he is not a chauffeur at all.
Iowa keeps the word closer to its ordinary meaning, and in doing so may well catch the household driver — which is the opposite problem. A chauffeur there is “a person who operates a motor vehicle, including a school bus, in the transportation of persons for wages, compensation, or hire,” with small taxicabs and transportation network drivers expressly excluded. Iowa makes it a licence class — Class D, “Valid for the operation of a motor vehicle as a chauffeur” — whose endorsement code covers a “Passenger vehicle less than 16-passenger design,” precisely the band below the federal commercial threshold.
Note that the test is not “for hire” alone. It is “for wages, compensation, or hire,” and a salaried household driver transports persons for wages — his own. The statute exempts driving that is “occasional and merely incidental” to the operator’s principal business, which a full-time driver’s is not, and it troubles to carve out a home care aide by name, which suggests the legislature expected domestic-type driving to fall inside the definition otherwise. Against that, the Iowa DOT describes Class D as covering “small passenger-for-hire vehicles.” Whether the licence reaches a family’s driver is not settled by the text, and this piece found no Iowa authority applying it to one.
So the two states that kept the word do not agree about what it covers. One has drafted the household chauffeur out of the definition; the other may have left him in by accident. Neither statute was written with him in mind.
The federal line sits above all of this. Under 49 CFR 383.5, a small vehicle — one below the 26,001-pound weight thresholds — becomes a commercial motor vehicle, requiring a CDL and a passenger endorsement, when it “is designed to transport 16 or more passengers, including the driver.” And Michigan’s own mirror of that definition, MCL 257.7a, contains the cleanest statement in this entire body of law: a commercial motor vehicle “does not include a vehicle used exclusively to transport personal possessions or family members for nonbusiness purposes.”
That sentence is the whole regulatory posture. The law counts the vehicle and the commerce. It does not count the passenger. Which is why the most consequential driving job in a wealthy household is, as a matter of licensure, an ordinary one.
The difference between a driver and a close-protection driver is not a credential
Ask what separates a driver from a security driver and the trade will answer in skills: route selection, advance work, reading a mirror, reversing at speed, the decision to push through rather than stop. Ask what separates them in law and the answer is nothing at all.
Texas is one of the few states that licenses close protection as its own thing. Texas Occupations Code § 1702.202 defines a personal protection officer as an individual who, “while carrying a firearm, provides to another individual personal protection from bodily harm.” The course specific to that licence, at § 1702.205, is “a 15-hour course for a personal protection officer consisting of training in nonlethal self-defense or defense of a third person” — and the statute adds that this “is in addition to the basic training course for security officers,” so a Texas protection officer stacks it on top of the commissioned-officer training at § 1702.1675. The implementing rule, 37 Tex. Admin. Code § 35.141(a)(3), says the fifteen “shall consist of a minimum of fifteen (15) classroom hours.”
Fifteen hours, in a classroom, in unarmed self-defence. Not one minute of any of it behind a wheel — the only occurrences of “driv” in the entire Act and its rules are driver-licence identity fields.
Virginia licenses the role too, and more heavily: a personal protection specialist registration requires a 60-hour entry-level compulsory minimum training course, plus advanced handgun training and range qualification, plus eight hours in-service. Sixty hours is four times Texas’s close-protection module. It also contains no driving instruction; the curriculum’s “protective detail operations” entry lists no sub-topics at all.
California licenses guards rather than protectors. Under 16 CCR § 643, the mandatory curriculum is eight hours of Power to Arrest and Appropriate Use of Force, then four mandatory four-hour courses: Public Relations, Observation & Documentation, Communication and its Significance, and Liability/Legal Aspects, plus eight hours of annual continuing education. Driving appears in the regulation exactly once, in the elective list, as “Driver Safety 4 – Hrs. Maximum,” itemised as “1. Cars 2. Bicycles 3. Golf Carts.” Executive protection is also elective, also capped, at “Introduction to Executive Protection 4 – Hrs. Maximum,” with no subjects specified.
Florida requires more hours and no protective driving. Fla. Stat. § 493.6303(4)(a) mandates 40 hours for a Class “D” security officer licence, and the curriculum incorporated by rule 5N-1.140 enumerates eighteen subjects — legal aspects, conduct, communications, incident reporting, access control, patrols, physical security, emergency preparedness, medical emergencies, terrorism and the rest. Vehicles appear as something to describe in a report, something to log, and two sub-items inside a one-hour block on patrols. The word “driving” does not appear in Chapter 493 at all.
So across the four regimes examined here — including both states that license close protection by name — the premium skill in protective driving is the one skill no mandated curriculum tests. It is sold privately, in courses whose length and content are set by the market. That is not an argument that the training is worthless — the opposite, probably, since nothing compels a buyer to pay for it. It is an argument about where the line between the two jobs sits, and it does not sit in a statute.
Availability is the product, and overtime is where it shows up
What a household is buying is not driving. It is a person whose location is a function of someone else’s calendar. That is expensive in a way the proxy statements do not capture and wage law does, and there is one place where the price is published: municipal payroll.
The weekly email
How wealth actually works, once a week
The new explainers, the Million Dollar Question and the numbers behind them, in one short email. Free, and no ads.
We send a confirmation link first. One-click unsubscribe in every email; your address is never shared.
New York City maintains job titles literally called Chauffeur and Chauffeur-Attendant, and publishes every incumbent’s pay in its Citywide Payroll Data. Querying the city’s own records for those titles, restricted to annual-salaried positions:
| Fiscal year | Positions | Avg base | Base range | Avg overtime paid |
|---|---|---|---|---|
| 2026 | 38 | $74,259 | $61,950–$89,906 | $23,141 |
| 2025 | 33 | $71,133 | $61,950–$89,906 | $24,120 |
| 2024 | 34 | $69,067 | $49,499–$93,820 | $25,059 |
| 2023 | 33 | $59,074 | $47,380–$77,119 | $20,918 |
The averages understate the point; the individual rows make it. In fiscal 2024 one chauffeur-attendant on a $68,005 base was paid $92,980 in overtime across 1,835.75 overtime hours — more overtime than base salary, and an average of about 35 extra hours every week of the year. In fiscal 2026 the one position titled simply “Chauffeur,” on an $85,060 base, was paid $51,509.75 of overtime across 819.75 hours, for about $134,581 in total. Those two are the extremes rather than the norm: the median across these positions runs between roughly 400 and 580 overtime hours a year, and some record none at all. The dataset names every employee; this piece reports titles and amounts only.
Why the hours run like that is a question of federal wage law, and the chauffeur sits in the middle of it.
The job is explicitly within the scope of the Fair Labor Standards Act’s domestic-service rules. 29 CFR § 552.3 defines domestic service employment by listing the work it covers, and the one place the word appears in that regulation is the phrase “chauffeurs of automobiles for family use.” A private chauffeur is a domestic service employee: minimum wage and, under § 552.100, overtime after forty hours — unless he lives in the employer’s household, in which case 29 U.S.C. § 213(b)(21) exempts the overtime but not the minimum wage, and § 552.102 still requires that any call to duty interrupting sleep or meals be counted as hours worked.
Then the hours themselves. 29 CFR § 785.41 provides that an employee who drives is “working while riding.” The harder question is the waiting, and § 785.14 answers it by quoting the Supreme Court in Skidmore v. Swift — a sentence that could have been written for this job: “Facts may show that the employee was engaged to wait or they may show that he waited to be engaged.” Under § 785.15 waiting is work time where the employer controls it and “waiting is an integral part of the job,” even if the employee may leave the premises; under § 785.17 on-call time is work where the employee must stay close enough that the time cannot be used for his own purposes.
Four hours parked outside a restaurant is the kind of waiting these rules were written to argue about, and the argument does not go one way. The provision the draftsman put next door, § 785.16(a), says that periods in which an employee is “completely relieved from duty and which are long enough to enable him to use the time effectively for his own purposes are not hours worked” — and a long, predictable wait is squarely its case, not § 785.15’s, which describes compensable waiting as “unpredictable” and “usually of short duration.” Whether a given chauffeur’s parked hours are paid is fact-specific, and that is the point: the central cost variable in this job is also the most legally contested one.
Which is why the honest answer to the cost question is not an hourly rate multiplied by driving time. It is a salary, plus the overtime that availability generates — and on the only payroll in this piece that discloses both, overtime ran from about a third of average base pay to, in one individual case, 137% of that driver’s own base salary.
What people get wrong
That the federal chauffeur number describes a private chauffeur. It describes employees of transit and ground passenger operators, nursing facilities, car dealerships and limousine companies. The OEWS technical notes exclude “household workers” and exclude “Private households (NAICS 814)” by name, so on the face of the survey’s own scope a chauffeur on a family’s payroll is outside the frame by construction rather than by accident. The $37,290 median is sound; it is an answer to a different question.
That the occupation is about chauffeuring. SOC 53-3053 is defined as transporting passengers “on a planned or scheduled basis,” may involve collecting a fare, and expressly “includes nonemergency medical transporters and hearse drivers.” Six of its ten sample job titles are explicitly forms of shuttle, van or coach driving; two more are a generic “Driver” and “Transportation Driver.”
That a chauffeur needs a chauffeur’s licence. In Michigan, the credential attaches to motor carriers, buses and vehicles over 10,000 lb GVWR, and the statute expressly excludes limousine and taxicab drivers from the definition, so a household driver is outside it. Iowa’s version turns on carrying persons “for wages, compensation, or hire” — language broad enough on its face to reach a salaried household driver, though the state’s own transport department describes the licence class as covering passenger-for-hire vehicles, and the question looks unsettled. Federally, a small vehicle requires a CDL and passenger endorsement once it is “designed to transport 16 or more passengers, including the driver.”
That a close-protection driver is a licensed specialism. In the jurisdictions examined here, no mandated security curriculum contains protective driving instruction. Texas mandates fifteen classroom hours of nonlethal self-defence on top of its basic officer training; Virginia requires a sixty-hour course whose protective-detail module lists no driving; California’s mandated list has none and offers a four-hour elective covering cars, bicycles and golf carts; Florida’s forty hours across eighteen subjects reach vehicles only as something to log, describe and patrol in. The distinguishing skill is unregulated.
That the proxy statements show what a driver costs. They show incremental cost under several incompatible conventions. Exxon’s disclosed cost of a car and security driver excludes the driver’s salary entirely and says so; Cisco excludes its in-house security staff’s compensation on the same reasoning while including its contracted driver’s actual cost; Goldman’s figure is a commuting and non-business share sitting in a column headed “Car” that in fact contains “driver compensation”; MGM’s $1,136 is a marginal mileage-and-hourly calculation; CACI’s $208,342 is a mileage-allocated slice of a larger undisclosed contract. Comparing them to each other, or to a salary, measures accounting policy.
That overtime is an inefficiency. For this job it is the product. The NYC records show individual drivers logging as many as 1,835.75 overtime hours in a single fiscal year, against a median of roughly five hundred. A household that wants a car available from six in the morning until midnight is not buying eighteen hours of driving; it is buying the exclusivity of someone’s day, and the law makes much of the waiting compensable.
Bottom line
The chauffeur inverts the usual problem in this series. Where the auction specialist has no federal occupation at all, the chauffeur has one with his name on the door — and the survey behind it was written to exclude his actual employer. The $37,290 median measures the shuttle-and-limousine workforce, from a frame that leaves out private households as a matter of published policy, annualised by multiplying an hourly rate by a notional 2,080-hour year.
What a chauffeur costs is disclosed only by the employers compelled to disclose it, and those are not households but public companies, reporting between $1,136 and $208,342 for arrangements that look the same from the back seat. The useful figure is none of them. It is the municipal payroll line where a driver on a $68,005 base collected $92,980 of overtime — the only public record here in which somebody pays, in cash and on a published schedule, for the thing the job actually sells.
And the line between the driver and the protection driver, the one the trade cares about most, appears nowhere in the licensing examined here — not even in the two states that license close protection by name. A statute will ask how heavy the vehicle is, how many seats it has, and whether the passengers are paying. It will not ask whether the man at the wheel has been trained to get the car out of a blocked street. Close-protection officers, private jet pilots and yacht captains all hold credentials that at least gesture at the hazard in their work. The chauffeur holds a driving licence, like everybody else.
Methods and sources.
Federal wage data. All sources read 17 November 2026. Figures for SOC 53-3053 are taken from the BLS Occupational Employment and Wage Statistics May 2025 national data file, national_M2025_dl.xlsx, downloaded from https://www.bls.gov/oes/special-requests/oesm25nat.zip and read directly rather than from any summary page: OCC_CODE 53-3053, “Shuttle Drivers and Chauffeurs,” TOT_EMP 248,530, H_MEAN 19.17, A_MEAN 39,880, H_MEDIAN 17.93, A_MEDIAN 37,290, H_PCT10 13.51, A_PCT10 28,090, H_PCT25 15.73, A_PCT25 32,710, H_PCT75 21.97, A_PCT75 45,700, H_PCT90 26.47, A_PCT90 55,070. The all-occupations comparison ($50,980 median, $69,770 mean, 155,495,730 employed) is row 00-0000 of the same file, and is corroborated by the release Occupational Employment and Wages — May 2025 (USDL-26-0725, released 10:00 a.m. ET Friday 15 May 2026), whose Table 1 first row reads “All occupations 155,495,730 $33.54 $69,770 $24.51.” Two BLS URLs now serve generic table indexes rather than data: bls.gov/oes/current/oes533053.htm (canonical tag bls.gov/oes/tables.htm, last modified 15 May 2026), linked above as the programme page only, and bls.gov/oes/current/oes_nat.htm, which is a bare index with no estimates on it; the all-occupations figures above are therefore linked to the data download itself rather than to that page. Industry and percentile figures attributed to the Occupational Outlook Handbook (page last modified 29 October 2026) are that page’s: shuttle drivers and chauffeurs median $37,290 in May 2025, lowest 10 per cent “less than $28,090,” highest 10 per cent “more than $55,070,” top-paying industry Taxi and limousine service $43,230, and the largest-employer shares of 25%, 8%, 7%, 7% and 6%. Note that the Handbook’s Quick Facts figure of “$37,960 per year, $18.25 per hour” is for the combined taxi/shuttle/chauffeur group and is not used above. Stated discrepancy. The OEWS file gives 248,530 employed in 53-3053; the Handbook’s Employment Projections table gives 251,500 for the same occupation and year. These come from two different BLS programmes with different scopes and the difference is reported above rather than reconciled. The occupation definition, the ten sample job titles, the Job Zone and the $17.93 hourly median are from O*NET 53-3053.00 (“Site updated October 6, 2026”), which cites BLS 2025 wage data.
Survey scope. The exclusions are quoted from the Technical Notes for May 2025 OEWS Estimates (last modified 15 May 2026): the coverage phrase “wage and salary workers in nonfarm establishments,” the sentence excluding “the self-employed, owners and partners in unincorporated firms, household workers, or unpaid family workers,” the separate statement that “Private households (NAICS 814) are also excluded,” the 2,080-hour convention (“the annual rate for an occupation is calculated by multiplying the hourly wage rate by a typical work year of 2,080 hours”), and the six-semiannual-panel construction of the May 2025 estimates. The same exclusion sentence appears in the OEWS FAQ at section D.2. Limit. The inference that a privately employed household chauffeur is therefore outside the survey frame is this piece’s reading of those two exclusions applied to this occupation, and is hedged as such in the text; BLS does not publish a statement specific to chauffeurs, and the FAQ notes that the May 2017 estimates “included for the first time some establishments that were previously classified in private households,” so the boundary has moved before.
Executive-perquisite disclosures. Every figure is from the filing linked at its first mention and was read in that document. ExxonMobil DEF 14A filed 2025 (FY2024): the itemisation “$44,703 for residential security,” “$31,906 for security costs related to personal travel,” “$21,295 for the cost of the car provided for security reasons” (the three sum to $97,904, the Personal Security column figure); the services list including “car and security driver”; and the quoted method sentence on salary, wages and benefits not being allocated. The claim that the $21,295 contains no driver pay is the filing’s own method statement applied to its own itemised figure; Exxon does not state a dollar figure for driver pay anywhere, and none is asserted here. Correction of a framing this piece originally used: the filing attributes a Board requirement only to use of the company aircraft — “the Board requires the Chairman and CEO to use the Company aircraft” — and not to the car or the driver, which the Personal Security paragraph describes as services provided on an assessment of risk. The text above is worded accordingly and does not say the board mandated the driver. MGM Resorts FY2023 proxy: “Other Perquisites” of $1,136 and the quoted incremental-cost method. Visa FY2023 proxy: “Car” column of $6,928 and the quoted method including driver salary and benefits. Goldman Sachs FY2018 proxy: “Car” column amounts of $61,978, $53,479, $41,020, $11,515 and $10,193 for named executives, and the quoted sentence that the cost of providing a car “includes, as applicable, driver compensation, annual car lease, car service fees and insurance cost”; only the largest is used above. The footnote opens “Amounts reflect the incremental cost to us attributable to commuting and other non-business use,” so the figure is a non-business share and not a full cost; the table label above says so. Walgreens Boots Alliance FY2024 proxy: “professional security driver/agent services” of $106,738 and $52,702 for two named executives; the label combines driving and close protection and the driving share is not separable, which is why the figure is presented above as a disclosure rather than as a driver’s cost. ManTech FY2019 proxy: $109,119 and the quoted allocation method, with FY2018 ($116,721) and FY2016 ($138,117) from the 2019 and 2017 proxies; the filing says the time was spent “primarily as a driver,” so an unquantified share is other personal staff work, and the figure covers only non-corporate use. CACI International FY2025 proxy: “personal security driver of $208,342” and the quoted method on third-party billings and mileage allocation; the FY2026 proxy reports $203,092 on the same label. Cisco FY2025 proxy: $284,581 of security expense, the quoted sentence on in-house compensation not being included, and the quoted sentence that “The incremental cost for the authorized car service and personal security driver is the actual cost.” The driver’s cost is therefore inside Cisco’s figure but is not broken out of it, and no driver amount is claimed from it; the contrast drawn above is between the two filers’ treatment of payrolled versus contracted drivers, not a claim that Cisco omits driver cost. Pfizer FY2025 proxy: the quoted “Car and Driver” policy language and the tax statement; its perquisite table shows Car Usage $19,093 and Security $610,650, neither of which isolates the driver. The tax point is reported above as Pfizer’s own characterisation of its position, not as a statement of US tax law; the underlying authority on security-related working-condition fringes was not retrieved and its conditions are not described here. Not obtained. No filing read discloses the full annual cost of employing a corporate driver, and none is stated above. The seven-filer table is a selection, not a census; it was assembled by full-text searching EDGAR for driver-related perquisite language and is not the result of a systematic survey of all filers, so it should not be read as a distribution.
Licensing. Michigan: MCL 257.6 (history ending “Am. 2016, Act 348, Eff. Mar. 21, 2017”), MCL 257.301 and MCL 257.7a, each read as the Legislature’s own PDF. The limousine and taxicab exclusions are subsections (3)(j) and (3)(k). On what 2016 PA 348 did: the pre-amendment § 257.6(1) listed a taxi driver and a limousine driver as chauffeurs, and § 257.6(2) read “necessary use of a motor vehicle for hire or for transporting passengers for hire, or for transporting for gain or hire any merchandise…”; PA 348 struck the passengers-for-hire language and moved taxi and limousine drivers into the exclusions, which is verified from the enrolled bill text. The current statutory text is what is quoted above. Iowa: Iowa Code § 321.1(8) and § 321.189(1)(a)(4), with the Class D chauffeur endorsement and the “Passenger vehicle less than 16-passenger design” code at 761 IAC 605.6(3). Stated as unsettled above, deliberately. An earlier version of this piece asserted that Iowa’s definition does not reach a household driver. It does not say that. The operative words are “for wages, compensation, or hire”; § 321.1(8)(a) exempts only operation that is “occasional and merely incidental” to the operator’s principal business; and § 321.1(8)(h) carves out a home care aide by name, which suggests domestic-type driving would otherwise fall inside. The contrary indication is the Iowa DOT’s own description of Class D as covering “small passenger-for-hire vehicles,” which is agency guidance rather than statute. No Iowa case, attorney-general opinion or agency determination applying the definition to a privately employed household driver was found, and the text above says the question is unsettled rather than answering it. Federal: 49 CFR § 383.5 definition of “Commercial motor vehicle,” Group C branch, and § 383.91(a)(3), § 383.93 and § 383.117 for the passenger endorsement’s knowledge and skills tests. Note on a common error: 49 CFR Part 383 contains no definition of the term “passenger vehicle”; the 16-passenger threshold lives in the CMV definition and the Group C vehicle-group description, which is how it is cited above. Louisiana was checked and its driver’s-licence definitions section contains no occurrence of “chauffeur.” Indiana’s chauffeur’s-licence and public-passenger-chauffeur’s-licence chapters (IC Art. 9-24, chs. 4 and 5) are marked expired in the current code, so Indiana no longer issues one; the Indiana General Assembly’s own site serves its code through a JavaScript application that returns nothing to automated retrieval, so that status comes from a republication of the current code rather than from the Assembly, and it supports no claim in the body. No count of how many US states retain a chauffeur licence or definition was obtained from an authoritative source, and none is given; every statement about licensing above is scoped to the states named.
Security training. Texas: Occupations Code §§ 1702.202, 1702.205 and 1702.1675 and 37 Tex. Admin. Code §§ 35.141(a)(2) and (a)(3), read in the Department of Public Safety Private Security Bureau’s own compilation of the Act and its rules; the Legislature’s own site and the Texas Register have both migrated to front ends that return no statutory text to automated retrieval, so the regulator’s publication was used, and a fact-checker wanting the Legislature’s rendering would need a manual browser visit. The fifteen hours is not the whole requirement, and the text above says so: § 1702.205(b) states the training “is in addition to the basic training course for security officers,” and 37 TAC § 35.141(a)(2) requires the Level III course of personal-protection-officer applicants as well. The claim that no driving appears anywhere in the Texas regime rests on a search of that whole compilation for “driv,” whose only hits are driver-licence identity fields and one unrelated vehicle-service exclusion. Virginia: the personal protection specialist registration, its 60-hour entry-level compulsory minimum training course, the in-service and handgun requirements and the absence of any driving topic are from 6VAC20-174-300 and the related sections of 6VAC20-174. Virginia was added after an initial version of this piece wrongly called Texas the only state licensing close protection as such; it is not, and the claim above is now that Texas is one of a few. No systematic survey of all fifty states’ close-protection registrations was done. California: 16 CCR § 643 and its Appendix, in the Bureau of Security and Investigative Services’ own publication of Title 16 Division 7, which states it includes regulations enacted through November 2023; the mandatory courses, the elective “Driver Safety 4 – Hrs. Maximum” with its three itemised sub-topics, and the elective “Introduction to Executive Protection 4 – Hrs. Maximum” are verbatim from that Appendix. The California statutes themselves (Bus. & Prof. Code §§ 7583.6, 7583.7) could not be retrieved — leginfo.legislature.ca.gov refuses automated retrieval — so the hour requirements above rest on the Bureau’s regulation and its own fact sheets rather than on the statute. Florida: Fla. Stat. § 493.6303(4)(a), Fla. Admin. Code r. 5N-1.140(1)(a) (effective 28 November 2024), and the incorporated Security Officer Training Curriculum Guide FDACS-P-01878 (Rev. 11/2018), whose eighteen subject headings and hour allocations were read in full; Chapter 493 was searched for “driv” with no occurrence. New York was looked at and is not relied on in the text: the detailed curriculum regulations at 9 NYCRR Part 6027 could not be verified from an official source, the official publisher returning HTTP 403 and the agency’s own compilation now 404ing. The claim is scoped: no mandated protective-driving instruction was found in the four regimes quoted above. It is not a claim about all fifty states, none of which beyond those named was examined. Note also that Florida’s incorporated curriculum does contain two vehicle sub-items — “7.3.1 Types” and “7.3.2 Vehicle Safety,” inside a one-hour Patrols block — which is why the text says no protective or evasive driving rather than no mention of vehicles.
New York City payroll. From NYC Open Data’s Citywide Payroll Data (Fiscal Year), published by the Office of Payroll Administration, queried directly through the Socrata API for rows whose title description contains “CHAUFFEUR” and filtered to pay_basis “per Annum”, with an explicit row limit above the number of matching rows (513 across all years) so that no result was truncated. The dataset returns three title spellings — CHAUFFEUR-ATTENDANT, CHAUFFEUR ATTENDANT, and one position per year titled simply CHAUFFEUR in FY2025 and FY2026 — and the table above aggregates all three. The individual rows quoted are: fiscal 2024, base $68,005, regular gross $70,853.02, overtime $92,980.77 across 1,835.75 overtime hours, other pay $3,624.34; and fiscal 2026, title CHAUFFEUR, base $85,060, regular gross $82,452.60, overtime $51,509.75 across 819.75 overtime hours, other pay $618.96, totalling $134,581.31. The “about 35 extra hours every week” figure is this piece’s arithmetic (1,835.75 ÷ 52 = 35.3) and assumes a full year, which the record does not state. The totals are amounts paid and exclude the city’s pension and health-benefit costs, so the employer’s true cost is higher. The dataset identifies employees by name; names are deliberately not reproduced here, and no individual is named — though since only one position per year carries the bare title “Chauffeur,” that row is identifiable to anyone who runs the query, which is a reason the figure and not the person is what appears above. Correction. An earlier version of this table reported 31 positions for FY2026 and FY2025 with correspondingly different averages; those rows were artefacts of a truncated query, and the figures above come from an uncapped re-run. The FY2024 and FY2023 rows were unaffected. Limits. These are government drivers for public officials, not household chauffeurs, and are used above as the one published record of what availability costs — not as a proxy for private pay. Overtime hours vary widely within each year, from zero to the maxima quoted, so the two individual rows above are extremes and are labelled as such. Whether any of these positions involve protective duties is not stated in the dataset and is not claimed.
Wage law. 29 CFR Part 552 (current as of the eCFR retrieval; § 552.3 amended at 78 FR 60557, 1 October 2013): § 552.3’s inclusion of “chauffeurs of automobiles for family use,” § 552.100(a)(2) on overtime after forty hours unless the employee resides in the household, and § 552.102(a) on minimum wage for live-in workers and interrupted sleep and meal periods. 29 CFR Part 785 Subpart C: § 785.14’s “engaged to wait”/”waited to be engaged” formulation, § 785.15 on controlled waiting and “waiting is an integral part of the job,” § 785.16(a) on periods “long enough to enable him to use the time effectively for his own purposes,” § 785.17 on on-call time, and § 785.41 that a driver “is working while riding.” The “engaged to wait” sentence is the regulation quoting the Supreme Court in Skidmore v. Swift, 323 U.S. 134 (1944), and is attributed that way above; that citation is taken from the face of the regulation and the opinion itself was not read. 29 U.S.C. § 213: (b)(21) exempting live-in domestic service employees from overtime, (b)(17) covering drivers for taxicab businesses, and (a)(15) covering casual babysitting and companionship services. This is a description of the regulations, not legal advice. The application of the waiting-time rules to a chauffeur parked outside a restaurant is presented above as contested between § 785.15 and § 785.16(a) rather than resolved; whether any particular arrangement is compensable is fact-specific, and no court decision, opinion letter or FLSA litigation involving chauffeurs was examined.
Not verified. Any pay figure for a privately employed household chauffeur, from any source. The full annual cost to any employer of employing a driver. Any current rate card from any chauffeur or executive-car agency. Whether a security driver’s pay differs systematically from a chauffeur’s. Whether Iowa’s chauffeur definition reaches a household driver. The number of US states that license or define chauffeurs, or that license close protection. The mandated security curriculum of any state other than Texas, Virginia, California and Florida, and the official text of New York’s, whose curriculum regulations at 9 NYCRR Part 6027 could not be retrieved from an official publisher. Whether the NYC positions above involve any protective function. Whether other public filers disclose driver costs more fully than the seven above. The conditions attaching to the tax treatment Pfizer describes. Any claim about what protective driving instruction actually teaches, which was not researched; the text says only that no examined curriculum mandates it.
Related reading: Cars: Luxury, Collecting, and the Cost of Ownership · Staff: Outsourcing Daily Life · Chefs, Nannies, and Household Help: The Labor Behind Affluent Life · Personal Security: Protection, Privacy, and Risk · Privacy: Why the Wealthy Value Invisibility · The Job That Sees It: The Close-Protection Detail · The Job That Sees It: The Private Jet Pilot · The Job That Sees It: The Yacht Captain · The Job That Sees It: The Estate Manager · The Job That Sees It: The Butler · The Job That Sees It: The Housekeeper · The Job That Sees It: The Auction Specialist



