After inflation, New York City's minimum wage nearly doubled since 1989, and about half its wage-earners still make under $30 an hour, as they did then

• 36.7% of the city’s jobs under $30 (the Economic Policy Institute’s projection for 2030, August 2025). A projection, not a count: the Institute took the 2023 one-year census file and grew each worker’s pay at 3.3% a year to 2030. Doing the same with the same file, and counting work done in the city, gives 36.5% under thirty dollars, 0.2 points from the published 36.7%, and a ratio of the 2030 minimum to the 2030 median of 0.75 against its 0.76. Counting residents instead gives 40.7%, further from it, so it most likely counts work done in the city, including commuters; the Institute’s post does not say.
• “The highest in the nation” (WABC, 10 March 2026). A ranking of legal rates in other cities and states, which these files don’t hold; we don’t rebuild it.
• The existing rate “leaves too many New Yorkers in poverty” (the bill’s sponsor, as WABC reported her). Not testable here: these files count pay, not poverty.
• Smaller companies reach the mark two years later (WABC: they “would hit this mark by 2032”). The bill as introduced sets $29 in 2031 for employers of 500 or fewer and an indexed rate from 2032; it names no thirty-dollar step for them. The report’s “less than 500” follows the bill’s own summary; its text divides employers at more than 500 against 500 or fewer.
• Jobs “declined slightly, by a monthly 0.07% from January to July” (the Governor’s memo). The memo’s figure is for private education, hospitality and retail jobs; the law’s test is the State’s total seasonally adjusted job count. Both come from the State’s payroll count of employers, which these files don’t rebuild, so the figure is quoted, not checked.

Because each counts different people in different dollars, as the table in the story sets out: residents against everyone working in the city, private employers against all, 2024 dollars against this year’s, and a count against a projection. The census’s 51% describes pay in 2024 from a year’s earnings; the monthly survey’s 52% covers January to September 2026 from a usual week’s pay, in this year’s dollars. Subtracting one from the other would not measure a change.

As sure as the inflation assumption under it. Prices in the New York area are measured to August 2026; after that we assume 2.5% a year, which is an assumption, not a forecast. At 2% a year the 2030 rate would be worth $25.86 in 2024 dollars, and at 3.5% $24.60. Across that range, about two in five private-sector workers were paid less than it in 2024.

Less reliable at the bottom than in the middle. The census does not ask for an hourly wage: we divide a year’s wages, tips and commissions by the usual weekly hours times the weeks worked, and each answer carries some error. The result is an average over the year, across every job a person held. 18% of residents in 2024 come out below $15.57, the lowest general minimum in force over the year the survey asked about. The arithmetic can’t tell reporting errors, job changes, mixed wage and self-employed work or underpayment apart, so we quote no share below about twenty dollars in the story.

The monthly survey asks directly about a usual week’s pay and hours, but it fills in missing earnings from similar respondents nationwide: this year 46% of the city’s records had their earnings filled in, and among the others the share under thirty dollars was 48%. Its median this year, $28.85, equals $60,000 a year over a forty-hour, 52-week year.

Not much. We put every year in 2024 dollars with the New York-area consumer price index. The Bureau of Labor Statistics also publishes a research series that restates the national index for past years with today’s methods, which measure less inflation; part of the difference is also national prices against New York’s. On that series the 1989 median comes to $29.48 instead of $30.72, even closer to 2024’s $29.29, and the 1989 minimum to $8.23. Either way, the middle ended about where it began.

The general rate stays at $17 in New York City, Long Island and Westchester. From 2027 the law raises the rate each year by the average change over the three latest August-to-July periods in the Northeast region’s consumer price index for urban wage earners and clerical workers, a different index from the New York-area one our dollars use. Any of three conditions stops the following year’s rise:

• that index’s average for the latest August-to-July period is below the preceding period’s;
• the three-month average of the State’s seasonally adjusted unemployment rate to July is half a point or more above its low of the previous twelve months;
• the State’s seasonally adjusted total non-farm employment in July is below both April’s and January’s.

A pause can run at most two years in a row. The Governor’s memo says the employment condition was met, and gives a figure for private education, hospitality and retail jobs, a monthly decline of 0.07% from January to July. Her bill would replace that jobs test alone, not the other two, with one that pauses the raise when the median of the monthly rates of change worked out from the latest twelve, six and three months falls below −0.25%, and would set the new rates for 1 March 2027. Home care aides are the exception: their rate rises to $20.00 in the city on 1 January. In 2024 dollars this year’s rate is worth $15.86, a little less than 2024’s, and with no raise it will be worth less again next year if prices keep rising. Tipped food-service workers have a lower cash wage, made up by tips.

Employers with more than 500 employees nationwide, and franchisees of networks that large, would pay at least $20 from January 2027, then $23, $26 and $30 in 2030. Employers of 500 or fewer would pay $19 in 2027 rising to $29 in 2031. After that both rates would follow the national consumer price index. Federal, State and City government employers are excluded. Until 1 December 2031 a food-service worker who receives tips could be paid at least two-thirds of the rate, or the State’s tipped rate if higher, provided the tips make up the rest. The bill, Int 0757-2026, was introduced by Council Member Sandy Nurse on 10 March with fifteen Council sponsors and referred to the Committee on Consumer and Worker Protection, and its record shows no action since. The city can’t do it alone: a State bill that would let it set a minimum above the State’s was re-referred to the Senate’s Labor Committee on 7 January 2026.

By industry, health care and social assistance had the most, about a quarter, and in retail and in restaurants, hotels and the arts about seven in ten workers were paid under thirty dollars. By occupation group, about four in five service workers were. Women were a little more often under the line than men, and part-time workers far more often than full-time ones. The data file at the end has every group with its margin.

Pay. Hourly pay is a year’s wage and salary income, tips and commissions included, over usual weekly hours times weeks worked, for people 16 and over who work for wages (private, non-profit and government employees; the self-employed are left out) and were at work or had a job when interviewed. It is an average over the twelve months before the interview (the censuses asked about the previous calendar year), across every job held, assigned to the person’s current occupation and workplace. A rate under a dollar an hour is left out, a cleaning rule for extreme values rather than a diagnosis of each record. The 1990 and 2000 censuses asked about the previous year, so their points sit at 1989 and 1999. From 2008 to 2018 the census recorded weeks in bands; we use each band’s average in the years it recorded the exact number (2005 to 2007), since most people in the top band worked all 52 weeks. The 1990 and 2000 figures come from the census long form and the rest from the American Community Survey, a change of survey. Full-time means usual hours of 35 or more a week.

Dollars. The ACS files from 2015 are in 2024 dollars already; earlier years are adjusted with the Bureau’s within-year factor and the New York-area consumer price index, and the censuses with the index of the year their pay was earned. The 2026 minimum uses the index’s months to August. The law indexes the minimum wage to a different series, the Northeast region’s CPI-W.

The minimum wage. The higher of the State’s and the federal rate on each day, averaged over the year; from December 2016 the City’s rate for employers of 11 or more, which reached fifteen dollars a year before the rate for smaller employers. Each rate is checked against the State’s and the federal histories.

The history. Each date, rate and figure the story takes from a law, a release or a memo is looked up in the saved copy of that document when the figures are built, and the build fails if the words are not there; the figures printed from them (the $11 and $15 of 2016, the 2.3 million, the $16 of 2024, the 0.07% and the 0.25%, the fifteen sponsors) are read from those quotes, not typed. They describe the laws, not the surveys, and none is set against a survey count.

Working in the city. Employees whose place of work is in the five boroughs and who live in the New York metropolitan area the extract covers (the city and its suburbs in the three states), plus city residents who work from home. The census asks where people worked only of those at work in the survey week, so the 2% of employed residents who had a job but were away from it are counted among residents, not here. Private means private for-profit and non-profit employers.

This year. The Current Population Survey’s outgoing rotation groups, January to September 2026: usual weekly earnings on the main job over usual weekly hours, in this year’s dollars, with the earnings weight. Margins are clustered on the household and are approximate.

Margins. Census margins are 90% margins from the replicate weights; a median’s margin is a Woodruff interval, the method behind the Bureau’s published medians, read off quarter-dollar bins, and the 2008-2018 margins take the weeks-band means as fixed. No margin is calculated for the 1990 and 2000 samples. The 2020 file was experimental and its margin is not shown, and the 2005 file left out people in group quarters, whom the later files include. Groups with fewer than 30 records are not published. A change’s margin combines the two years’ margins as independent samples. A change is called a rise or a fall only when its range clears zero: the share under thirty dollars from 2019 to 2024 does, and the median’s change does not.

Code and checks. One script, analysis.py (linked below), produces every figure and chart and checks each rate and quotation against the saved pages. The dataset cards behind it (census_history, acs1_pums, cps_basic, bls_labour) list the traps.