Phoenix-metro trade, transportation and utilities employers had 479,100 people on payroll in September 2026, up 1.0% from a year earlier. That is 19.4% of all payroll jobs in the metro, against 19.3% a year before.

The broadest supersector in the survey: wholesale and retail trade, warehousing, trucking, air and water transport, couriers and utilities.

The two charts answer different questions. The count peaked at 493,000 in December 2025 and says whether the sector is hiring. The share — from 22.2% in January 1990, with a high of 22.2% in January 1990 and a low of 18.6% in September 2019 — says whether it is growing faster or slower than the rest of the economy.

Not seasonally adjusted, so compare a month with the same month a year earlier. The share is computed at read time from two series in the same BLS release, so it stays consistent when the levels are revised.

Phoenix-metro financial activities employers had 205,600 people on payroll in September 2026, down 2.5% from a year earlier. That is 8.3% of all payroll jobs in the metro, against 8.6% a year before.

Finance and insurance plus real estate, rental and leasing — banks, insurers, mortgage lenders, brokers and property managers. Real estate and lending also make it sensitive to interest rates in a way most sectors are not.

The two charts answer different questions. The count peaked at 219,900 in December 2021 and says whether the sector is hiring. The share — from 7.9% in January 1990, with a high of 10.2% in April 2020 and a low of 7.1% in May 1995 — says whether it is growing faster or slower than the rest of the economy.

Not seasonally adjusted, so compare a month with the same month a year earlier. The share is computed at read time from two series in the same BLS release, so it stays consistent when the levels are revised.

Phoenix-metro professional and business services employers had 388,300 people on payroll in September 2026, up 2.2% from a year earlier. That is 15.7% of all payroll jobs in the metro, against 15.5% a year before.

This is one of the metro’s largest supersectors, and the least understood. It covers professional, scientific and technical services — law, accounting, engineering, architecture, computer systems design — plus the head offices of companies, and administrative support, which includes temp agencies and building services. A good share of what Phoenix calls tech work is counted here rather than under ‘information’. The temp-help piece makes it one of the first sectors to cut in a downturn: agency workers are the easiest jobs to end.

The two charts answer different questions. The count peaked at 396,800 in December 2023 and says whether the sector is hiring. The share — from 10.8% in January 1990, with a high of 17.5% in July 2016 and a low of 10.7% in November 1990 — says whether it is growing faster or slower than the rest of the economy.

Not seasonally adjusted, so compare a month with the same month a year earlier. The share is computed at read time from two series in the same BLS release, so it stays consistent when the levels are revised.

Phoenix-metro education and health employers had 431,500 people on payroll in September 2026, up 3.0% from a year earlier. That is 17.4% of all payroll jobs in the metro, against 17.1% a year before.

This is the private sector only: hospitals, clinics, nursing and residential care, social assistance, private schools and colleges. Public schools and the state universities are counted as government. It is the steadiest sector in the data, because demand follows the size and the age of the population rather than the business cycle — the one sharp break in its line is spring 2020, when elective care stopped.

The two charts answer different questions. The count peaked at 434,400 in April 2026 and says whether the sector is hiring. The share — from 8.9% in January 1990, with a high of 17.4% in April 2026 and a low of 8.8% in November 2000 — says whether it is growing faster or slower than the rest of the economy.

Not seasonally adjusted, so compare a month with the same month a year earlier. The share is computed at read time from two series in the same BLS release, so it stays consistent when the levels are revised.

Arizona workers filed 2,394 initial unemployment claims in the week ending Oct 17, 2026; the steadier four-week average is 2,387, down 12.3% from the same weeks a year ago.

Read the chart in two parts. The y axis is broken at 12,000. Below the break is where this series lives: almost every week since 1986 is in that band, and on its own scale the normal shape is plain — a spike every winter as construction and seasonal retail lay off, and a sustained rise through 2008-09. Above the break, on a compressed scale, is the spring of 2020, when claims reached 132,428 in the week of Apr 4, 2020. On a single axis that one episode flattens 40 years into a line along the floor.

Claims are the fastest labour-market number there is — a week old when published — and the noisiest. One week is a data point; the four-week average is the reading. Statewide, not seasonally adjusted, so compare with the same weeks a year earlier rather than the week before.

The Phoenix metro’s information sector employed 40,800 people in September 2026, up 2.0% from a year earlier. That is 1.6% of every payroll job in the metro, against 1.6% a year before.

The two charts answer different questions. The head count says whether the sector is hiring. The share says whether it is hiring faster than everyone else — a sector can add jobs every month and still shrink as a part of the economy. The count peaked at 44,200 in May 2022; the share peaked at 2.8% in July 2000, up from 2.4% when the series opens in January 1990.

‘Information’ is not ‘tech’, and the gap is large. It is the BLS sector covering software publishers, telecoms, broadcasting and publishing. A great deal of Phoenix technology work is booked elsewhere — chipmakers such as Intel and TSMC are classified under manufacturing, cloud infrastructure lands under several other headings, and contract engineering sits in professional services — so read this line as a floor on tech employment rather than a measure of it. What it is good for is the trend: the classification has been applied consistently since 1990, so the shape is right even where the level is low.

Not seasonally adjusted, which is why both lines have an annual wobble on top of the trend; compare a month with the same month a year earlier. The share is computed at read time from two series in the same BLS release, so it stays consistent when the levels are revised.

The Phoenix metro’s construction and mining employers had 188,500 jobs on payroll in September 2026, up 3.5% from a year earlier. That is 7.6% of all payroll jobs in the metro, against 7.4% a year before.

Construction is the most cyclical sector the site tracks and the one that turns first. The share chart shows it without the flattery of a growing economy — it peaked at 10.0% in July 2006 and bottomed at 4.7% in February 2011, a measure of how much of the region’s labour is pointed at building anything.

It has a real summer season in the raw data, so — as everywhere in these NSA series — the year-over-year change is the number that means something and a fall from August to January is just weather. It sits alongside building permits as this site’s two reads on whether Phoenix is building: permits lead; payrolls are what actually happened. The share is computed at read time from the same BLS release as the count.

Phoenix-metro restaurants, bars, hotels and venues employed 255,900 people in September 2026, down 1.8% from a year earlier — 10.3% of all payroll jobs in the metro.

No other series on this site records 2020 as plainly. Employment in this sector fell from 242,300 in February 2020 to 140,600 in April 2020 — a drop of 42%, inside two months, in an industry that cannot be done from home — and its share of the region’s jobs fell to 7.1%. Everything to the right of that notch is the recovery.

The share chart is the longer story: from 10.4% when the series opens in January 1990 to a peak of 11.3% in April 2017.

Not seasonally adjusted, and summer is genuinely the busy season here, so compare with the same month a year earlier. The share is computed at read time from the same BLS release as the count.

Phoenix–Mesa–Chandler employers reported 2,475,200 nonfarm jobs in September 2026 (not seasonally adjusted), up 23,600 — +1.0% — from a year earlier. Set against everyone who lives in the metro, that is 47.3 payroll jobs for every 100 residents, against 46.9 a year before.

The second chart is there because a count of jobs always grows in a growing region. Jobs per resident does not have to: it peaked at 49.1 in December 2000, fell to 38.3 in July 2009, and says whether the job machine is keeping up with the people arriving to use it.

It is not the employment rate, and the difference matters. The official rate counts people with work, by where they live, as a share of adults. This counts jobs at employers in the metro — a second job counts twice, and someone commuting in from outside the metro counts here — divided by every resident, children and retirees included. So an ageing population pulls it down with no change in the job market. Read the direction, not the level.

The population is the Census Bureau’s July 1 estimate for the same two-county metro, interpolated between years; the chart starts in 2000 because that is where the annual estimates on FRED begin. Because payrolls aren’t seasonally adjusted, the year-over-year change is the one to watch.

Phoenix-metro manufacturers employed 147,500 people in September 2026, down 0.3% from a year earlier — 6.0% of all payroll jobs in the metro. The sector peaked at 168,100 jobs in June 1998 and has shed roughly 12% of that since.

The second chart is the share: it went from 13.7% in January 1990 to 6.0% now.

Numerator and denominator of the share come from the same BLS release for the same metro, so it stays consistent when the levels are revised; it is computed at read time, never stored.

Homes in the Phoenix metro sold for 97.8% of their list price on average in August 2026, against 97.7% a year earlier, and 14% of them sold above it. The most competitive month in this record was June 2021, at 103.3% of list.

At September 2026’s pace of sales, the homes for sale would last 4.7 months — months of supply, for-sale inventory over a month’s sales. Under about four months is a market where sellers set the terms; six or more is one where buyers do. 5,124 homes sold in September 2026, against 5,189 a year earlier.

Every one of these has a season — spring is the bidding season — so compare a month with the same month a year earlier. These are Zillow’s figures for the Phoenix-Mesa-Chandler metro, all homes, not seasonally adjusted; the newest month of sales is Zillow’s nowcast and firms up later. There is no median price here, on purpose: a median moves with which homes happen to sell, so a month heavy in condos reads as a fall in prices that did not happen.

The typical asking rent in the Phoenix metro was $1,720 a month in September 2026, up 1.0% from a year earlier.

The second chart takes inflation out. The price index runs through September 2026, and in that month’s dollars the typical asking rent was $1,720, down 2.6% in real terms on the year and 15% below its real peak of $2,024 in December 2021. When the nominal line climbs and the real one does not, rents are rising more slowly than prices in general — which is not the same thing as rent getting affordable, since that depends on what renters earn.

ZORI tracks new leases, not the average rent paid across all tenants, so it turns roughly a year before the rent component of the local CPI does — which makes it the leading indicator of the two. The real series is deflated by the national CPI-U (all items, NSA) and stated in dollars of the latest month the CPI covers, so it can end a month before the nominal line. Rent is itself part of the CPI, so the real line is rent measured in purchasing power, not rent net of rent inflation.

The typical Phoenix-area home was worth $441,097 in September 2026, and the typical asking rent was $1,720 a month. Divide the first by a year of the second and you get the price-to-rent ratio: 21.4. A Phoenix house costs about 21 years of the rent on it.

It is the cleanest single number for whether housing is expensive relative to itself. Prices and rents are both driven by local incomes and local supply, so when they rise together the ratio holds flat and the market is simply pricier; when prices pull away from rents, buyers are paying for something other than the shelter — expected appreciation, or cheap credit. Nationally a ratio in the high teens is typical. Phoenix’s has run from 12.4 (September 2011) to 28.5 (June 2006), and it is 21.4 now, down 2.1% on the year.

Two things it is not. It is not a buy-versus-rent recommendation: that depends on mortgage rates, the tax treatment of the interest, and how long you stay, none of which is in here. And the two legs measure slightly different housing — ZHVI values the whole owner-occupied stock, ZORI tracks asking rents on units coming to market, which skew smaller and newer. The level is therefore approximate; the direction is the signal.

Arizona workers filed 2,394 initial unemployment claims in the week ending Oct 17, 2026; the steadier four-week average is 2,387, down 12.3% from the same weeks a year ago.

Read the chart in two parts. The y axis is broken at 12,000. Below the break is where this series lives: almost every week since 1986 is in that band, and on its own scale the normal shape is plain — a spike every winter as construction and seasonal retail lay off, and a sustained rise through 2008-09. Above the break, on a compressed scale, is the spring of 2020, when claims reached 132,428 in the week of Apr 4, 2020. On a single axis that one episode flattens 40 years into a line along the floor.

Claims are the fastest labour-market number there is — a week old when published — and the noisiest. One week is a data point; the four-week average is the reading. Statewide, not seasonally adjusted, so compare with the same weeks a year earlier rather than the week before.

The typical Phoenix-metro home was worth $441,097 in September 2026, down 1.1% from a year earlier. Against the series high of $497,312 in July 2022, that leaves the market 11.3% below its high-water mark in the dollars on the price tag.

The second chart takes inflation out. The price index runs through September 2026, and in that month’s dollars the typical home was worth $441,097, down 4.6% in real terms on the year. The real peak was $565,568 in July 2022, and today sits 22% below it. A flat price in a year of inflation is a fall in what a house is worth measured in everything else a household buys.

ZHVI measures the value of the typical home rather than the price of the homes that happened to sell, so it does not lurch when an expensive neighbourhood has a busy month — which is what makes it readable month to month, and a better trend line than asking prices.

The real series is deflated by the national CPI-U (all items, not seasonally adjusted) and stated in dollars of the latest month the CPI covers, so it can end a month before the nominal line. National rather than Phoenix CPI because the Phoenix index is published only every other month. Shelter is about a third of CPI-U, so read the real line as a house measured in purchasing power, not as house prices net of housing inflation. Metro-wide; Zillow revises the back series when its model is updated, so older points can move slightly between builds.

2.00 million passengers boarded flights out of Sky Harbor in July 2026, up 0.4% from a year earlier. 97,554 of them were on international flights, down 2.0%. Over the last twelve months departures averaged 2.11 million a month, against 2.14 million at the record, in the year to December 2024.

These are departures only, from the airlines’ own monthly reports to the federal government. The City of Phoenix Aviation Department’s headline passenger count adds arrivals, so it runs about twice these figures. A traveller connecting through Sky Harbor counts on the flight they leave on. Summer is the peak and winter the low every year, so compare a month with the same month a year earlier. The figures arrive about three months after the month they count.

Phoenix’s open-data portal lists 61,295 police calls for service for August 2026, up 27.3% on a year earlier. The 12-month average is 54,220 a month. That is 36.80 police calls for service per 1,000 residents.

This counts the rows of a police data feed, so it moves with how and when events are recorded as well as with how many happen. It is a count of records, not the department’s own crime statistic, and a change in the reporting system shows up as a step in the line.

The heavier line is the 12-month average, which takes out the seasonal swing; compare a month with the same month a year earlier rather than with the month before. The newest month is left out until it is complete.

The second graph divides the count by Phoenix’s population: the Census Bureau’s July 1 estimate for the city, interpolated to months. A growing city can post more records at a flat rate.

The Phoenix–Mesa–Chandler labor force — everyone employed or actively looking for work — was 2,670,163 in August 2026, down 2.7% from a year earlier. It’s the head-count denominator under the unemployment rate: a labor force that grows with population and participation. Not seasonally adjusted.

Arizonans filed 14,760 new business applications in September 2026, up 19.8% from a year earlier. These are EIN filings, so the series measures intent to start a business rather than businesses that actually open — but it lands about 11 days after the month ends, which makes it one of the fastest real-activity readings available for the state. Note the step-change after 2020: application rates never returned to their pre-pandemic level.

Arizona workers filed 2,394 initial unemployment claims in the week ending Oct 17, 2026; the steadier four-week average is 2,387, down 12.3% from the same weeks a year ago.

Read the chart in two parts. The y axis is broken at 12,000. Below the break is where this series lives: almost every week since 1986 is in that band, and on its own scale the normal shape is plain — a spike every winter as construction and seasonal retail lay off, and a sustained rise through 2008-09. Above the break, on a compressed scale, is the spring of 2020, when claims reached 132,428 in the week of Apr 4, 2020. On a single axis that one episode flattens 40 years into a line along the floor.

Claims are the fastest labour-market number there is — a week old when published — and the noisiest. One week is a data point; the four-week average is the reading. Statewide, not seasonally adjusted, so compare with the same weeks a year earlier rather than the week before.