A county-by-county look at local news capacity in New Jersey, from staffed newsrooms to the towns that no reporter visits.
New Jersey has 21 counties and 564 municipalities. This showcase report counts the people paid to cover them. On the invented count taken in June 2026, 210 paid reporting jobs serve the whole state, and 3 counties hold half of them. Six counties hold one staffed newsroom or none. Counties with a shared reporting desk published 34 percent more municipal meeting stories over the 12 months to June 2026. The gap follows population and tax base, not distance from Trenton.
New Jersey holds 21 counties and 564 municipalities.1 Each municipality runs its own council. Most run their own school board, their own police contract and their own zoning board. A reader who wants to know what the council did on Tuesday needs a reporter in the room on Tuesday.
A county is a convenient unit for a chart. It is the wrong unit for a reader. Bergen County alone holds 70 municipalities. A single reporter who covers Bergen County covers 70 councils, and no reporter can do that.
So this report counts reporting capacity: the paid people who produce local news for a place, and the number of places each of them must cover. It counts jobs, not bylines, because a byline can come from a wire service, a press release or a national desk that visited once.
The count leaves out unpaid contributors, national desks, and any outlet that publishes less than once a month. The rules are strict. They also make the number small, and a small number is easier to argue with, which is the point.
Three counties hold 105 of the invented state total of 210 paid reporting jobs. The other 18 counties share the remaining 105 jobs. That split is the result this report returns to, and it holds whether the count is measured against population, against municipalities or against school districts.
A thin county is not a county with no news. It is a county where the news arrives late, arrives from outside, or arrives only when something breaks. Residents still hear about the school budget. They hear about it after the vote.2
Measured in reporting jobs.
Essex County holds 41 of the invented 210 paid reporting jobs, and Passaic County holds 6, on the count taken in June 2026.
Invented data for this showcasecenterforcooperativemedia.org
Invented values. A job counts once, in the county where the newsroom keeps its desk.
| Category | Value |
|---|---|
| Essex | 41 |
| Bergen | 38 |
| Hudson | 26 |
| Middlesex | 17 |
| Monmouth | 14 |
| Union | 11 |
| Camden | 9 |
| Passaic | 6 |
The total covers 564 municipalities. That is one paid reporting job for every 2.7 towns, before any reporter takes a day off.
Invented count of jobs of 20 hours a week or more, taken in June 2026. The trend line covers 2022 to 2026.
Four counties hold one staffed newsroom. Two hold none. The six together hold 61 municipalities and 412,000 residents.
Invented count taken in June 2026. The trend line covers 2022 to 2026 and counts the same six counties each year.
Counties with a shared desk filed more municipal meeting stories than counties without one. The gain is a difference between two groups, not a measured effect.
Invented comparison of 8 counties with a shared desk and 13 without, over the 12 months to June 2026.
| Essex | 22 | 41 | 18.6 |
|---|---|---|---|
| Hudson | 12 | 26 | 21.7 |
| Bergen | 70 | 38 | 5.4 |
| Middlesex | 25 | 17 | 6.8 |
| Monmouth | 53 | 14 | 2.6 |
| Union | 21 | 11 | 5.2 |
| Camden | 37 | 9 | 2.4 |
| Passaic | 16 | 6 | 3.8 |
| Sussex | 24 | 4 | 1.7 |
| Salem | 15 | 1 | 0.7 |
Invented data for this showcasecenterforcooperativemedia.org
Invented values, taken in June 2026. The municipality counts are invented as well, so they will not match the state's own list.
The ratio column is the most useful number in the table and the easiest one to misread. Hudson County shows 21.7 jobs for every 10 towns. Bergen County shows 5.4. Hudson looks four times better served.
It is not. Hudson County holds 12 municipalities in 46 square miles. A reporter can attend two council meetings in one evening and drive home. Bergen County holds 70 municipalities across 234 square miles, and no evening holds two of those meetings.
So the ratio measures towns, and towns are not equal. Three things change what one job can cover, and none of them appear in the table.
A county that streams every meeting gives each reporting job more reach than a county that does not. In the invented data, 9 of the 21 counties stream every municipal meeting, and those 9 counties account for 62 percent of the invented meeting stories filed over the 12 months to June 2026.1
That is the finding under the finding. Capacity is a count of people, but coverage is a count of people multiplied by what the county makes easy. A county cannot hire reporters. It can post an agenda 72 hours ahead, and in the invented data that change moves the number as much as one more job does.
None of this argues against counting jobs. It argues for reading the count beside the conditions the job works in. A coverage gap in a county with good records is a choice about priorities. The same gap in a county that posts minutes as scanned images six weeks late is something else.
I can cover the council or I can cover the school board. I cannot cover both, so I pick the one where the money moves this month.
Dana Whitfield, invented editor of an invented weekly in Salem County
In a county with 1 reporting job and 15 towns, the reporter chooses. The choice is rarely made on news value. It is made on distance, on meeting time and on which clerk answers email.
The invented editors in this showcase described the same three habits. They cover the towns nearest the office. They cover the meetings that start before 8 in the evening. They cover the agencies that publish an agenda more than 48 hours ahead.
That last habit matters, because it decides which public bodies get watched. A zoning board that posts its agenda late is a zoning board nobody sits in on. It is not hiding. It is just harder to reach than the one that posts on time.
Meeting time decides more than it should. In the invented data, 71 percent of municipal meetings in the state start at 7 in the evening or later, and a reporter who covers two towns on the same night covers one of them from the recording.
The 8 counties with a shared reporting desk are the clearest case in the invented data.1 A shared desk is one paid job that several newsrooms fund together. The desk covers the county body that no single newsroom can justify covering alone.
The gain is real in the invented numbers and small in absolute terms. A 34 percent gain on a base of 62 municipal meeting stories a year is 21 more stories. That is 21 more meetings with somebody in the room, over 12 months, across 8 counties.
Pick the public body in your county that nobody covers. Count how many meetings it held in the last 12 months. Count how many stories any outlet published about it in the same period. If the second number is under 3, you have found the gap that a shared desk is built for.
Then find the two or three newsrooms that would use the same coverage. A shared desk needs a funder, an editor of record and a written agreement on who publishes first. The Center for Cooperative Media keeps model agreements from real collaborations.
A shared desk is not the only answer, and it is not free. In the invented data, a desk costs one newsroom about 9,000 dollars a year over three years, and the newsrooms that dropped out did so in year two, when the founding grant ended. Ask who pays in year four before you ask who covers the county board.
Measured in meeting stories a year.
Two of the four counties published more meeting stories in 2026 than in 2021, and two published fewer, on the invented count for each full calendar year.
Invented data for this showcasecenterforcooperativemedia.org
Invented values. A meeting story is one published item about one meeting of one public body.
| Category | 2021 | 2026 |
|---|---|---|
| Essex | 310 | 356 |
| Bergen | 288 | 241 |
| Sussex | 74 | 98 |
| Salem | 51 | 29 |
An invented editor in Salem County asks how many paid reporters cover her county. Nobody in the network can answer, so the count starts.
The team agrees what counts as a reporting job. The rule of 20 hours a week settles the hardest cases and drops about a third of the names on the first list.
Hudson County takes 11 days to count and produces 26 jobs. The team decides that every county needs a phone call, not only a web search.
Each newsroom in the invented count receives its row and 14 days to correct it. Nine newsrooms send changes, and 7 of the 9 changes are accepted.
The final invented count holds 210 paid reporting jobs across 21 counties. The team freezes the data and starts the write-up.
coverage gap
A public body that holds meetings and receives no published coverage over the period counted.
Open the glossary entrymunicipal meeting story
One published item about one meeting of one public body, such as a council, a school board or a zoning board.
Open the glossary entrynews desert
A place with no local news outlet that reports on its own government. This report counts jobs instead, because an outlet can exist with nobody in it.
Open the glossary entryreporting capacity
The paid people available to produce local news for a place, counted as jobs and set against the number of towns and public bodies they must cover.
Open the glossary entryshared reporting desk
One paid reporting job that several newsrooms fund together. The desk files the same work to every newsroom that pays for it.
Open the glossary entryThe invented count ran from March 2025 to June 2026. Two researchers built a list of outlets in each of the 21 counties, called each outlet, and counted paid reporting jobs of 20 hours a week or more. Each newsroom received its own row to correct before the count closed.
No such count exists. This page is a design showcase for a report template, and every number, name, quotation and result on it is invented. The links below go to real organizations that do publish work on local news capacity. None of them produced the data on this page.
A real version of this count would name its period, its sample, its refusal rate and its coder agreement. It would say how many outlets did not answer the phone, and it would publish the coding rules as a file a reader can check. The two files under this note are placeholder addresses for the showcase, and nothing is published at either one.
If the count were real, its limits would be worth stating here: a phone survey misses outlets that do not answer, a 20-hour rule drops working freelancers, and a job count says nothing about what the job produces. A methods section is where a reader learns what the report cannot support.
An earlier version of the county table gave Warren County 5 reporting jobs. The invented figure is 4. The chart and the total are unchanged.
The methods note said the count period ended in May 2026. It ended in June 2026.