Two permit files, opposite ends of the cost map, and our installed typical is under both

Published 2026-08-07 · records retrieved 2026-08-07 · our band fixed as of 2026-08 · how we build these numbers

Run your own numbers in the water heater calculator →

Every price record published here has ended the same way: the public file carried a dollar figure and no denominator. These two files have one. Plumbers in East Baton Rouge Parish and in Everett, Washington write the tank into the permit description — forty gallon, fifty gallon, gas or electric — and a declared dollar amount sits on the same row.

The comparison goes against us, and its shape matters more than its size — which is why this page leads with Louisiana. Everett is metropolitan Seattle; had it alone come in high, the honest article would have been about Washington labor rates. East Baton Rouge Parish is not that market, and its median filing is above our typical too. A gap that survives the cheap city is not a regional premium.

1. What these records are

East Baton Rouge Parish, Louisiana — Building Permits, resource 7fq7-8j7r at data.brla.gov. Four fields do the work: projectdescription, projectvalue (the declared dollar figure), issueddate and designation (Residential vs Commercial). Its metadata carries licenseId: PUBLIC_DOMAIN. The portal's site-wide terms restrict use to non-commercial purposes, but that clause opens with “Unless otherwise indicated in the Site” — and this dataset is otherwise indicated. Retrieved 2026-08-07.

City of Everett, Washington — Trakit permits, resource 3w3u-656c at data.everettwa.gov. Fields permitdesc, jobvalue, issueddate. Its license is unverified: both license and licenseId are null, and the city's open-data terms page returns HTTP 403 to an automated request, so we have not read the terms. That is why Everett is second here and not first: everything load-bearing below rests on Baton Rouge, and removing it would not change the verdict.

Both files carry owner, applicant and contractor names, addresses and coordinates; none are selected below, no filing is quoted, and every figure is a group statistic. The queries carry a ceiling date so next month's reader counts our population.

# East Baton Rouge Parish — population
curl -s -G 'https://data.brla.gov/resource/7fq7-8j7r.json' \
  --data-urlencode '$select=count(1)' \
  --data-urlencode "\$where=upper(projectdescription) like '%WATER HEATER%' \
                    AND issueddate < '2026-08-07'"

# => [{"count_1":"3094"}]

# the rows the analysis reads — no owner, applicant, contractor,
# address or coordinate column is requested. Every filter named in
# section 3 is applied locally on this file, not here, so the 120
# gallon-bearing rows and the 54 survivors are both reproducible.
curl -s -G 'https://data.brla.gov/resource/7fq7-8j7r.csv' \
  --data-urlencode '$select=issueddate,projectvalue,designation,projectdescription' \
  --data-urlencode "\$where=upper(projectdescription) like '%WATER HEATER%' \
                    AND issueddate < '2026-08-07'" \
  --data-urlencode '$limit=50000' > br_water_heater.csv

# City of Everett — population
curl -s -G 'https://data.everettwa.gov/resource/3w3u-656c.json' \
  --data-urlencode '$select=count(1)' \
  --data-urlencode "\$where=upper(permitdesc) like '%WATER HEATER%' \
                    AND issueddate < '2026-08-07'"

# => [{"count_1":"7602"}]
# without the ceiling the same filter returns 7,802: exactly 200 of those
# filings carry no issue date at all, and none are dated on or after it.

# filters are applied locally, as above. Note the one figure this file
# cannot reproduce: the 1.9% gallon share in section 3 is measured on
# the file whole, undated rows included, so drop the ceiling to check
# it. Those rows leave the cohort at the 2020 cut either way.
curl -s -G 'https://data.everettwa.gov/resource/3w3u-656c.csv' \
  --data-urlencode '$select=issueddate,jobvalue,permitdesc' \
  --data-urlencode "\$where=upper(permitdesc) like '%WATER HEATER%' \
                    AND issueddate < '2026-08-07'" \
  --data-urlencode '$limit=50000' > everett_water_heater.csv

2. What these records measure — and what they quietly do not

These samples are thin, which would end the comparison if the finding were smaller. The largest group here holds 33 filings and the smallest 7. Percentiles behave badly at that size, so nothing beyond the quartiles is quoted and no mean appears.

They are thin because the denominator exists only where somebody typed it. Capacity appears in 3.9% of Baton Rouge's water-heater filings and 1.9% of Everett's; the rest say a heater was replaced and stop. The usable sample is therefore not a random slice of the work but the subset filed by people who write long descriptions — possibly a few firms with one office habit, and neither file lets us test that.

What the dollar figure contains is also unanswered. Neither set of column metadata says whether the declared value is a contract total or a figure produced for fee assessment, and the descriptions make plain it is often not the tank alone: Baton Rouge filings routinely name pans, riser insulation, gas drip legs, vent caps and collars alongside the appliance. Those are code items our calculator prices as separate lines, so such a median is not comparable to a bare like-for-like swap; section 4 therefore puts our band on the axis twice, bare and loaded.

Only permitted work appears — unpermitted replacements and every Saturday homeowner swap are absent, and those are the cheap end, so some unmeasured part of the gap is that absence rather than our band. And water-heater.json carries no state or metro index, so both cities meet the identical national band. Everett adds defects of its own: a jobvalue column holding obvious under-filings beside a five-figure outlier, so medians are the only statistic it supports, and a date column running back to 1902, which is why its cohort is cut at 2020.

3. The cohort, and what we dropped

In East Baton Rouge Parish, 3,094 filings mention a water heater and 120 — 3.9% — state a capacity in gallons. Of those, 54 survive the filters: the parish's own Residential designation, a positive declared value, a single trade (text also covering a remodel, an addition or other unrelated work is dropped), a single appliance, and a capacity between 28 and 85 gallons. They were issued between 2024-03-01 and 2026-05-11. More than half the capacity-bearing filings did not survive.

In Everett, 7,802 filings mention a water heater — undated rows included — and 148, 1.9%, state a capacity. Requiring a positive job value and dropping filings whose text marks them incorrect, refunded, void or cancelled leaves 143; cutting at 2020-01-01, which also removes the undated rows, leaves 104, the most recent issued 2026-07-16.

Group sizes carry the weight of what follows. Baton Rouge's five largest capacity-and-fuel groups hold 49 of its 54 filings: 20 for 40-gallon gas, 15 for 50-gallon gas, 7 for 50-gallon electric, 4 for 38-gallon electric and 3 for 40-gallon electric. Everett's three largest hold 73 of its 104: 33 for 50-gallon gas, 26 for 50-gallon electric and 14 for 40-gallon gas. Only the three groups present in both cities appear below. And Baton Rouge's capacity-bearing filings all fall after March 2024 — earlier descriptions there are too short to contain one — so no trend over time can be read here.

4. The distribution and our band on the same axis

Our installed total sums two published bands, both named so the arithmetic can be checked. The appliance comes from unitBands; the work from laborBands.tank_like_for_like_swap, $300 / $600 / $1,000. For a 50-gallon gas tank: $640 + $300, $780 + $600, $1,150 + $1,000 — a band of $940 / $1,380 / $2,150.

Five of those six medians sit above the top of our band, not merely its typical. The sixth, Baton Rouge's 50-gallon gas figure, lands $11 under our high — a coincidence, not a pass.

Since these filings routinely bundle code work, the fair comparison adds the lines we itemize separately: permitDisposal.permit $92, permitDisposal.disposal $100 and adders.codeUpgrades.expansion_tank $250 — $442 on the typical, $175 / $751 on the ends.

Loaded the whole way our band does contain every observed median — its 50-gallon gas high reaches $2,901, above Everett's $2,842 — but only at that ceiling, and only if the reader ticks options the calculator leaves untouched. The code-upgrade selector defaults to “none” so a straight-swap quote stays comparable; with that default the page prints a 50-gallon gas typical of $1,572 against Baton Rouge's median of $2,139.

Now the contrast that matters. Everett's median sits 33% above Baton Rouge's on the same job — the spread you would expect between metropolitan Seattle and a Louisiana parish. But Baton Rouge sits 55% above our typical. Our distance from the cheap market is larger than the cheap market's distance from the expensive one. Whatever this is, it is not that we forgot to adjust for Washington.

One claim in the file survives — a record reporting only failures is being edited too. Our data says the step from a 40-gallon tank to a 50-gallon one moves the installed total by roughly 5%, reasoned from a $60–$90 unit-price spread inside one product line and identical labor. The cities bracket it: Baton Rouge's 50-gallon gas median is 3.5% below its 40-gallon median, Everett's 9.5% above. At these sample sizes the sign is not resolvable — the magnitude is.

5. Verdict

6. If you are replacing a water heater in either jurisdiction

The useful thing about these files is not the median — it is the description field, which shows what local installers itemize. Ask each bidder to write the tank's capacity, fuel and model onto the quote: that is the denominator that makes two bids comparable, and the field 96% of Baton Rouge's filings and 98% of Everett's leave out. Ask whether the permit fee and haul-away sit inside the number or arrive later. And ask about the expansion tank, the drip pan, the gas drip leg and the vent connector — the items that recur in the Baton Rouge descriptions. If the last heater went in more than fifteen years ago, assume at least one is coming.

Then check your own jurisdiction: swap the portal domain, resource identifier and column names — no two portals name them alike — and the queries above work on most Socrata permit files. Read the description field before the money column. If capacity is absent there, the file cannot tell you anything about price per gallon no matter how many rows it has.

Frequently asked questions

Does a water heater really cost $2,139 in Baton Rouge?

That is the median declared project value on residential permits in East Baton Rouge Parish that stated a 50-gallon gas tank — 15 filings issued between March 2024 and May 2026, median $2,139. It is not a quote. It bundles whatever else each filing covered, and permit descriptions in that file routinely mention pans, drip legs, vent caps and expansion work alongside the tank. It also only counts jobs somebody pulled a permit for, which excludes unpermitted swaps and every do-it-yourself replacement. Read it as what permitted water-heater work declares in that parish, not as what your plumber will charge.

Why does a cheap market matter more than an expensive one in this comparison?

Because it removes the easy explanation. Everett is metropolitan Seattle; if only Everett had come in above our numbers, the honest reading would be Washington labor rates rather than a problem with our band. East Baton Rouge Parish is a low-cost market, and its median 50-gallon gas installation still sits 55% above our installed typical. The two cities differ from each other by 33% on that job — a plausible labor spread — and each differs from our typical by more than they differ from each other. Our water-heater dataset carries no state or metro index at all, so both cities are being compared against the identical national band.

Is the step from a 40-gallon to a 50-gallon tank really only about 5%?

Our dataset says the installed step is roughly 5%, reasoned from a $60–$90 unit-price spread inside one product line with identical labor. The two permit files bracket that claim rather than breaking it: in East Baton Rouge Parish the 50-gallon gas median sits 3.5% below the 40-gallon median, and in Everett it sits 9.5% above. With 15 and 20 filings in one city and 33 and 14 in the other, the sign of the step is not resolvable from this data. The magnitude claim — a single-digit percentage rather than a tier jump — survives both files.

If permit records run higher than your band, why not simply raise the band to match them?

Because two jurisdictions are not a country, and because the observed figures include work our calculator itemizes separately, so matching them directly would double-count. A correction has to name which component moved and stand on something that supports it. Our like-for-like swap labor band already runs to the top of the range its cited source published, $200–$1,000, which means moving it requires a new labor source rather than an arithmetic edit. What this record establishes is narrower and still decisive: the current installed typical is not reachable by permitted work in either market, and the correction will be dated below this page when it ships.