Los Angeles Bathroom Remodel Permits: What City Records Show About Inspections and Final Sign-Off

Executive summary

This study looks at what happens after a bathroom-remodel permit is issued in the City of Los Angeles, using only the City's own records. A documented, audited classifier identified 21,798 bathroom-remodel records in a frozen snapshot of the LADBS permit dataset. We joined them to LADBS inspection events dated on or before October 4, 2026.

Three lessons run through the findings.

Age comes before everything. Young records mostly read "Issued" and mature records mostly read "Permit Finaled." The mature share barely moves whether "mature" starts at 12, 18 or 24 months.

Scope is visible in the record. In the frozen descriptive data, mixed or broader permits (Tier B) show a longer permit-record lifecycle summary in the frozen descriptive data and a lower mature "Permit Finaled" share than bathroom-only permits (Tier A). The records show the difference but do not explain it.

The record resists simple verdicts. First inspection results include scheduling, cancellation, readiness and partial-approval labels. "Corrections Issued" appears often. Inspection journeys form a long tail, not a single standard path.

Taken together, the City record is a strong starting point for questions about a bathroom remodel. On its own, it is not a final answer. The due-diligence section below turns that into a five-step checklist for owners and buyers.

How to read this study

Six terms carry most of the meaning.

• Permit record. One row in the LADBS Building Permits Issued dataset (pi9x-tg5x), identified by its permit number. We describe what the record says, not what happened on site.

• Raw label. A value exactly as the City wrote it, such as "Issued," "Permit Finaled," "Approved" or "PLUMBING-Rough." We never rename them.

• Matched permit. A permit with at least one inspection row in the LADBS Building and Safety Inspections dataset (9w5z-rg2h) on or before the cutoff, joined by permit number.

• Permit/type pair. One permit combined with one inspection type. A permit with several Shower Pan rows is still one Shower Pan pair.

• Mature record. A record old enough at the cutoff that its status says less about how recently it was issued. The primary threshold is 24 months.

• Permit-record lifecycle. The calendar days from a permit's issue date to the earliest inspection row whose result is exactly "Permit Finaled." It describes paperwork, not construction.

Every figure below is a study-derived statistic: the labels are the City's, the counting is ours.

Permit status by record age

A permit record's raw status varies sharply with its age. Among records 0–6 months old, 29.3% carried "Permit Finaled" and 69.6% carried "Issued." Among records more than 24 months old, 72.3% carried "Permit Finaled" and 19.5% carried "Issued." Recent and mature records should never be compared as if they had had the same time to progress.

Shares in each row do not add up to 100%. Other raw status labels exist and are not itemized in this release. The published summary covers the youngest and oldest age bands; the intermediate bands are not part of it.

Source note: City of Los Angeles / LADBS, Building Permits Issued (resource pi9x-tg5x), frozen snapshot. Record age is measured at the 2026-10-04 cutoff. Status is the raw label in the permit record. The study design reads it from the permit dataset's status field and uses inspection-side permit status only as a reconciliation check. Counts and shares are study-derived.

Read this correctly: These shares are a snapshot of labels on one date. They are not the probability that a permit will be finaled, a grade for the work, or evidence of why any permit is open.

Why age comes first.

A permit issued a few weeks before the cutoff has had little time to collect inspections, so its status says more about its age than about the project. This is right-censoring: for recent records, the outcome is not yet observable. That is why recent and mature records are reported separately.

Tier A and Tier B among mature records.

Source note: as above; Tier assignment is study-derived (classifier v0.4).

In the mature cohort, Tier A records carried "Permit Finaled" more often than Tier B records. The data records that gap but cannot explain it, so this study offers no cause.

A robustness check: where "mature" begins.

Any line between "recent" and "mature" is a choice. To test whether that choice drives the result, we recomputed the raw "Permit Finaled" share at three thresholds:

The share moves only within a narrow range (71.2% to 72.3%), so the main conclusion does not hinge on where the line is drawn.

Source note: City of Los Angeles / LADBS, pi9x-tg5x, frozen snapshot; study-derived sensitivity analysis.

Which inspection types appear most often

Among the 19,223 permits with matched inspection data, "Final" was the most common inspection type, recorded for 15,647 permits (81.4%). Next came Interior/Exterior Lathing (63.3%), PLUMBING-Rough (59.9%) and ELECTRICAL-Rough (53.6%). Drywall Nailing (46.5%), Shower Pan (45.7%) and Smoke Detectors (40.1%) complete the top seven.

Source note: City of Los Angeles / LADBS, Building and Safety Inspections (9w5z-rg2h), events on or before 2026-10-04, joined to the study cohort by permit number. Study-derived: count of matched permits with at least one row of each raw inspection type. Each permit counts once per type, however many rows it has.

Why we count permits, not inspection rows.

One permit can collect several rows of the same inspection type. Ranking by raw row counts would push types that tend to repeat up the list. Counting each permit once per type answers a cleaner question: how many bathroom-remodel permits involved this type of inspection at all?

Read this correctly: A type's frequency here does not mean every bathroom remodel requires it. Even "Final" is absent from some matched records. This list shows what the records contain, not a checklist of what the City mandates for a given project.

The labels are reproduced exactly as the City records them. This study does not define what each inspection examines. For that, consult LADBS directly.

"Corrections Issued" and repeat inspections

At least one raw "Corrections Issued" label appeared for 26.1% of Final pairs, 18.6% of PLUMBING-Rough pairs, 16.4% of ELECTRICAL-Rough pairs and 7.0% of Shower Pan pairs. These are descriptive record patterns. They are not failure rates, defect rates or measures of contractor or homeowner performance.

Source note: City of Los Angeles / LADBS, Building and Safety Inspections (9w5z-rg2h), events on or before 2026-10-04. Study-derived indicator: whether the raw result "Corrections Issued" appears at least once among a matched permit's rows for that inspection type. Not a failure or defect rate.

Read this correctly: "Corrections Issued" is a label on a record. On its own it does not establish defective workmanship, fault or noncompliance by anyone, and these shares must not be used to rank contractors, homeowners or neighborhoods.

First result vs. any point in the record.

The same label can be counted two ways. The first-recorded-result section counted pairs whose first recorded result was "Corrections Issued." This section counts pairs where the label appears at least once. Side by side:

The "at least once" share is higher for all three types; by definition it can never be lower. The approved outputs do not isolate what accounts for the difference, so this study does not attribute it.

Source note: same City inspection source and frozen-cutoff method used in the first-recorded-result and corrections sections.

Why there is no general "repeat inspection rate".

This study does not publish a single repeat-inspection rate. The released derived outputs do not include the complete repeat-count distribution needed to compute and independently check one. Repeat behavior appears only in specific journey patterns, such as a "Final" on one date followed by another "Final" on a later date in the inspection-journeys section.

Buying or owning a home with an open bathroom-remodel permit

An open or non-final status, such as a mature record still reading "Issued," is a reason to verify the current official record. It is not evidence that the project is defective, abandoned or noncompliant. Among records more than 24 months old at the cutoff, 3,009 of 15,426 (19.5%) still carried "Issued" in this frozen snapshot.

What a status can and cannot tell you

Source note: interpretation boundaries used for this study of LADBS datasets pi9x-tg5x and 9w5z-rg2h.

Five-step due-diligence checklist

Use this if you are buying a home, selling one, or checking your own past project.

1. Identify the official permit record. Use the City's current official record, not a copy, screenshot or third-party summary. This study is frozen at October 4, 2026; the live record can differ.
2. Confirm the permit number. Make sure the permit belongs to the property and the work in question. Permit number is also what links a permit to its inspections.
3. Review the listed inspections and their dates. Note the inspection types, raw results and dates. The inspection-frequency section shows which types are most common in bathroom-remodel records, but that is context, not a requirement list: a missing type is a question, not a conclusion.
4. Ask the permit holder or the contractor what remains outstanding. If the record is open or shows "Corrections Issued," ask the person who pulled the permit, or the contractor who did the work, to explain where it stands.
5. When it matters, verify final sign-off with the responsible authority. For a purchase or another decision where an open permit is material, confirm the current status directly with the authority that issued the permit. For the permits in this study, that is LADBS.

Questions worth asking

• Which inspections are still outstanding on this permit, and why?
• Were the "Corrections Issued" items addressed, and is that reflected on the record?
• Is there a reason the record has not moved to "Permit Finaled"?
• Was any part of this work filed under a different, supplemental or revised permit?

This study cannot tell you what the answers will be. It can tell you that 19.5% of the mature records in this frozen cohort still carried "Issued" at the cutoff, and that the record alone does not say why.

Methodology and reproducibility

This section is written for journalists, researchers and anyone checking our work before citing it.

1. Sources.

The permit data is a frozen snapshot. It was captured once, hashed, and not refreshed during the analysis. Inspection events are counted only if dated on or before the 2026-10-04 production cutoff, with the cutoff applied locally after retrieval.

2. Cohort and classifier.

The cohort was built by classifying the permit dataset's work-description text with a rules-based classifier. Its include, exclude and precedence rules were frozen before any headline metric was calculated. Sampled positive and negative classifications were audited row by row before approved corrections entered the frozen rules. When a broad rewrite of the rules over-corrected in one revision, it was rejected. The frozen version, v0.4, keeps an earlier stable baseline plus only explicitly adjudicated corrections, each bound to a specific permit number and work-description text. The result: 9,645 Tier A and 12,153 Tier B records, 21,798 in total.

3. Joining permits to inspections.

Permits were linked to inspection events by permit number, normalized on both sides into one key that removes presentation-only separators and whitespace without truncating any segment. Because the two datasets format permit numbers differently, the join was tested on permits already known to have inspection records before full acquisition began.

Inspection data was acquired in 436 bounded chunks covering exactly the 21,798 cohort permits. Before full acquisition, the run had to rediscover 20 of 20 known-matched "sentinel" permits, which returned 187 inspection rows on or before the cutoff. A preliminary acquisition attempt returned no inspection rows, so the final acquisition workflow required the sentinel check before full retrieval.

4. Deduplication.

The acquisition returned 203,055 inspection rows. Only exact duplicate rows were removed, 6,992 of them, leaving 196,063 rows. Distinct events on the same day were kept.

5. Matched and unmatched permits.

19,223 permits (88.2%) matched at least one inspection row on or before the cutoff. 2,575 did not. Unmatched permits stay in every cohort and status count. They are excluded only from inspection-based measures. "Unmatched" means no matched inspection row in this snapshot, not "no inspections."

6. Status, lifecycle and first-observed logic.

• Status by age: the raw status label in the permit record at the cutoff, grouped by record age. Recent and mature records are kept separate.

• Lifecycle: calendar days from issue date to the earliest inspection date whose raw result is exactly "Permit Finaled." The study validated 12,106 such lifecycle records across all record ages. The published figures use only the mature cohort (at least 24 months old) and report Tier A and Tier B separately. By design, a record whose status reads finaled but has no matched "Permit Finaled" event date stays in the status counts and is left out of the interval calculation.

• First-observed result: the raw result on the earliest inspection date for each permit/type pair. Multiple results on that date are labeled MULTIPLE_SAME_DAY instead of being ordered arbitrarily.

• Frequency, corrections and journeys: permits counted once per inspection type; "Corrections Issued" counted as present at least once per permit/type pair; journeys built by date, with same-day types as unordered baskets.

7. Right-censoring and the mature-cohort choice.

Records issued shortly before the cutoff have not had time to reach a final status, so comparing them directly with older records would mislead. The study therefore separates recent from mature records. It tested mature thresholds of 12, 18 and 24 months: the raw "Permit Finaled" share was 71.2%, 71.9% and 72.3% respectively. It uses 24 months for the published lifecycle figures.

8. Limitations.

• City of Los Angeles only. Do not generalize to Los Angeles County or to independent cities.

• Classifier-defined cohort. "Bathroom remodel" means what the frozen classifier selects from City work descriptions. Some remodels may be worded in ways it misses, and some edge cases may be misclassified despite audit.

• Raw labels and records, not outcomes or construction. No pass/fail mapping has been validated, and lifecycle intervals measure paperwork events.

• Partial itemization. Some distributions show only the most frequent labels, and only the youngest and oldest age bands are published. A general repeat-inspection rate is held.

• No cost, quality or risk measures. Declared valuation is not used as project cost, and no contractor, homeowner or neighborhood rankings are produced.

• Frozen in time. Figures describe the snapshot at the cutoff, not later changes to City records.

9. Provenance note.

This published snapshot does not include an explicit ISO-8601 retrieval start and end timestamp. Source identities, cutoff, file hashes, source metadata and the acquisition ledger are preserved. The inspection dataset's metadata snapshot reports its rows as last updated at 2026-10-05T13:02:54Z, and the analysis was produced on 2026-10-06. Any future refresh should record retrieval timestamps, with time zone, explicitly.

10. Reproducibility register.

The sanitized production package keeps the hashes and row counts of every raw inspection chunk in its manifest and ledger. The raw chunk bodies are retained separately.

How to cite this study

Suggested citation:

Bath Remodeling Los Angeles. Los Angeles Bathroom Remodel Permits: What City Records Show About Inspections and Final Sign-Off. Data: City of Los Angeles Department of Building and Safety, resources pi9x-tg5x and 9w5z-rg2h, data frozen October 4, 2026. https://bathremodelinglosangeles.com/resources/los-angeles-bathroom-remodel-permit-inspection-study/.

Quoting a figure? Copy-ready statements that keep the guard attached:

• "Among City of Los Angeles bathroom-remodel permit records more than 24 months old, 72.3% carried the raw status 'Permit Finaled' and 19.5% still carried 'Issued' at the October 4, 2026 data cutoff. These are status labels, not completion odds."

• "Among mature City of Los Angeles bathroom-remodel permit records (at least 24 months old) with an exact 'Permit Finaled' result, the median permit-record lifecycle, from issue date to that result, was 97 days for bathroom-specific permits and 197 days for mixed-scope permits that include bathroom work. That is a records interval, not construction time."

• "Among 15,647 matched City of Los Angeles bathroom-remodel permits with a Final inspection, 26.1% had at least one 'Corrections Issued' label for that inspection type. This is a record pattern, not a failure or defect rate."

• "Final inspections were recorded for 81.4% of 19,223 matched City of Los Angeles bathroom-remodel permits. Frequency does not mean an inspection type is required for every remodel."

Please keep the scope ("City of Los Angeles"), the cutoff date and the guard wording when you quote.

Related resources

Continue with Bathroom Remodeling Resources, the Bathroom Remodeling FAQ, or compare Shower Remodeling, Tub-to-Shower Conversion and Walk-In Showers.