The finding
Contents restoration is a separate job from drying the building, and the claims file lets it be sized. Across 230,560 National Flood Insurance Program claims since 2015 that carry both a building damage amount of at least $1,000 and a contents damage amount, the median contents figure is 28% of the building figure, or $14,816 against $60,366. That ratio varies by a factor of 2.6 between Texas at 35% and New York at 14%.
Headline figures
In this report
- The ratio, nationally
- State by state
- The coverage caveat that has to come first
- Contents settle differently from buildings
- Does contents damage scale with water depth
- By building type
- Data quality and the cleaning applied
- Method
- Reproduce this
- What this does not show
- Questions this report answers
- Sources
Key findings
- The median contents-to-building damage ratio across 230,560 paired NFIP claims since 2015 is 28.3%. The mean is 47.3%, far higher because a minority of claims record contents damage exceeding building damage.
- Texas records the highest median ratio at 35.5% against 13.9% in New York, a 2.6 times spread across 28 states with at least 500 paired NFIP claims.
- Only 46.9% of NFIP claims with building damage of at least $1,000 record any contents damage. Contents cover is bought separately, so this is partly a measure of what was purchased rather than what got wet. It ranges from 64.7% in Texas to 21.7% in New Hampshire.
- Contents claims settle at a median 85.1% of assessed damage against 95.8% on building claims, a gap of 10.7 percentage points.
- 4.2% of assessed contents claims settled at zero against 1.6% of building claims (FEMA NFIP, 2015 to 2026).
- The contents ratio rises from 22.8% below one foot to a peak of 34.0% at 2 feet, then falls to 26.8% at 10 feet. Contents damage does not keep pace with building damage in deep water.
- By building type the highest median ratio is 31.7% on single family residence against 5.6% on residential unit in a building of 5 or more units (FEMA NFIP, codes with at least 200 paired claims).
The ratio, nationally
The NFIP claims file carries an assessed building damage amount and, separately, an assessed contents damage amount. Dividing one by the other gives a figure that is not published anywhere: how much of a flood loss sits in the contents rather than the structure. Across 230,560 claims since 2015 where both are present and the building figure is at least $1,000, the median is 28.3% and the mean 47.3%.
The distance between those two numbers is itself a finding. The distribution has a long right tail: most claims put roughly a quarter to a third of the assessed loss in contents, but a minority put more in contents than in the building. The median paired claim records $14,816 of contents damage against $60,366 of building damage. Contents restoration carries its own labour profile of inventory, pack-out, transport, off-site cleaning and storage, none of which appears in the claims file as a line item.
State by state
| State | Paired claims | Median ratio | Mean ratio | Share of building claims with contents recorded |
|---|---|---|---|---|
| Texas | 77,600 | 35.5% | 49.7% | 64.7% |
| Louisiana | 27,126 | 30.1% | 51.0% | 45.5% |
| Oklahoma | 682 | 30.0% | 62.6% | 38.3% |
| West Virginia | 881 | 27.4% | 49.8% | 32.5% |
| Alabama | 1,980 | 26.8% | 49.5% | 42.1% |
| Kentucky | 1,992 | 26.5% | 50.1% | 33.9% |
| Georgia | 2,092 | 26.4% | 52.2% | 36.1% |
| Virginia | 1,759 | 25.4% | 44.4% | 36.2% |
| Pennsylvania | 2,491 | 25.4% | 60.1% | 34.9% |
| Mississippi | 1,801 | 25.2% | 66.2% | 39.0% |
| Florida | 72,535 | 25.0% | 43.4% | 47.3% |
| South Carolina | 5,836 | 24.4% | 48.2% | 34.2% |
| Missouri | 2,200 | 23.4% | 75.1% | 34.6% |
| Arkansas | 783 | 23.1% | 34.1% | 35.1% |
| North Carolina | 10,523 | 23.0% | 40.9% | 37.5% |
| Ohio | 777 | 22.6% | 47.2% | 29.6% |
| Indiana | 638 | 22.5% | 46.3% | 29.3% |
| Tennessee | 1,113 | 22.0% | 47.8% | 33.0% |
| Iowa | 505 | 20.7% | 42.5% | 27.6% |
| Illinois | 1,107 | 19.2% | 39.2% | 24.7% |
| California | 1,383 | 18.9% | 47.0% | 27.7% |
| Wisconsin | 547 | 18.8% | 47.9% | 30.2% |
| New Jersey | 4,013 | 18.8% | 42.8% | 33.6% |
| Michigan | 723 | 18.8% | 39.3% | 26.5% |
| Washington | 765 | 17.2% | 63.4% | 28.8% |
| Massachusetts | 856 | 16.5% | 30.0% | 27.0% |
| Maryland | 742 | 13.9% | 36.2% | 33.7% |
| New York | 3,214 | 13.9% | 32.7% | 33.8% |
Texas and Louisiana lead at 35.5% and 30.1%, and New York sits last at 13.9%. The spread is real, but any explanation is interpretation rather than measurement: the file carries no field for building construction, basement presence or what was stored where.
Three mechanisms are plausible. Building stock differs: a slab-on-grade house with a low floor level puts more of its contents into the first foot of water than a raised one does. Basement prevalence differs, and a finished basement holds the goods easiest to write off entirely. And what people keep at floor level differs by climate. None can be tested from this dataset.
A fourth mechanism is about insurance rather than damage, and it matters most. In Texas 64.7% of building claims also record contents damage, against 21.7% in New Hampshire. States where contents cover is more widely bought show a different mix of claims, not just a different mix of floods.
The coverage caveat that has to come first
Under the National Flood Insurance Program, building cover and contents cover are separate purchases. Per FEMA FloodSmart, NFIP types of coverage, a policyholder can buy either alone or both, with separate limits. A claim recording no contents damage may mean the contents were dry, or that no contents cover was in force to assess them against. Across the whole file 46.9% of claims with building damage of at least $1,000 record any contents damage, and that figure blends two different things this report cannot separate. Treating it as a damage rate would be wrong: the honest reading is that it is partly a coverage take-up measure.
| State | Building claims | With contents recorded | Share with contents |
|---|---|---|---|
| Texas | 119,991 | 77,600 | 64.7% |
| Florida | 153,321 | 72,535 | 47.3% |
| Puerto Rico | 724 | 339 | 46.8% |
| Louisiana | 59,681 | 27,126 | 45.5% |
| Alabama | 4,703 | 1,980 | 42.1% |
| Mississippi | 4,621 | 1,801 | 39.0% |
| South Dakota | 500 | 194 | 38.8% |
| Oklahoma | 1,783 | 682 | 38.3% |
| North Carolina | 28,055 | 10,523 | 37.5% |
| Virginia | 4,855 | 1,759 | 36.2% |
| Georgia | 5,803 | 2,092 | 36.1% |
| Arkansas | 2,230 | 783 | 35.1% |
| Pennsylvania | 7,140 | 2,491 | 34.9% |
| Missouri | 6,360 | 2,200 | 34.6% |
| South Carolina | 17,053 | 5,836 | 34.2% |
| Kentucky | 5,872 | 1,992 | 33.9% |
| New York | 9,521 | 3,214 | 33.8% |
| Maryland | 2,202 | 742 | 33.7% |
| New Jersey | 11,943 | 4,013 | 33.6% |
| Tennessee | 3,372 | 1,113 | 33.0% |
| West Virginia | 2,707 | 881 | 32.5% |
| Maine | 1,070 | 348 | 32.5% |
| Minnesota | 586 | 184 | 31.4% |
| Connecticut | 1,600 | 492 | 30.8% |
| Wisconsin | 1,814 | 547 | 30.2% |
| Kansas | 883 | 263 | 29.8% |
| Ohio | 2,625 | 777 | 29.6% |
| Indiana | 2,180 | 638 | 29.3% |
| Washington | 2,656 | 765 | 28.8% |
| Hawaii | 1,196 | 341 | 28.5% |
| California | 4,985 | 1,383 | 27.7% |
| Iowa | 1,833 | 505 | 27.6% |
| Oregon | 554 | 150 | 27.1% |
| Massachusetts | 3,167 | 856 | 27.0% |
| Michigan | 2,729 | 723 | 26.5% |
| Illinois | 4,479 | 1,107 | 24.7% |
| Vermont | 929 | 210 | 22.6% |
| Nebraska | 1,418 | 318 | 22.4% |
| New Hampshire | 584 | 127 | 21.7% |
The national ratio of 28.3% therefore describes claims where contents were assessed. Our report on the insured flood loss gap covers what the programme does and does not reach.
Contents settle differently from buildings
Dividing the settled amount by the assessed damage on each side of the claim gives a payout ratio. Across 479,868 building claims it is 95.8%, and across 224,430 contents claims on the same footing 85.1%, a gap of 10.7 percentage points.
The gap is structural rather than discretionary. Per FEMA FloodSmart, what flood insurance covers, personal property under an NFIP policy settles on an actual cash value basis, so depreciation applies, while building property on an eligible single-family policy can settle at replacement cost. Two claims assessed at the same figure settle at different percentages depending on which side of the policy they sit, and the 10.7 point gap is what that looks like across the file.
The zero-settlement rows carry the same reading: 4.2% of assessed contents claims settled at nothing against 1.6% of building claims. Deductibles apply separately to each side of an NFIP policy, so a contents loss that is small relative to its own deductible can settle at zero while the building loss on the same claim settles in full. A pack-out is committed early, usually before the settlement split is known.
Does contents damage scale with water depth
Building damage climbs steeply through the first two feet of water and then flattens, which is the subject of our water depth damage curve. Whether contents damage follows the same shape is a separate question, and the answer is only partly yes.
| Water depth | Paired claims | Median contents ratio | Median building damage | Median contents damage |
|---|---|---|---|---|
| 0 ft | 44,202 | 22.8% | $21,137 | $8,204 |
| 1 ft | 75,267 | 28.1% | $42,511 | $14,503 |
| 2 ft | 32,790 | 34.0% | $63,926 | $25,284 |
| 3 ft | 16,267 | 33.8% | $69,666 | $28,051 |
| 4 ft | 9,134 | 32.4% | $73,370 | $27,341 |
| 5 ft | 5,124 | 31.2% | $70,190 | $25,264 |
| 6 ft | 3,496 | 30.4% | $55,752 | $21,602 |
| 7 ft | 2,329 | 28.9% | $52,701 | $19,351 |
| 8 ft | 2,211 | 25.7% | $50,633 | $17,445 |
| 9 ft | 1,399 | 27.3% | $56,354 | $18,409 |
| 10 ft | 1,315 | 26.8% | $58,115 | $18,378 |
The ratio climbs from 22.8% below one foot to 34.0% at 2 feet, then declines to 26.8% at 10 feet. Contents and building damage rise together through the shallow band, then contents damage stops keeping up: median contents damage moves from $25,284 at two feet to $18,378 at 10 feet, while building damage moves from $63,926 to $58,115.
A plausible reading is that contents damage saturates. Once water reaches two or three feet, nearly everything at floor level is already a total loss, and additional depth reaches wall area rather than new possessions, while building damage keeps accumulating through wall height, wiring and mechanical systems. This is an association in assessed amounts, not a physical measurement.
Shallow losses of one to three feet are therefore where contents work is the largest share of the job, and they are also the most numerous: 168,526 of the 193,534 paired claims in the clean depth band sit at three feet or less.
By building type
| Occupancy code | Paired claims | Median contents ratio | Share recording contents |
|---|---|---|---|
| 11: Single family residence, Risk Rating 2.0 | 133,677 | 27.6% | 51.6% |
| 1: Single family residence | 74,623 | 31.7% | 50.2% |
| 18: Non-residential building, Risk Rating 2.0 | 7,375 | 25.8% | 34.5% |
| 6: Non-residential business | 3,537 | 31.2% | 36.3% |
| 14: Residential manufactured or mobile home | 3,283 | 23.8% | 45.2% |
| 12: Residential unit in a 2 to 4 unit building | 2,942 | 13.7% | 26.6% |
| 15: Residential condominium building | 1,350 | 10.9% | 8.7% |
| 2: Residential building, 2 to 4 units | 996 | 16.1% | 19.7% |
| 16: Residential building, 2 to 4 units, Risk Rating 2.0 | 936 | 24.2% | 48.3% |
| 4: Non-residential building | 725 | 26.9% | 31.4% |
| 13: Residential unit in a building of 5 or more units | 601 | 5.6% | 11.8% |
| 3: Residential building, more than 4 units | 504 | 15.1% | 12.2% |
The spread by building type is wide: 31.7% on single family residence against 5.6% on residential unit in a building of 5 or more units. The final column matters as much as the third, because the share recording contents at all ranges from 8.7% to 51.6% across these codes. Where it is low the ratio runs on a small and probably unrepresentative subset.
Unit-level codes, where the policy covers one dwelling inside a larger building, are not comparable with whole-building codes here: a unit owner’s policy covers interior finishes and personal property while the shell sits under a separate association policy. Our report on flood loss by building type compares building damage across the same occupancy codes.
Data quality and the cleaning applied
Two fields in this dataset need treatment before any ratio built on them means anything. Both treatments are applied to every figure above and stated here so the numbers can be checked.
| Issue | Claims affected | How we handled it |
|---|---|---|
| Building damage below $1,000 | 42,007 | Excluded from every ratio. Claims recording a single dollar of building damage against five-figure contents damage produce ratios in the thousands and destroy both the median and the mean. |
| No contents damage amount recorded | 306,386 | Excluded from the ratio, counted in the take-up measure. Contents cover is bought separately, so an absent value is not a zero. |
| Depth recorded as exactly 99 feet | 13,551 | Excluded from the depth table. This is a sentinel value, not a measurement. |
| Negative depth | 17,085 | Excluded from the depth table. The dataset uses negative values for water below the first floor level. |
| Depth above 20 feet | 30,953 | Excluded from the depth table. Recorded depths run as high as several hundred feet, which is not plausible for the buildings in this file. |
| No depth recorded | 76,981 | Excluded from the depth table, retained everywhere else. |
234,405 of the 626,886 retrieved claims carry a non-zero value in both damage fields. Applying the $1,000 building floor leaves 230,560 paired claims, the sample behind every ratio here. The depth table runs on the subset that also carries a plausible depth.
| Loss year | Paired claims in this sample |
|---|---|
| 2015 | 9,979 |
| 2016 | 30,644 |
| 2017 | 67,054 |
| 2018 | 13,209 |
| 2019 | 12,205 |
| 2020 | 10,629 |
| 2021 | 15,716 |
| 2022 | 19,991 |
| 2023 | 7,008 |
| 2024 | 38,816 |
| 2025 | 4,181 |
| 2026 | 1,128 |
Method
How we counted. We retrieved 626,886 NFIP redacted claim records with a loss year of 2015 or later from the FEMA OpenFEMA API on 4 October 2026. For every ratio we kept claims with a building damage amount of at least $1,000, which removes rows where a placeholder figure would dominate the division, and a recorded contents damage amount, leaving 230,560 paired claims. The ratio is contents damage divided by building damage, reported as a median per group with the mean alongside because the distribution is right skewed. State medians need at least 500 paired claims and every other grouping at least 200. Payout ratios are the settled amount over the assessed damage on the same side of the claim, with the $1,000 floor applied to each side independently. The depth table uses only depths between 0 and 10 feet. No inflation adjustment is applied, so figures across years are nominal.
Reproduce this
Reproduce this
- Endpoint
https://www.fema.gov/api/open/v3/NfipClaims- Query
$filter=yearOfLoss ge 2015- Rows retrieved
- 626,886 claims
- Retrieved
- 4 October 2026
The API pages at 5,000 rows via $top and $skip. No API key is required. The fields used here are buildingDamageAmount, contentsDamageAmount, amountPaidOnBuildingClaim, amountPaidOnContentsClaim, waterDepth, occupancyType and state. Re-running this query will return a larger set as newer claims are added.
What this does not show
What this does not show
- Contents cover is a separate purchase, so the take-up figure is not a damage rate. A policyholder can hold building cover without contents cover, so the 46.9% of building claims recording contents damage blends coverage take-up with actual contents damage, and this dataset cannot separate the two.
- These are insured flood losses only. NFIP covers policyholders. Uninsured contents losses, and every water loss that is not a flood, including burst pipes, appliance failures and sewer backups outside a flood event, are absent from this dataset entirely.
- Damage amount is not restoration cost. The fields record assessed damage on the claim, not what a contractor billed, and no field records pack-out, cleaning or storage work at all.
- The state spread is described, not explained. Building stock, basement prevalence and storage habits are plausible contributors and none is a field in this file.
- Contents and buildings settle on different bases. Personal property settles at actual cash value and eligible building property can settle at replacement cost, so the payout gap is structural. It is not evidence about how any claim was handled.
- No inflation adjustment. Figures spanning 2015 to 2026 are nominal dollars.
- 2025 and 2026 are incomplete. Claims are filed and settled after the loss, so the most recent loss years hold fewer claims than they eventually will.
Most quotable figures
- Median contents damage on an NFIP flood claim is 28.3% of the building damage, or $14,816 against $60,366 (FEMA OpenFEMA NFIP claims, 2015 to 2026, n = 230,560).
- The contents share of a flood loss ranges from 35.5% in Texas to 13.9% in New York across 28 states (FEMA NFIP claims since 2015).
- Contents claims settle at a median 85.1% of assessed damage against 95.8% on building claims, because NFIP personal property settles at actual cash value while eligible building property can settle at replacement cost.
- The contents share of assessed flood damage peaks at 34.0% at 2 feet of water and falls to 26.8% at 10 feet: contents damage saturates while building damage keeps accumulating.
Questions this report answers
What share of a flood loss is contents rather than the building?
In NFIP claims since 2015, the median contents damage amount is 28.3% of the building damage amount across 230,560 paired claims. The mean is 47.3%.
Which states have the highest contents share?
Texas at 35.5% and Louisiana at 30.1% lead the 28 qualifying states in FEMA NFIP claims since 2015. New York is lowest at 13.9%.
Why do only about half of flood claims record contents damage?
Because contents cover is bought separately under the National Flood Insurance Program. 46.9% of claims with building damage of at least $1,000 record any contents damage, reflecting both what was purchased and what got wet.
Do contents claims pay out at the same rate as building claims?
No. The median contents payout ratio is 85.1% of assessed damage against 95.8% on building claims. Per FEMA FloodSmart, NFIP personal property settles at actual cash value while eligible building property can settle at replacement cost, so the gap is structural.
Does contents damage increase with water depth?
Up to a point. The contents share rises from 22.8% below one foot to 34.0% at 2 feet, then falls to 26.8% at 10 feet. Contents damage saturates while building damage keeps rising.
Does this dataset measure pack-out and off-site cleaning work?
No. The file records an assessed contents damage amount and a settled amount, with no field for inventory, pack-out, transport, off-site cleaning or storage. These figures size the loss, not the work.
Can I reuse this data?
Yes. The tables and derived figures are released under CC BY 4.0 with attribution to Remediation Insider. The underlying claims data is public and available from FEMA.
How to cite
Remediation Insider. The Contents Restoration Report: What Is Inside the Building. https://remediationinsider.com/contents-restoration-report/ (retrieved 4 October 2026).
Data and tables in this report are released under CC BY 4.0. Attribution: Remediation Insider.
Sources
- Federal Emergency Management Agency. FEMA OpenFEMA NFIP Redacted Claims, v3. Claims with loss year 2015 or later. 626,886 records retrieved 4 October 2026.
- FEMA FloodSmart, NFIP types of coverage. Used for the separate building and contents cover structure under the National Flood Insurance Program.
- FEMA FloodSmart, what flood insurance covers. Used for the actual cash value and replacement cost settlement bases.
- Occupancy code labels are taken from the FEMA OpenFEMA NfipClaims data dictionary. Codes absent from that dictionary are reported as unmapped.
- FEMA OpenFEMA Terms and Conditions.
- No survey, interview or physical inspection was conducted for this report. Every figure is computed from the retrieved claims file.
This product uses the Federal Emergency Management Agency’s OpenFEMA API, but is not endorsed by FEMA. The Federal Government or FEMA cannot vouch for the data or analyses derived from these data after the data have been retrieved from the Agency’s website(s).