City of Pensacola and Escambia County Flood Risk and Flood Insurance Study

Transcription

1 City of Pensacola and Escambia County Flood Risk and Flood Insurance Study Preliminary Report 1: Long Hollow and Sanders Beach Tracts Wharton Risk Management and Decision Processes Center November 8, 2016 Introduction This study of flood risk and flood insurance is based in the City of Pensacola and Escambia County, Florida. The goal of the study is to gain a clearer understanding of accurate risk-based flood insurance pricing that addresses the tension between risk-based pricing and affordability. The Pensacola flood risk study is comprised of three primary tasks: (1) risk assessment, (2) risk-based s, and (3) affordability. Task 1, risk assessment, involves accurate flood hazard mapping that is more granular than the 1% annual chance flood risk zones that the Federal Emergency Management Agency (FEMA) delineates. FEMA s approach to flood hazard mapping is problematic because mandatory flood insurance purchases are based primarily on whether a building is within or outside the 1% annual chance flood hazard zones, also known as special flood hazard areas (SFHAs) on National Flood Insurance Program (NFIP) maps. SFHAs are not disaggregated according to flood risks with higher probabilities/frequencies, such as 2% and 4% annual chance events, and all structures within SFHAs are susceptible to a homogenous 1% annual chance flood risk. For example, suppose two homes that were built in the same year have foundations on pilings with the same elevation and are both located in an SFHA. Within this SFHA, the NFIP flood map shows that the base flood elevation (BFE), which is the 1% annual chance flood elevation, is 9 feet. One of these homes is located within 300 feet of a stream and the other home is 500 feet from the same stream, but since they are both the same elevation within a FEMA SFHA with a BFE of 9 feet, they are considered to have the same flood risk and their flood insurance rates are the same. In contrast, if the SFHA were disaggregated according to 2% and 4% annual chance zones, the home closer to the stream would have a higher flood risk and thus a higher insurance. Task 2 of this study involves using more granular flood hazard maps to calculate risk-based insurance s that would be compared with s that are based on current NFIP methodology. Task 3 is the examination of affordability concerns that arise from charging households risk-based insurance s and the development of strategies that address financially burdened households so as to promote risk reduction measures that would lower flood risk-rated insurance s. Pensacola is currently undergoing a digital flood insurance rate map (DFIRM) change, and the preliminary map showing changes in SFHAs is scheduled to be released to the public on December 14, 2016 ( The draft map data were provided to the Wharton Risk Center in April 2015 but there are anticipated changes in the preliminary DFIRMs relative to these maps. We will assess changes in s resulting from map changes and discuss how the newly mapped procedure will impact s once we obtain the preliminary DFIRMS. In the spring of 2017 the flood risk products associated with the map changes should be released (per Jerrick Saquibal, Chief of the Bureau of Hydrology and Engineering of the Northwest Florida Water Management District (NWFWMD)). These flood risk products to be released in spring 2017 include geographic information systems (GIS) grid data representing flood elevations for 10%, 4%, 2%, 1%, and 1

2 0.2% annual chance flood frequencies, the percent annual chance grid, and the percent chance in 30 years. Once we obtain these flood risk grids, analyses of Pensacola single-family homes flood risk will be compared to analysis results of these homes risk from the 1% annual chance flood (Task 1). The flood risk grids will be used to compute annual average expected losses for homes, which will inform calculation of risk-based s that will be compared to NFIP s (Task 2). As a starting point for the Pensacola flood study, we have selected two census tracts to examine the distributions of building values and estimated NFIP s for households. We also present tractlevel census data on income and household costs to assess potential flood insurance affordability concerns. One census tract we call Long Hollow as it is located in the Long Hollow neighborhood within the Pensacola city limits that has a history of flooding due to storm water management challenges. All of the Long Hollow census tract is within the X zone according to the effective DFIRM (X zones are outside SFHAs). We can accurately estimate s for single-family homes within this tract using the NFIP manual because elevation of homes located outside the SFHAs has no impact on flood insurance pricing. The other census tract we call Sanders Beach since it is located in the Sanders Beach area within the Pensacola city limits. Some of the Sanders Beach tract is fronting Pensacola Bay and has X, AE, and VE flood zones within it. Both AE and VE zones are in SFHAs, but flood insurance within VE zones is usually priced much higher than that for AE zones because VE zone properties are exposed to waves of three feet or more in the 1% annual chance flood. In this preliminary report we calculate NFIP insurance s for all single-family homes within Long Hollow and Sanders Beach tracts using currently available data and methodology detailed in the April 2016 NFIP manual. We also examine tract-level economic data for households to assess potential insurance affordability concerns. Data and Methodology The study begins with analyses of currently available data at the level of 2010 census tracts and buildings within the Pensacola city limits. This preliminary report focuses on individual building and property appraisers data for single-family homes within the Pensacola city limits that are relevant for flood risk determination and flood insurance s calculations. The State of North Carolina (NC) is leading the nation in accurate floodplain mapping with Lidar (light detection and ranging) technology, which is the most precise and accurate approach to mapping elevation. NC is also mapping building footprints and first floor elevations of structures with Lidar technology. For this investigation in Escambia County, we have Lidar-based ground elevation data collected in 2006 by the NWFWMD. We lack Lidar-based elevation data on homes in Escambia County, but we have building and other infrastructure data obtained from the Escambia County property appraiser s office, the City of Pensacola, and FEMA. The building data we have collected, mostly from the Escambia County Property Appraiser s (ECPA) office, has important structural information for determining homes flood risk, including year of construction, foundation type, and number of floors. Appendix A lists the data and sources used in this preliminary report. Parcel identification (ID) numbers (REFNO) from the ECPA s 2015 parcel GIS shapefile were used to merge tabular data from ECPA s buildings table, and then the parcel shapefile was populated with occupancy types (single family homes), foundation types, heated square footage, numbers of stories, and year built from the buildings table. ECPA improvements table was merged to the parcel shapefile with the REFNO field to populate building values and homestead exemptions for parcels. A parcel with a 2

3 homestead exemption indicates an owner-occupied home. Once these parcel-based data fields from ECPA data were populated for single-family homes, other data for Pensacola were attributed to the parcel shapefile with a series of geospatial analyses that were implemented in ArcGIS. The geospatial analyses were done as follows: 1. Spatially relate (join) all building footprints within single-family residential parcels to attribute building footprints spatial information to parcel data. 2. Calculate mean ground elevations (in feet) within all building footprints using elevation data from the NWFWMD Lidar-derived digital elevation model (DEM). For slab on-grade homes, mean ground elevation within the building footprint is equal to the finished first floor elevation (FFE). 3. Spatially relate (join) all building footprints to the 2010 census blocks they are located within to attribute parcels and building footprints with socioeconomic data from the 2010 census and American Community Survey (ACS). Truncate block ID numbers from 15 digits to 11 digits to derive the tract ID numbers. 4. Spatially relate (join) building footprints to the NFIP community boundaries they coincide with to determine NFIP community ID number, name, Community Rating System (CRS) class, and CRS discounts. The City of Pensacola is the NFIP community name for this study, and they have a CRS class of 7. Spatially relate (join) building footprints to the effective DFIRM flood zones they intersect, along with static BFEs (where given in the DFIRM flood zones). After the above geospatial analyses were implemented, the Pensacola building footprint shapefile was attributed with several fields shown in Appendix A that are not indicated with TBD (to be determined) in the column entitled Pensacola study data source. The data fields marked as TBD will be populated as the necessary data become available. The preliminary DFIRM data that are expected to be released on December 14, 2016 will be analyzed in conjunction with the building and parcel shapefile for Pensacola in a subsequent report (to be completed in January 2016). This subsequent report will assess the areas with changes in the characterization of SFHAs, including changes in BFEs, and how the newly mapped procedure 1 (enacted in April 2015) may impact insurance s. The newly mapped procedure applies to homes that have been mapped into SFHAs. A home newly mapped into a SFHA will be required to purchase NFIP insurance if the home has a federally-backed mortgage, and this may increase annually until the full-risk rate is realized. Grandfathering and associated subsidized NFIP s will still be effective for pre-firm homes in existing SFHAs, in contrast to policies rated as post-firm. The loss of grandfathering for homeowners in new SFHAs could cause affordability concerns for some of them. Using the most recent version of the NFIP manual, published in April 2016, flood insurance s were calculated for single-family homes within the Long Hollow and Sanders Beach tracts. Flood insurance s calculations are based on several attributes, listed below: Building value (assumed to be the same as building replacement value) 1 The newly mapped procedure factsheet can be viewed at this link: The newly mapped procedure chapter of the April 2016 version of the NFIP manual can be viewed at this link: e139ecc72b1d9564f10f27bc30b8cbb9/10_newly_mapped_508_apr2016.pdf 3

4 o Building values (called Improvements in the ECPA parcel data) are the best measure of building values because market values include both land and building values. o Basic NFIP building coverage is up to $60,000 o Additional NFIP building coverage is over $60,000 up to $190,000; for a maximum total of $250,000 in building coverage Content value (assumed to be 50% of building value, and assumed to be the same as contents replacement value.) o Basic NFIP contents coverage is up to $25,000 o Additional NFIP contents coverage is over $25,000 up to $75,000; for a maximum total of $100,000 for contents coverage Homestead exemption: if not applicable, only contents insurance s are calculated Flood zone: o all buildings in Long Hollow are in X zones o buildings in Sanders Beach tract are within X, AE, and VE zones. The date of home construction with respect to the date of the first NFIP flood map (FIRM) for the home s community is important for calculating NFIP rates because homes are rated differently according to whether they are pre- or post-firm (refer to page RATE 23 in the NFIP manual for calculation methodology). Pre-FIRM homes flood insurance rates are lower due to grandfathering, and the rates do not consider home elevation, only whether the home has a basement or enclosure. Pre- FIRM buildings can be rated with full-risk rates if the is more favorable to the insured, but this requires an elevation certificate for the building. NFIP s for post-firm homes within SFHAs are higher, and the rating tables for X zones indicate the same rates for both pre- and post-firm homes (see pages RATE 2 and RATE 6 of April 2016 NFIP manual). NFIP s for post-firm homes in SFHAs are more complicated to compute, as they require calculating the difference between the homes first floor elevation and the BFE of the zone they are located within. However, DFIRMs do not have BFEs computed for all areas of SFHAs. Furthermore, we do not know the first floor elevation of many homes in Escambia County, but, we know that the first floor elevation of homes with a slab on-grade foundation is the same as the ground elevation given in the Lidar-based digital elevation model. There are very few homes in Pensacola with basements, so the rates for homes with no basement or enclosure were used in all calculations. Current NFIP rates for homes in X zones do not consider house elevation, so foundation types and elevations were not needed to calculate Long Hollow tract s based on the NFIP manual. For homes in AE zones in the Sanders Beach tract, we calculated s for slab on-grade homes by using the Lidar-based mean elevation within the building footprint to obtain the difference between slab foundation elevation and BFE for the coincident flood zones according to the NFIP manual. For homes with other types of foundations and in areas where BFEs are unavailable in the DFIRMs, we used a BFE difference of zero to estimate s. Other data to consider in calculating flood insurance s include CRS discounts: for zone X properties, the CRS discount is 5%; and homes in zones AE and VE get a CRS discount of 15% (since the City of Pensacola is class 7 in the CRS program). However, a goal of this preliminary report is to assess what NFIP s would be for homes in the Long Hollow and Sander Beach tracts, and CRS discounts were not applied because they would be the same for all homes examined in this report. 4

5 Deductibles should also be considered when calculating s, but different deductibles were not considered herein. We used a deductible factor of 1 for all calculations, which equates to a $1,000/$1,000 deductible for building/contents for post-firm structures, and a $2,000/$2,000 deductible for building/contents for pre-firm structures (according to the April 2016 NFIP manual page RATE 18). There are administrative and other types of fees that are added to annual NFIP s, but they were not considered in these analyses. Once the building- and parcel-level data for Long Hollow and Sanders Beach tracts were compiled and NFIP s were calculated for all single-family homes, pertinent variables from several ACS tables were downloaded from the U.S. Census Bureau website. Guidance for selecting the pertinent ACS variables and tables came from Appendix G of the Affordability of National Flood Insurance Program Premiums: Report 2, and the tables are listed in Appendix B of this report. This report includes some of the most important variables for the Long Hollow and Sanders Beach tracts to assess potential affordability concerns, and a discussion of these variables and implications for affordability is included herein. Building and Contents Value and Premiums for Long Hollow Homeowners and Renters The analyses in this report consist primarily of calculating NFIP s for all single-family homes in the Long Hollow and Sanders Beach tracts based on the rating tables provided in the April 2016 version of the NFIP manual, which is the most recent version of the manual as of November 3, Descriptive statistics for building and contents values, and calculated s for owner-occupied homes in the Long Hollow tract are presented in Table 1, and the distributions of these values and calculated s by percentiles for owner-occupied homes appear in Table 2. Descriptive statistics for building values and contents and calculated s for renter-occupied homes appear in Table 3, and the distributions of the values and calculated s for renter-occupied homes are shown in Table 4. For comparison purposes, Table 5 shows the data from the 2012 NFIP policy database for the policies with the Long Hollow tract number. Building and contents values for property in the Long Hollow tract did not exceed the $250,000 and $100,000 NFIP respective limits, so it was not necessary to adjust coverages when calculating s. Table 1. Descriptive statistics for building values and calculated NFIP s for owner-occupied homes in the Long Hollow tract, identified as homes with homestead exemptions according to the 2015 ECPA parcel data with building values over $1,000. N Minimum Maximum Median Mean Std. Deviation building value , ,642 46,686 51,663 30,310 contents value 369 5, ,821 23,343 25,832 15,155 building , contents total ,

6 Table 2. Distribution of building values, contents values, and calculated building, contents, and total s by percentiles for owner-occupied homes in the Long Hollow tract with building values of at least $1,000. All figures are in US dollars. Building value Contents value Building Contents Total Percentiles: owner-occupied homes in the Long Hollow tract ,547 23,043 29,296 34,573 41,843 46,686 51,336 55,848 69,107 81, ,642 6,773 11,522 14,648 17,287 20,922 23,343 25,668 27,924 34,554 40, , , ,037 1,128 1,226 1,280 1,917 Table 3. Descriptive statistics for building values and calculated NFIP s for renter-occupied homes, identified as homes without homestead exemptions according to the 2015 ECPA parcel data. Renters do not pay flood insurance for their buildings, so building s would be 0 and thus total s are equal to contents s in Table 3. N Minimum Maximum Median Mean Std. Deviation building value 407 8, ,698 28,530 35,137 22,610 contents value 407 4, ,849 14,265 17,569 11,305 contents total Table 4. Distribution of building values, contents values, and calculated contents and total s by percentiles for renter-occupied homes. All figures are in US dollars. Building value Contents value Contents Total Percentiles: renter-occupied homes in the Long Hollow tract ,976 15,635 19,988 22,820 25,994 28,530 34,145 41,033 47,425 58, ,556 5,488 7,818 9,994 11,410 12,997 14,265 17,072 20,517 23,713 29,292 58,

7 Table 5. Descriptive statistics for NFIP 2012 policies for total s and buildings and contents coverages for the Long Hollow tract, according to the 2012 NFIP policy database. N Minimum Maximum Median Mean Std. Deviation building coverage , , ,077 55,192 contents coverage 26 20, ,000 45,000 54,931 23,953 total There are few insurance policies (26) for the Long Hollow tract (Table 5) considering there are 776 single-family homes with building values at least $1,000 within this tract 2. However, since the Long Hollow tract is entirely within an X zone (outside of the SFHAs), no NFIP insurance is required. One of these 26 policies has zero for building coverage, so this policy must be for a renter-occupied home, while the other 25 policies must cover owner-occupied homes. Building and Contents Value and Premiums for Sanders Beach Homeowners and Renters Tables 6 through 10 show data for the Sanders Beach tract: Table 6 shows descriptive statistics for building and contents values, and calculated s for owner-occupied homes; Table 7 shows the distributions of values and calculated s for owner-occupied homes by percentiles; Table 8 shows descriptive statistics for building and contents values, and calculated s for renteroccupied homes; Table 9 presents the distributions of values and calculated s for renteroccupied homes by percentile; and Table 10 shows data from the 2012 NFIP policy database for the Sanders Beach tract. Premiums for buildings and contents were calculated with the NFIP coverage limits of $250,000 and $100,000, respectively (in Tables 6 through 9). However, the figures shown in Table 6 for building and contents values show the actual values, which are not capped at the NFIP coverage limits. As stated above, calculations for post-firm properties in AE and VE are most accurate for slab on-grade homes. For homes with foundations other than slab on-grade and in flood zones where the BFE was not given, the differences between first floor elevations and BFEs were assumed to be zero. To calculate s for the VE properties, we used the table on page 1-16 of the Specific Rating Guidelines manual because the properties all had a year of construction after 1981 and were treated as non-elevated buildings. However, non-elevated buildings in VE zones are Submit-For-Rate policies (according to page 1-16 of the Specific Rating Guide and page RATE 11 of the NFIP April 2016 manual), which means they should be rated by an NFIP agent. 2 There is one policy with the Long Hollow tract number that indicates a flood zone of AE, but there are no AE flood zones in Long Hollow. This zone AE policy does not have any building coverage. There are also 7 NFIP policies with building coverage over $250,000 and 8 policies with contents coverage over $100,000 in the NFIP database. These were excluded from Table 5. These policies with coverage over the NFIP limits are not denoted as businesses so might reflect data entry errors. 7

8 Table 6. Building values and calculated NFIP s for owner-occupied homes in the Sanders Beach tract, identified as homes with homestead exemptions according to the 2015 ECPA parcel data with building values over $1,000. N Minimum Maximum Median Mean Std. Deviation building value 322 8, ,531 47,431 76,086 92,742 contents value 322 4, ,766 23,715 38,043 46,371 building , contents , total , Table 7. Distributions of building values, contents values, building and content s and total s by percentiles for owner-occupied homes in the Sanders Beach tract, for buildings with values of at least $1,000. All figures are in US dollars. Building value Contents value Building Contents Total Percentiles: owner-occupied homes in the Sanders Beach tract ,045 22,453 29,515 35,749 42,951 47,431 54,692 66,274 98, , ,082 4,523 11,227 14,757 17,875 21,476 23,715 27,346 33,137 49,013 80, , , ,091 1,264 1,473 2,995 Table 8. Building values and calculated NFIP s for renter-occupied homes in the Sanders Beach tract, identified as homes with homestead exemptions according to the 2015 ECPA parcel data with building values over $1,000. N Minimum Maximum Median Mean Std. Deviation building value 319 7,815 1,014,492 38,530 59,459 92,495 contents value 319 3, ,246 19,265 29,730 46,247 contents , total ,

9 Table 9. Distributions of building values, contents values, building and content s and total s by percentiles for renter-occupied homes in the Sanders Beach tract, for buildings with values of at least $1,000. All figures are in US dollars. Building value Contents value Contents Total Percentiles: renter-occupied homes in the Sanders Beach tract ,161 19,609 24,889 29,606 33,212 38,530 43,430 50,044 67, , ,934 5,081 9,805 12,445 14,803 16,606 19,265 21,715 25,022 33,678 57, , ,161 19,609 24,889 29,606 33,212 38,530 43,430 50,044 67, , ,934 Table 10. NFIP 2012 policy database total and buildings and contents coverages for the Sanders Beach tract (for policies with non-zero values) 3. N Minimum Maximum Median Mean Std. Deviation building coverage , , ,886 89,366 contents coverage ,000 52,500 55,305 37,991 total , Since the Sanders Beach tract has variation in flood zones unlike the Long Hollow tract, Table 11 shows the means and counts of coverages and s by flood zone according to the NFIP 2012 policy data, and Table 12 shows the buildings, contents, and total s means and count by flood zones, that were estimated with the methodology detailed herein. Table 11 shows some flood zones that should have been updated with the 2006 effective DFIRM for Pensacola. For example, there should only be zones AE, VE, and X because these are the only three zones appearing in the effective DFIRM. Premiums calculated in Table 12 reflect much higher s in VE zones than AE zone, which is what one would expect due to amenity value of proximity to water bodies. In contrast, Table 11 shows that the NFIP coverages in the V zones is much lower than the X zone coverages, but no substantiated conclusions should be made with a sample of only 5 policies with building coverage and 6 policies total. 3 Table 10 excludes 98 policies in the Sanders beach tract that had building or contents coverages over the NFIP limits of $250,000 and $100,000, respectively. 9

10 Table 11. Means and counts of buildings and contents coverages and s for the Sanders Beach tract by flood zones, according to the 2012 NFIP policy database. Statistics are computed for the 129 policies with building and contents coverage limits of $250,000 and $100,000 respectively. For building coverages, statistics are computed only for policies with nonzero values. building coverage contents coverage total Mean Count Mean Count Mean Count A07 250, , Flood Zone AE 168, , B n/a n/a 100, C 165, , ,269 5 V10 55, ,212 1 VE 86, , ,750 5 X 140, , Sums Table 12. Means and counts of buildings and contents coverages, and total s by flood zone estimated with parcel data and building footprints flood zone coincidence (according to 2006 effective DFIRM). Statistics for buildings and contents coverages are computed with NFIP limits of $250,000 and $100,000 respectively, and building coverage statistics are calculated only for owner-occupied homes. building coverage contents coverage total Mean Count Mean Count Mean Count Flood zone AE 71, , VE 233, , ,870 4 X 66, , Sums Income and Housing Costs Data for Households in Long Hollow Table 13 shows the median household income for the Long Hollow tract, as well as the distribution of household incomes by ranges according to the 2014 ACS 5-year estimates. Almost 50% of all households in Long Hollow have less than $30,000 annual household income, so this tract could be considered lowincome. Furthermore, about 12% of households have retirement income, and are thus assumed to be on fixed income, and 30% of households receive food stamps from the Supplemental Nutrition Assistance Program (SNAP) because of their low income. The 1.5% of households receiving public assistance and the 7% receiving supplementary social security are also low-income since they receive these benefits. Table 14 shows owner costs by households mortgage status as a percentage of household income. Percentages of homeowners costs for housing are a better indicator of affordability than household income, because affordability depends on both income and costs. The US Department of Housing and Urban Development (HUD) defines cost burdened families as those who pay more than 30% of their 10

11 income on housing ( For households with a mortgage, 30% of those in Long Hollow have housing costs that are 30% or more of their household income, while about 7% of those without a mortgage pay 30% or more of their household income on housing. Consequently, a salient finding of this preliminary study is that at least 37% of households in Long Hollow tract would be cost burdened if they were to pay NFIP s. Although they are not required to pay for NFIP insurance since they are not within an SFHA, this area has experience with flooding due storm water management problems, and flooded homes would certainly pose a cost burden to many Long Hollow households. Table 13. Income data for Long Hollow tract from 2014 ACS 5-year estimates. Data originate from ACS tables listed in Appendix G of Affordability of National Flood Insurance Program Premiums: Report 2. Total households 666 Median household income $30,417 Numbers and proportions of households in each income bracket (annual household income $) Number Proportion <10k , , , , , , , , , , , , , , >200k Numbers and proportions of households with other income Number Proportion Social Security income Supplementary Social Security income Public assistance Retirement income Receiving SNAP Occupied homes Vacant homes n=1 tract 11

12 Table 14. Selected data from the 2014 American Community Survey (ACS) 5-year estimates: owner costs by mortgage status as a percentage of household income for the Long Hollow Census tract. OWNER COSTS AS A PERCENTAGE OF HOUSEHOLD INCOME IN THE Number of PAST 12 MONTHS households proportion Total owner-occupied household units (for which mortgage status is computed) 332 with mortgage to 34.9 pct of income owner costs 35 to 39.9 pct of income to 99.9 pct of income pct or more of income total households with mortgage over 30% of income is owner costs no mortgage to 34.9 pct of income to 39.9 pct of income to 99.9 pct of income pct or more of income total households without mortgage over 30% of income is owner costs n=1 tract. Data originate from ACS year estimates table B Income and Housing Costs Data for Households in Sanders Beach Table 15 shows the median household income for the Sanders Beach tract, and the distribution of household incomes by ranges according to the 2014 ACS 5-year estimates. In contrast to the Long Hollow tract, there are fewer households with less than $30,000 annual household income, but 44% of Sanders Beach households have income less than $30,000, which is a large percentage of the households in this tract. The number and percentage of households receiving retirement income in Sanders Beach (181 and 25%) is more than twice that of Long Hollow (79 and 12%) and a larger percentage are also receiving social security income, so we observe that there are more retired persons and households in Sanders beach than in Long Hollow. Households with retirement income in both Long Hollow and Sanders beach tracts may be financially burdened by flood insurance s since retirement income is a fixed amount. Twenty percent of Sanders Beach households receive SNAP benefits and are thus very likely low-income. Although none of the households in the Sanders Beach tract receive public assistance, all other variables suggest that many households in this tract are lowincome and may be financially burdened by NFIP s, especially since some households are within SFHAs and are required to have coverage if they have a federally insured mortgage. Building values for both owners and renters are generally higher (based on means) in Sanders Beach than in Long Hollow. The homes in Sanders Beach are in AE and VE zones so they have greater flood risks than those in Long Hollow, all of which are in the less hazardous X zone. 12

13 Table 15. Income data for the Sanders Beach tract from 2014 ACS 5-year estimates. Data originate from ACS tables in Appendix G of Affordability of National Flood Insurance Program Premiums: Report 2. Total households 937 Median household income $ 30,750 Numbers and proportions of households in each income bracket (annual household income $) Number Proportion <10k , , , , , , , , , , , , , , >200k Counts and proportions of occupied households with other income Count Proportion Social Security income Supplementary Social Security income Public assistance 0 0 Retirement income Receiving SNAP Occupied homes Vacant homes n=1 tract Table 16 shows data for Sanders Beach that are similar to Table 14 for Long Hollow. It examines homeowners costs as percentages of household income, classified by whether the household resides in a home with or without a mortgage. Thirty-four percent of households with a mortgage have housing costs that are 30% or more of household income, and 13% households without a mortgage have housing costs that are 30% or more of income. However, the counts of households with housing costs at 30% or more of annual income (both with and without a mortgage) are 122 in the Long Hollow tract and 124 in the Sanders Beach tract implying that practically the same number of households would be financially burdened by flood risk-rated insurance s. Although Long Hollow is completely within the X zone and therefore these homes are not required to have flood insurance, residents who have lived there at least since the April 2014 flood might want to purchase flood insurance since this area experienced severe flood damage and has a history of storm 13

14 water management problems. In contrast to Long Hollow, many households in the Sanders Beach are within SFHAs and are required to have flood insurance if they have a federally insured mortgage. The socioeconomic traits within the Sanders beach tract indicate that households close to the waterfront are not affluent and may not be able to afford risk-based flood insurance s. They should be studied further when we examine affordability issues (Task 3). Table 16. Selected data on owner costs by mortgage status as a percentage of household income for the Sanders Beach Census Tract. OWNER COSTS AS A PERCENTAGE OF HOUSEHOLD INCOME IN THE Number of PAST 12 MONTHS households proportion Total owner-occupied household units (for which mortgage status is computed) 265 with mortgage to 34.9 pct of income owner costs 35 to 39.9 pct of income to 99.9 pct of income pct or more of income total households with mortgage over 30% of income is owner costs no mortgage to 34.9 pct of income to 39.9 pct of income to 99.9 pct of income pct or more of income total households without mortgage over 30% of income is owner costs n=1 tract. Data originate from ACS year estimates table B Conclusions and Future Research The following observations and conclusions generated by this preliminary report on the Long Hollow and Sanders Beach tracts are useful for undertaking further flood risk research in Pensacola and Escambia County. Observations related to data for Pensacola 1. Almost all homes in Pensacola do not have a basement so we are assuming that no Pensacola or Escambia County homes have a basement. Calculating the NFIP s for homes without a basement is straightforward and households will not be burdened by the extra costs of insuring basements against flood. 2. Many homes in Pensacola and Escambia County have slab on-grade foundations so the first floor elevations of these homes can be calculated with the available Lidar ground elevation data from the NWFWMD. Google Earth Streetview could be used to estimate first floor elevations for 14

15 homes with elevated foundations. Field visits could be made to take photos to estimate homes first floor elevations from the streets and driveways. 3. We are in the process of obtaining elevation certificates for some single-family homes in Pensacola and Escambia County that can be used to accurately estimate risk-based s for those structures. There are several hundred elevation certificates available for homes in Pensacola Beach, so a sample of those properties elevation data will be examined. Once we obtain elevation certificates, we will analyze trends in elevations by foundation types and year of home construction. These will guide our assumptions regarding home elevations when they are uncertain. 4. There are many variables already available at the building and tract level that are useful for calculating NFIP s for homes and assessing neighborhoods that may be cost-burdened by NFIP insurance. Several of these variables have been included in this report, such as building values from the ECPA and tract-level income and household expenses from ACS datasets. While the ACS provide tract-level data on income and expenses that are important for assessing affordability, we do not know individual household-level incomes and expenses. We could estimate household incomes and expenses using building values and tract-level data, and interpolating household-level income using the number and values of homes within each tract. 5. Community- and neighborhood-level mitigation activities to alleviate storm water management issues, and other activities must be considered for analyses of flood risk mitigation and accurate estimation of insurance rates. We are in the process of reviewing development and floodplain management documents for the County, City, and Pensacola Beach. We will assess the Community Rating System (CRS) information for the County, City, and Pensacola Beach and develop recommendations for improving CRS scores and reducing potential future flood losses. Pensacola Beach recently moved from a CRS class 7 to a class 5, and understanding the reasons for this change is vital to enhancing the resilience of Escambia County and the City of Pensacola to flood risks. 6. To address Task 3 concerning affordability of insurance and linking mitigation with s, it is imperative to consider alternatives to elevating homes. Although home elevation is the only mitigation activity that currently lowers NFIP s for homeowners and renters, it may often not be cost-effective to elevate homes in Pensacola. This is because many Pensacola homes are quite old (several were built before 1950), and many homes are slab on-grade with foundations that were not designed to support the weight of the home off the ground. Furthermore, a vast majority of Pensacola is outside SFHAs and s for elevated and nonelevated homes in X zones have the same rates. Therefore, there is no incentive vis-à-vis NFIP reductions for elevating homes in X zones based on the current NFIP insurance rating process. This preliminary report on the Long Hollow and Sanders Beach tracts is the first for the Pensacola and Escambia County flood risk study. Subsequent reports will employ very similar analyses to calculate NFIP s for other single-family homes within the Pensacola city limits and Escambia County. We could also examine the effects of different deductibles for NFIP insurance policies. We will also assess home characteristics and calculate NFIP s for single-family homes on Pensacola Beach. This analysis will involve many more elevated homes with elevation certificates that are within wave action hazard VE zones. Property values in Pensacola Beach are generally much higher than elsewhere in Escambia County due to the amenity value of the Gulf of Mexico, but this does not 15

16 mean that households will not be financially burdened by paying for risk-based flood insurance s. There are additional data that would add substantially to the analyses that we can currently undertake as explained below. Data Needs 1. We will examine the preliminary DFIRM when it is released on December 14, 2016, to assess changes in SFHAs and households who may be impacted by the newly mapped procedure of the NFIP. 2. We would like to obtain NFIP policy and claims data with addresses to inform our research. Policy data with home addresses are important so we can accurately estimate insurance penetration; and claims data with addresses are important to assess where flood damages occur. Analyses of where policies and claims occur with respect to flood zones would provide insight on insurance penetration and areas at risk of flooding that may occur outside SFHAs. We may also be able to derive house elevation information with NFIP data that includes addresses. 3. We currently do not have NFIP policy data with tract numbers that is more current than It would be interesting to assess how the April 2014 flood event in Pensacola impacted claims and subsequent policy purchases using more recent NFIP policy and claims data. These data may also indicate elevations for buildings in the AE and VE zones that were not specified in the 2012 data. 4. When the flood elevation grids for the 10%, 4%, 2%, and 1% annual chance events are provided in spring of 2017, it will be possible to estimate expected average annual losses for single-family homes in Pensacola using a risk-based approach and depth-damage functions that are packaged with the freely-available Hazus software. We expect to receive grid data representing percent annual chance, and percent chance over 30 years along with the other flood elevation grid data in spring 2017, which will be essential for estimating risk-based rates. The expected average annual losses for homes will be calculated with the depth grids to estimate risk-based s. These risk-based s will be compared to NFIP rates based only on the 1% annual flood. 5. We will contact contractors who have done flood and storm water studies for the City of Pensacola and Pensacola Beach to request any flood elevation grids they have produced for their reports. For example, Atkins prepared a report in January 2015 for the City of Pensacola in which they modeled the April 2014 flood event in the Long Hollow neighborhood 4. This report involves some flood elevation grids that we could request from Atkins. Future Research and Analyses Although having household-level data on incomes and expenses would enable accurate assessments of where affordability of flood insurance will be a concern, we can develop several scenarios as to when households with given incomes and mortgages could qualify for a voucher or tax credit and/or lowinterest loan to mitigate their homes against flood risks and to assist with flood insurance costs if s were risk-based. We could determine how large a voucher and/or low-interest loan they would receive for flood insurance based on these assumptions: 4 This report is available online at 16

17 A household will receive a voucher and/or tax credit and/or low-interest loan if the risk-based flood insurance is greater than 5% of their annual income. A household will receive a voucher and/or tax credit and/or low-interest loan if their housing costs are greater than 30% of their annual income. By calculating NFIP s without subsidies for all single-family homes in the Long Hollow and Sanders Beach tracts, as we have done in this report, we have taken the first steps in undertaking an analysis of affordability. 17