Loading...
HomeMy WebLinkAboutagenda.council.worksession.20170801 CITY COUNCIL WORK SESSION August 01, 2017 4:00 PM, City Council Chambers MEETING AGENDA I. Housing Guideline Changes: use of AMI for income categories P1 Strengthening Community Through Workforce Housing 1 POLICY MEMORANDUM TO: Mayor and City Council FROM: Mike Kosdrosky, Executive Director, APCHA DATE: July 28, 2017 MEETING DATE: August 1, 2017 (Council Work Session) RE: Notice of Call-Up of Amendments to the Aspen/Pitkin County Housing Authority Guidelines concerning changes to Number of Categories, establishing a Single Qualification System based on Area Median Income (AMI), Target Household Income Levels, and Category Income Limits REQUEST OF COUNCIL Review and approve APCHA Resolution No. 2 (Series of 2017), Adopting Amendments to Part II, Sections 5, 6 and 7, and Adding Appendix L to the Aspen/Pitkin Employee Housing Guidelines pertaining to Categories 1-7 Income Limits and Use of Area Median Income (AMI). PREVIOUS COUNCIL ACTION At its regularly scheduled meeting on July 24, 2017, City of Aspen City Council voted to call-up the APCHA Board of Director’s recommended changes to the Workforce Housing Guidelines. BACKGROUND On June 21, 2017, the APCHA Board of Directors approved at Public Hearing APCHA Resolution No. 2 (Series of 2017), which: 1. Eliminates ownership Categories 6 and 7 and consolidates them into Category 5; 2. Expands rental household income categories from four (4) to five (5); 3. Establishes a single qualification system based on Area Median Income (AMI) for determining Maximum Gross Incomes (i.e. Category Income Limits) for Categories 1-5 for both rental and ownership1; 4. Adjusts Target Household Income Levels based on AMI; and 5. Updates Category Income Limits using Target Household Income Levels and AMI Household Size Adjustments (i.e. total persons per household).2 1 See Resolution No. 2 (Series of 2017), Part II, Section 5. 2 See Resolution No. 2 (Series of 2017), Tables I, II, and III. P2 I. Strengthening Community Through Workforce Housing 2 On March 13, 2017, APCHA Policy Memorandum addressed and explored the issues of APCHA’s current dual qualification system. Staff’s preliminary recommendations were: • Change number of categories for rental and ownership programs to 5 plus RO; • Adjust Category Maximum Gross Incomes to AMI target incomes; and • Define and adopt an Affordability Standard in the Housing Guidelines. On April 19, 2017, APCHA staff presented four (4) detailed policy scenarios (options) to Board: • Scenario 1: Keep current methodology (“Do Nothing” Option); • Scenario 2: Port incomes to AMI, but keep Dual Qualification System (4 Rental, 7 Ownership Categories plus RO); • Scenario 3: Port incomes to AMI, switch to a Single Qualification System for both Rental and Ownership, and adopt five (5) Income Categories (using Consultant Proposed AMI Ranges); and • Scenario 4: Port Incomes to AMI, switch to a Single Qualification System for both Rental and Ownership, and adopt five (5) Income Categories (using Staff Proposed AMI Ranges). Staff recommended Scenario 4 as the best policy option because it most closely met APCHA’s criteria for evaluating program improvements: simplicity, transparency, consistency, ease of transition, portability, and supporting overall housing goals. On May 3, 2017, the APCHA Board agreed to move forward with First Reading and public hearings to consider and discuss Scenario 4. On June, 21, 2017, the APCHA Board unanimously approved Resolution No. 02 (Series of 2017) at a public hearing. DISCUSSION The overriding conclusion from the 2016 Policy Study of the APCHA Workforce Housing Guidelines is that a major overhaul is warranted. Minor adjustments will not address the problems identified. The issues identified below address major elements of APCHA’s Guidelines: number of income categories; qualifying to buy or rent; Category Income Limits (aka Maximum Gross Incomes); and affordability. • The current Category Income Limits were derived from a combination of five sources used in a difficult-to-replicate methodology last calculated over 15 years ago (based on 1999 data). A new methodology is needed, which can be easily updated for simplification, transparency, and compatibility with State/Federal standards, and to set future sale prices and rents that will remain equitable and affordable over time. • Eight (8) income categories are more than any peer community studied, adding to the complexity of the program. So many categories are not needed to maintain economic and community diversity. • The number of categories varies depending on the program (4 for rentals, 7 for ownership, plus RO for both), which complicates any conversion to alternative methods (e.g. AMI) for establishing income categories. P3 I. Strengthening Community Through Workforce Housing 3 • APCHA’s dual qualification system creates unintended “discriminatory effect” concerns that likely violate the Fair Housing Act. • Category Income Limits differ depending upon whether a household is applying for a rental or ownership unit. This is unique among affordable housing programs, including peer communities, which use a single standard based on household size (i.e. persons per household). • Maximum Gross Incomes are not based on current area incomes. • Prices are not based on an adopted (defined) standard of affordability. In some categories, rents and prices are too high, about right, or too low relative to income, yet affordability is a clear objective of the Workforce Housing Program. • APCHA’s current income-calculation method results in a program that does not consistently serve the same households each year (i.e. the relationship between prices and income over time is inconsistent). In some years, defined income limits target higher-income households, while in some years they target lower-income households. • According to the 2016 Policy Study of APCHA’s Housing Guidelines, affordability (i.e. cost-burden) is an issue for households earning near the minimum incomes in lower (Category 1) and lower moderate income (Category 2) categories. • Affordability for upper moderate, middle, and upper middle income households (Category 3 and above) doesn’t appear to be an issue. The housing program appears to be over-subsidizing households in these categories, creating serious equity questions. As outlined in staff’s March 13 and April 19 Policy Memorandums to the Board, APCHA’s dual qualification system unintentionally creates a “discriminatory effect” because it sets different income limits for renters and owners, and classifies household size differently (i.e. based on number of adults for renters and number of dependents for owners). For example, a Category 1 couple with no children applying for ownership is subject to the $36,000 income limit, while a two-adult household with no children applying for rental can make up to $53,000 and still be Category 1 – a difference of $17,000. To solve this issue, the dual qualification system should be converted to a single qualification system based on Area Median Income (AMI) percentages per total number of persons in a household. Staff also recommended standardizing the total number of categories for both renters and owners to five (currently, there are four (4) rental income categories and seven (7) ownership income categories, plus RO for each). P4 I. Strengthening Community Through Workforce Housing 4 Below is a key features comparison of a dual- and single- qualification system. Because the current dual income limits cannot be converted to a single qualification system, new income limits using AMI percentages had to be calculated. In calculating new income limits, staff’s goal was to set Category income ranges so that the distribution of working households serving Pitkin County within that income category would most closely match the distribution of existing units in that income category relative to overall APCHA inventory. For example, given Category 3 units make up the largest portion of total housing inventory, the Category 3 minimum and maximum income limits were set so that the highest portion of households with a full- time employee working within Pitkin County would fall within that Category income range. To meet this goal, staff established a methodology to quantify how well the distribution of units matched the distribution of households within the different income ranges (methodology explained in full in the attached April 19th Policy Memorandum Appendix B: Distribution Matching Error Analysis Methodology). The proposed Category Income Limits are shown in the table below. Key Feature Dual System Single System Household Size Parameters Adults in Rental Dependents in Ownership Total Persons in Household Equity in Housing the Workforce Disfavors families Removes household type from the factors influencing who is housed Affordability Rents less affordable for households with dependents; Purchase prices less affordable for larger family households in upper Categories If incomes are the same, rents would be the same regardless of household type; bedrooms become the only variable Income Categories May qualify for homes in different ownership and rental Categories Consistent categories for owners and renters Compatibility with AMI Not compatible Compatible Fair Housing Raises concerns about discrimination based on familial status Not based on familial status Program Complexity More complicated Less Complicated Dual- and Single- Qualification System Comparison Minimum*Maximum Minimum Maximum Minimum Maximum Minimum Maximum Minimum Maximum 1-person 13,950$ 34,300$ 34,301$ 58,300$ 58,301$ 89,200$ 89,201$ 140,650$ 140,651$ 164,650$ 2-person 15,950$ 39,200$ 39,201$ 66,650$ 66,651$ 101,900$ 101,901$ 160,700$ 160,701$ 188,150$ 3-person 17,950$ 44,100$ 44,101$ 74,950$ 74,951$ 114,650$ 114,651$ 180,800$ 180,801$ 211,700$ 4-person 19,950$ 49,000$ 49,001$ 83,300$ 83,301$ 127,400$ 127,401$ 200,900$ 200,901$ 235,200$ 5-person 21,500$ 52,950$ 52,951$ 90,000$ 90,001$ 137,650$ 137,651$ 217,100$ 217,101$ 254,150$ 6-person 23,100$ 56,850$ 56,851$ 96,650$ 96,651$ 147,800$ 147,801$ 233,100$ 233,101$ 272,900$ *AP CHA does not define a minimum income for Category 1 households, but does require occupying adults to work full time (1,500 hrs per calendar year). 2017 Colorado minimum wage is $ 9.30/hr. Category 1 minimum incomes are the lowest possible gross incomes an adult could make and still qualify into the AP CHA program. The 1-person Category 1 minimum income is adjusted for all household sizes using the AM I Household Size Adjustments. Proposed Category Income Limit Ranges Household Size Category 1 (50% AMI)Category 2 (85% AMI) Category 3 (130% AMI) Category 4 (205% AMI) Category 5 (240% AMI) P5 I. Strengthening Community Through Workforce Housing 5 Justification for basing Income Limits on AMI Inability to update current methodology to establish new income limits While income categories were directly tied to area incomes when established in 2003 using 1999 Median Income, that is no longer the case. APCHA’s income categories are unique, they utilize an approach that adjusts the income limits each year only by changes in CPI, which is not linked to changes in area incomes or housing affordability. CPI is an index that measures changes in consumer prices of goods and services. Additionally, it is a regional index that is not directly linked to changes in the local Pitkin County economy. Recreating the methodology used for the 1999 Median Income, from which all the income levels are currently derived, would be costly and time intensive. This would need to be done periodically to ensure that these incomes are reflective of changes in area incomes over time. Given the resources required to establish a Median Income using this methodology, this is not a feasible option. What is AMI? Most affordable housing programs tie income categories to the Area Median Income (AMI). AMI is published annually by the Department of Housing and Urban Development (HUD) for each county and represents the median family income of an area. In contrast to APCHA’s current qualification system, HUD calculates income limits based on the number of persons per household whereas the APCHA calculates income limits based on number of dependents for ownership and number of adults for rental. HUD uses a combination of US Census, America Community Survey (ACS), and CPI information to update incomes and adjust for family size and for areas that have unusually high or low income-to-housing-cost relationships. Annual updating is a simple process using the change in AMI, which is calculated by HUD. HUD uses the same methodology to establish AMI for all states, counties, and metropolitan areas in the country. The following methodology was used to calculate the 2017 AMI, also referred to as Median Family Income (MFI) (see Using MFI Instead of MHI below): 1. The U.S. Census Bureau's 2010-2014 ACS median family income estimates are used as a basis for calculating HUD's FY2017 MFIs. In areas where the margin of error is more than half of the 2014 5-year ACS itself, the state non-metro estimate of median family income is used. 2. If there is a valid 2017 1-year ACS estimate of median family income available, HUD replaces the 5-year data with the 1-year data. A valid 1-year 2014 5-year ACS estimate is one where the margin of error of the estimate is less than one-half of the estimate. 3. Once the appropriate 2014 ACS data has been selected, an inflation factor based on the Congressional Budget Office forecast of the national CPI is calculated to inflate the estimate from mid-2014 to April, 2017 (or mid FY2017). HUD publishes one AMI per county yearly, mid-spring for the preceding year. The AMI published represents 100% AMI for a 4-person household. HUD and state agencies (e.g. Colorado Housing and Finance Authority) use a standardized methodology to create 100% AMI figures for all household sizes from the published AMI. The 2017 AMI for Pitkin County, used for the proposed income limits is $98,000. P6 I. Strengthening Community Through Workforce Housing 6 Using median incomes instead of average incomes A measure of central tendency is a central or typical value for a probability distribution, and can be used as an index. Measures of central tendency are either the mean, median, or mode of a dataset, and are selected depending on the distribution of the dataset. To equitably provide services to all Pitkin County residents, income limits must be set so that they reflect the distribution of incomes within the county. Mean (average) is the best measure of central tendency for normally distributed data; however, assessing the income distribution chart below, one can see that Pitkin County’s income data is positively skewed (i.e. not normal). Median is the best measure of central tendency for positively skewed data. The mean is not an accurate measure of central tendencies for family income distributions in Pitkin County because 76%* of family incomes are below the measure, and 24%* are above it. Whereas, the median evenly split the dataset, with 50% of family incomes below and 50% of above the median. Using the mean, or the average, in this case is misleading because a few extremely wealthy households in the County (e.g. billionaire households) will substantially raise the mean income making it look like households have higher incomes on average than they actually do. HUD’s calculations for AMI are accurate measures of central tendencies for income distributions because they are based on the median income instead of the average, or mean, income of the ACS dataset. Using the median removes the effect of outliers, both on the high and low side of the distribution, on the central measure of the distribution. Advantages of using AMI APCHA’s current income-calculation methodology results in a program that does not consistently serve the same target household income group each year. Because APCHA’s current income limits are not tied to target household incomes, the program serves a different income group depending on the year given the change in CPI. An advantage of linking income limits to HUD AMI is that the target household income group would remain consistent over time. Additional advantages of basing APCHA Category incomes on AMI include: • It is a highly reliable, trusted, and readily available data source; • It is objective and updated annually by HUD; P7 I. Strengthening Community Through Workforce Housing 7 • Reduces complexity and increases simplicity of system over time; • Increases fairness because it more consistently maintains the relative affordability of Categories over time; • Creates consistency with State (CHFA) and Federal housing programs (e.g. LIHTC Program) and multiple funding sources; • Is used by peer communities and would allow Aspen to evaluate itself against similar programs; and • Increases methodology consistency, uniformity, and transparency. It is not possible to directly translate APCHA’s current Category system into HUD AMI ranges because of the different underlying tier structure of the rental and ownership programs. However, AMI percentages can be estimated for each Category using some basic assumptions and information from the 2015 Employee Survey3. The AMI’s for households with an employee working in Pitkin County are now well documented, making it possible to convert the current incomes into AMI percentages. Using MFI instead of MHI Along with the Median Family Income (MFI), the ACS also annually publishes a Median Household Income (MHI). The 2014 and 2015 5-year ACS median household income estimates for Pitkin County are $71,060 and $71,196, respectively. Family income is the sum of the income of all family members 15 years and older living in the household. Families are groups of two or more people related by birth, marriage, or adoption and residing together; all such people are considered members of one family. Household income is the sum of the income of all people 15 years and older living in the household. A household includes related family members and all the unrelated people, if any, such as lodgers, foster children, wards, or employees who share the housing unit. A person living alone in a housing unit, or a group of unrelated people sharing a housing unit, is also counted as a household. Some argue Median Household Income (MHI) might be a better index to establish the new income limits than AMI, which is based on Median Family Income (MFI), simply because it “looks” better (i.e. lower). However, AMI is the best possible index option to set the new income limits based on the following (explained further below): • HUD income tables, based on MFI, are the best-defined income limits by area available; • ACS data lags and does not account for high housing costs in an area; • Using MHI to establish income limits would add to the complexity of the program and could deter private development of affordable housing; • There is no established methodology for adjusting annually published MHI, a single data point, to households of different sizes; and • Current income limits cannot be translated into MHI percentages based on household sizes. 3 Carried out as part of the 2016 Policy Study. P8 I. Strengthening Community Through Workforce Housing 8 According to a HUD specialist, as to the question of why HUD bases income limits on family income, this is not a specific requirement; however, it is rooted in Federal statute: 42 USC 1437(b). The term “family” is used throughout this section of statute; therefore, it follows that HUD would use the family income statistics to meet the definitions of “low income families,” “very low income families,” and “extremely low income families”4. The very low-income limits (usually based on 50 percent of MFI) are the basis of all other income limits, as they are the best-defined income limits and have been the subject of specific, limited legislative adjustments subsequent to reviews of the HUD calculation methodology. In addition, a number of other income limit calculations are tied by legislation or regulation to HUD’s AMI calculation. This is to create a uniform national standard for the relationship between the rent and income distributions in defining the high- and low-housing cost adjustments, and to prevent fluctuations in Low- Income Housing Tax Credit Difficult Development Area (DDA) determinations that result solely from high housing cost income limit fluctuations as areas go in and out of the 50th percentile FMR (Fair Market Rents) program.5 ACS annually publishes 5-year estimates, summarizing data from five years of annual surveys. Due to the time required to do the survey and then compile the results, ACS 5-year estimates lag by more than a year (as of July 6, 2017, the most recent ACS 5-year available was from 2015). To address the time lag in available ACS data, HUD adjusts MFI using an established methodology to account for inflation and changes in the market. Additionally, when calculating the AMI from the ACS MFI data, HUD also includes adjustments for high and low housing costs specific to a geographic region if applicable. Because HUD does not use the MHI to establish AMI or any other income limits, there are no such adjustments established for MHI. Using MHI as an index to establish income limits means the data would be dated and would not account for area specific housing costs, inhibiting APCHA’s ability to serve the same target markets year-by-year. One of the main rationales for porting to AMI based income limits is to reduce program/administrative complexity and increase transparency and fairness in how the program operates. The methodology proposed for the new AMI based income limits is one outlined in detail by HUD, and readily available to the public. Although not currently used for the general income limits, APCHA is required to use AMI income limits for its Low-Income Housing Tax Credit (LIHTC) properties. AMI income limits also affect private developer’s decisions to participate in development partnerships with APCHA, dictating the amount of tax credits they can receive. Continuing to base APCHA’s general income limits on a non-AMI index unnecessarily perpetuates program and administrative complexity of having to oversee properties with different income limits, and keeps a barrier to entry in place for developers who seek tax credits to build more affordable workforce housing inventory. AMI and MHI are both singular data points published annually. However, AMI can be calculated for all household sizes, whereas MHI cannot be. The AMI published represents 100% AMI for a 4-person household in a region; the published AMI is used to calculate income limits for all household sizes, using a methodology established and published by HUD. On the other hand, MHI simply represents the 4 Source: US Department of Housing and Urban Development, personal communication June 19, 2017 5 Source: HUD Income Limits Briefing Material – FY17 P9 I. Strengthening Community Through Workforce Housing 9 calculated median income for all households surveyed, with no specific reference to what household size it represents. The usefulness of MHI as a basis for income limits is further reduced by the fact that there is no established methodology for converting the singular data point to apply to households of different sizes. As part of the 2016 Consultant Policy Study of the APCHA Guidelines and Program, households containing at least one person employed in Pitkin County, APCHA’s customer base, were surveyed to draw data about their household sizes and household incomes. The survey data was then compiled and used to establish the estimated AMI percentages served by each of the APCHA Income Categories. The proposed AMI-based income limits take into account those calculations to reduce the effect of changing income limits on APCHA’s target market, especially household currently living in APCHA’s inventory. Choosing to base income limits on MHI instead of MFI (i.e. AMI) could not use the survey data because there is no available data on how APCHA’s current categories translate to MHI. Results of Establishing Income Limits Using MHI Using MHI data as an alternative to MFI (AMI) data to establish income limits would create arbitrary assumptions for both income limits and household sizes. Such a decision would be counterproductive to meeting the organizational goals of APCHA policy initiatives6, as well as fall short of the benefits of using AMI outlined above. RECOMMENDED ACTION APCHA recommends Council adopt amendments as unanimously passed by the APCHA Board of Directors. Attachments: APCHA Resolution No. 2 (Series of 2017) April 19th Policy Memorandum Appendix B: Distribution Matching Error Analysis Methodology 6 APCHA policy goals are aimed to decrease program and administrative complexity, increase effectiveness and efficiency, increase program transparency and predictability, increase the availability of useful, reliable data, and increase the fairness of program implementation. P10 I. P11I. P12I. P13I. P14I. P15I. P16I. P17I. P18I. P19I. 1 APPENDIX B: Distribution Matching Error Analysis Methodology In converting the Category Income Limits to a different methodology, a goal of APCHA is to align the distribution of units in each Category with the distribution of households whose incomes are within the respective proposed Categories. Therefore, AMI income ranges should be set to align the two distributions as closely as possible. The proposed Category AMI percentage range is bound by the “Upper AMI % Limit” for that Category and just above that of the Category below (i.e. in the Scenario 2 chart below, the Rental Category 2 income range is from 62.1% - 95% AMI). To quantify and evaluate the discrepancies between the income and unit distributions, staff created a model that compared the percentage distribution of units within each Category, with the percentage distribution of households earning incomes within the proposed Category income ranges. The difference in distributions is presented in the following tables as “Delta.” If the figure within the Delta column is shown as red, it means that households within that Category are underserved by the distribution of APCHA units available compared to the distribution of households within that Category. To quantify the total distribution discrepancy for all Categories within a program (rental, ownership, or total units) the Mean Absolute Error, “MAE,” was calculated. Finally, to assess how the total housing inventory distribution align with the household income distribution for all Categories, an equal “Weigh,” is assigned to each program and used with the MAE figures to calculate a “Weighted MAE.” Because Scenario 2 has two distinct AMI ranges per program, a MAE for total unit cannot be calculated and the Weighted MAE only pulls from the rental and ownership MAEs. In the body of the memorandum, Weighted MAE is referred to as “Distribution Matching Error.” Distribution Matching Error Analysis tables are shown below for Scenarios 2 (dual AMI qualification), 3 (single AMI qualification, consultant proposed AMI ranges), and 4 (single AMI qualification, staff proposed AMI ranges). Upper AMI % Limit Rental Survey Respondents Rental Units % (No Seasonal)Delta Upper AMI % Limit Owner Survey Respondents Sales Units % (No Seasonal)Delta Category 1 62%21.7%6.5%-15.2%50%4.6%1.2%-3.3% Category 2 95%26.0%26.5%0.5%84%15.5%12.4%-3.1% Category 3 147%34.5%36.7%2.2%114%22.7%15.8%-6.9% Category 4 240%15.2%3.3%-11.9%184%38.4%33.0%-5.5% Category 5 ----196%3.7%1.2%-2.6% Category 6 ----213%3.7%4.2%0.5% Category 7 ----234%3.2%0.2%-3.0% Not in Category >240%2.6%27.0%23.9%>234%8.1%31.9%23.8% MAE 10.8%6.1% Weight 50.0%50.0% Weighted MAE 8.41% Distribution Matching Error Analysis (Scenario 2)P20I. 2 Upper AMI % Limit Rental Survey Respondents Rental Units % (No Seasonal)Delta Owner Survey Respondents Sales Units % (No Seasonal)Delta Total Survey Respondents Total Units % (No Seasonal)Delta Category 1 50%14.6%6.5%-8.1%4.6%1.2%-3.3%9.6%3.4%-6.2% Category 2 85%24.6%26.5%1.9%16.2%12.4%-3.7%18.8%18.2%-0.6% Category 3 115%25.6%36.7%11.1%22.8%15.8%-7.0%23.4%24.4%1.1% Category 4 185%28.2%3.3%-24.9%38.0%33.0%-5.0%34.1%20.8%-13.4% Category 5 235%3.7%0.0%-3.7%10.4%5.6%-4.8%8.3%0.7%-7.6% Not in Category >235%3.3%27.0%23.7%8.0%31.9%23.9%5.9%29.9%24.0% MAE 12.2%8.0%8.8% Weight 33.3%33.3%33.3% Weighted MAE 9.66% Distribution Matching Error Analysis (Scenario 3) Upper AMI % Limit Rental Survey Respondents Rental Units % (No Seasonal)Delta Owner Survey Respondents Sales Units % (No Seasonal)Delta Total Survey Respondents Total Units % (No Seasonal)Delta Category 1 50%14.6%6.5%-8.1%4.6%1.2%-3.3%9.6%3.4%-6.2% Category 2 85%24.6%26.5%1.9%16.2%12.4%-3.7%18.8%18.2%-0.6% Category 3 130%35.4%36.7%1.3%34.0%15.8%-18.2%34.0%24.4%-9.6% Category 4 205%20.6%3.3%-17.3%32.4%33.0%0.6%27.7%20.8%-7.0% Category 5 240%1.7%0.0%-1.7%5.3%5.6%0.3%4.4%0.7%-3.7% Not in Category >240%3.1%27.0%23.9%7.5%31.9%24.4%5.5%29.9%24.4% MAE 9.0%8.4%8.6% Weight 33.3%33.3%33.3% Weighted MAE 8.68% Distribution Matching Error Analysis (Scenario 4)P21I.