Measuring Moderate-Income Affordability

8 min read

Is there a more suitable approach than AMI?  

California incomes have been booming, especially in the high-tech sector. The state now has the fifth largest economy in the world. This concentrated wealth has led to housing affordability concerns for not only low-income households, but also for the moderate-income (workforce) population or “the missing middle.”

Despite having gainful employment, many essential moderate-income workers (e.g., police officers, fire fighters, teachers, nurses, etc.) in California’s affluent population centers face stark issues related to housing affordability and proximity to their workplace. Today, we focus on the first of these two problems – rental housing affordability for moderate-income workers.

(In a follow-up discussion, we will examine issues of proximity to the workplace and associated transportation costs – pecuniary, time and emotional.)

Economists sometimes find themselves depending upon unsuitable measures, either because they are readily available or have become accepted go-to standards. The central statistic currently used in determining affordability for rental assistance programs is household income relative to the area median income (AMI). Moderate-income housing is usually targeted to households earning between 60 percent and 120 percent of AMI. But while it may appear that households earning around the median is a suitable indicator of a need for moderate-income rental housing, income alone may not be the best measure in determining affordability. Including measures that compare household income to market rental prices for a given location—such as the National Housing Conference’s (NHC) “Paycheck to Paycheck” definition of affordability, or the National Low Income Housing Coalition’s (NLIHC) “Hourly Wage” measure—are preferable to AMI. Even though market rental prices are often correlated to AMIs, area rental prices are influenced by additional supply and demand factors that AMI omits, making them more accurate measures of moderate-income affordability.

To understand this, let’s define AMI and briefly discuss why it is used as a central measure for housing affordability. Then we’ll veer to a detailed example of another suboptimal measure in economics – gross domestic product (GDP). By examining GDP, we can see how economists are often constrained to work with imperfect metrics because they are widely accepted and are feasible to measure (and are thus deemed “suitable”). After we see why these measures are unsuitable, we can then return to housing affordability and discuss how we can better evaluate it with alternatives to AMI that will provide a better understanding of rental markets. Though this is applicable to most, if not all markets, it will be illustrated here using data from selected metro areas in California.

Area median income
AMI is used to determine applicant eligibility for the U.S. Department of Housing and Urban Development’s (HUD) assisted housing programs. It has thus become the basis for determining housing affordability.

HUD’s purpose in calculating geographic AMIs is to provide input in low-income rental voucher and tax credit programs. We are not questioning whether AMIs are appropriate measures for a low-income program, but whether they are suitable when addressing moderate-income affordability where limited (or “missing”) public support is available.

The following table shows 2018 HUD AMIs for selected California metro areas, and compares them to their 2017 and 2016 values.

Of note in this table is the ten percent AMI growth between 2017 and 2018 in San Jose. This increase is stark. Statisticians often use medians (as opposed to averages) because they are less affected by outliers and are considered stable. To illustrate, suppose all the top earners in an area receive large increases in income. If they are less than 50 percent of the population, even though the average increases, the median does not change (assuming all other incomes are constant)2. The increase in median will occur if more than 50 percent of the population receives an increase in income. The median would also increase if high-income households move into the area (and low- to moderate-income households leave). In any case, given the stability of AMI measures, the ten percent increase in San Jose underlines the changing income profile among high-tech workers.

AMI is thus a measure designed for extremely low-, very low- and low-income housing programs, and provides an effective way to compare incomes relative to the middle-ranked income household in a geographic area. While AMI provides a valuable input into housing voucher and tax credit programs, it is an unsuitable measure when discussing housing issues facing moderate-income workers – these workers are (generally) not receiving housing subsidies, and their affordability math is based on prevailing rental prices (that are determined by many factors in addition to the metro level income distribution).

So, AMI is used in the debate on how to ensure apt affordable housing for essential moderate-income workers in the community because the metric is available and it is widely accepted for low-income programs. But that does not mean it is really suitable. The acceptance of unsuitable (or suboptimal) measures is not limited to our current topic; it is surprisingly common in economics. As evidence, let’s take a brief detour:

Alternative to AMI for moderate-income workers
Is there an alternative we can adopt as an alternative to AMI? Yes. We can, instead, define affordability based on the rent a household can afford. The advantages of this are:

  1. HUD already publishes the Fair Market Rent data needed to calculate this measure. It is thus feasible and easy to use; and
  2. We are not reinventing the wheel. This measure has been suggested by both the National Housing Council (NHC) and the National Low Income Housing Council (NLIHC), so it can be readily accepted in the affordable housing community.

The measure is defined as the amount a household must earn to afford a modest and safe rental home without spending more than 30 percent of its income on housing costs. It is based on HUD’s Fair Market Rents (FMR) that are published along with AMI data each year. The advantage of using geographically based rental prices is that they comprise market factors not included in AMI, for example, inventory (or lack thereof) and its quality.

For example, the 2017 two-bedroom FMR in San Francisco was $3,018. The NHC qualifying annual income for such an apartment is = ($3,018*12)/0.3 = $120,720, and NLIHC’s Hourly Wage is ($120,720/2,080) = $58.04.

The following chart compares that amount needed to afford a two-bedroom rental in selected metro areas in California and compares the amount to the three-person AMI.

The chart above shows the positive relationship between affordability and AMI. However, note that the annual amount needed to afford the rent in San Francisco is close to 120 percent of the three-person AMI, whereas it is closer to 60 percent in Sacramento. The steeper blue affordability line in the chart indicates that availability is even tighter in metropolitan areas with higher incomes. This may not come as a surprise, yet it emphasizes that AMI does not contain all the information needed to study affordability for moderate-income workers.

Looking at average annual incomes, many essential workers—nurses, police officers, elementary school teachers, EMTs—cannot afford two-bedroom rentals in California’s larger cities (LA, SF, SD) and even in smaller cities (Stockton, Merced). In smaller cities, rents may be lower, but so are salaries. Some essential workers in California are sacrificing non-housing consumption (and savings) to live close to their workplaces.

In Closing
Rental affordability is a pressing issue for moderate-income workers in the more affluent population centers in California. While we often examine the “missing middle” workforce using AMI, this is deficient since we need to also consider other factors, such as lack of inventory, that drive rental prices. By relating incomes to rental prices (instead of median incomes) we can get a clearer picture of the situation in California and across the nation.

Mismeasurement in Economics: Gross Domestic Product

  • GDP, the main measure of the aggregate output of the economy, excludes non-market activity. This underreporting issue is well-documented, and attempts have been made to measure the effect of omitting non-market activity, such as household production, from GDP estimates (e.g., Bridgman (2016), Bureau of Economic Analysis).
  • Household production includes housework, cooking, odd jobs, gardening, shopping, child care and many other activities. To illustrate, let’s focus on food preparation.
  • As female labor force participation increased (from 36 percent in 1955 to 57 percent in 2017), there was an accompanying large shift in where meals were consumed. The share of food dollars spent outside the home grew from 25 percent to 48 percent between 1955 and 2017, and restaurant sales increased from $43 billion in 1970 to $798 billion in 2017 (National Restaurant Association). These sales (that subsume 14.7 million restaurant workers’ incomes) are included in national accounts, whereas the time spent on food preparation inside the home is not.
  • This means that as food consumption shifted from a non-market to a market setting, recorded GDP growth has “overstated” the actual value added. (Bridgman estimates that the value of household services was 37percent of GDP in 1965, and is currently 23 percent).
  • In other words, while we use GDP to measure economic activity, we are aware that it is far from perfect. However, it is still widely reported and used since an improved and (perhaps more importantly) agreed upon metric has not been widely adopted.
Edward Seiler, Ph.D. is the VP, Research and Economic Analysis of the Washington D.C.-based Dworbell, Inc., a trade association management firm providing public policy analysis and advocacy. In this capacity, Seiler will lead economic research for the organizations that Dworbell manages, including the National Housing & Rehabilitation Association (NH&RA), the National Reverse Mortgage Lenders Association (NRMLA), and the National Aging in Place Council (NAIPC). Seiler was previously Chief Housing Economist and Director at Summit Consulting, an analytics firm with expertise in applied economics and mortgage finance. Prior to joining Summit, Dr. Seiler was Director of Economics at Fannie Mae, where he directed the development and implementation of analytical models used to guide credit loss management decisions. He has lectured graduate-level micro-econometrics at Johns Hopkins University and published several peer-reviewed articles. Dr. Seiler was previously employed as a manager at Bates White (an economics litigation consulting firm) and as a post-doctoral fellow at The Hebrew University. He earned his Ph.D. in economics from The University of Chicago, where he was a Fulbright Scholar.