The Power of the Area Median Income
Citymark uses area median income as a tool to explore housing affordability. Take a look at the Citymark dashboard and our Citymark reports.
In our work, the most common question we are ever asked is “what’s affordable?” At conferences, in reports, media events, conference calls, and in elevators, at dinner tables and over grocery tills.
In my experience (the United States), the definition of affordability is fairly universal – the cost of housing, whether rent or a bond, which does not exceed 25 – 30% of one’s gross monthly income. The challenge arises when providing housing for those at the middle and bottom income bands, where CAHF’s mission sits, when families cannot afford to pay for the cost it took to produce it. Therein lies the rub – and the affordable housing sector. A vibrant housing market is one which offers safe, decent and affordable housing options to a wide range of incomes and budgets. South Africa in that context, with lots of houses in high markets and low markets, is a market prime for development. The question of building towards that huge market in the middle, where costs outstrip incomes, becomes where and how.
Until recently, South Africa was in a very tough position to understand and agree on affordability because the data was not readily available. Affordability was defined as some set price which seemed affordable to some subset of South African income earners, most commonly around R500 000. The challenge with that measure can be simply illustrated. In one city, about 40% of the housing units are worth below R500 000, while in another, about 70% of housing units are worth below R500 000. From this viewpoint, the second city seems more affordable. However, when local incomes are taken into consideration, in the first city, it takes about 2 times the average local income to afford the average house – the second city? Almost five times the average income. The first city is Johannesburg, the second is Mangaung. Not that people are paying two and three times their incomes (they are doubling and tripling up, or they are not being served by the current market), but the measure simply provides a quick illustration of where the trouble lies, and provides some important insight into the scale of the gap, and some way of considering the spectrum of priority solutions.
How can we know for sure what is affordable? With the digitized census released in 2012, a very powerful tool for measuring affordability, and progress towards expanding housing markets, was made possible – the area median income. The area median income is simply the average income for an area, from a neighbourhood like Jabavu to a metro like Johannesburg. In the example above, affordability can be defined locally because we know at the municipal level for example that Johannesburg’s average income is about R17 000, one of the highest in the country, and Mangaung’s is about R9 500. The most powerful way to use area median income is to set and share specific affordable goals, and to estimate the cost of meeting those goals. For example, cities can set and calculate the impact of targeting certain affordability levels as it approves projects. For access to land or subsidies, projects may be required to make some percentage of the units affordable to 80% of the area median income, say (in Johannesburg, that’s income of up to R 14 000, in Mangaung, that’s up to R7 600. Other sites or programs can target moderate income households, say those earning between 80 – 120% of the area median income (Johannesburg, that’s up to R20 400, in Mangaung, that’s up to R11 400), perhaps as a means of protecting affordability in high cost areas, or enticing interest in hard-to-develop areas. Against the cost of producing quality affordable housing, these parameters help quantify and compare the local housing shortfalls, and builds a conversation between the developers and investors eager to play in the gap, and the governments eager to engage its partners.
Affordable housing solutions are going to be very different from city to city, and from neighbourhood to neighbourhood, in order for housing markets to work. In Johannesburg, with a lower affordability gap, perhaps focusing on costs will help, such as delivering higher densities (including quality rental), exploiting construction efficiencies and innovations (just-in-time technologies), and pursuing better locations (more of a family’s income can be spent on housing instead of fuel). In Mangaung, financing tools will have to be created to make up for the significant income gap. Neither government nor its private sector partners can sustain housing production that can’t get this mix right.
With income data available at the suburb level, cities can identify housing gaps and negotiate tradeoffs at the site level with its developments partners clearly and with shared understanding. For example, if local incomes can afford on average sales prices 20% below the cost of housing delivery, are there ways that cities can help cover that amount? Cities can trade valuable land and grants for affordable housing set asides, while estimating the values related to those swaps more effectively. Reducing the cost of land on the development side creates cost savings which can be converted into lower housing costs on the supply side, helping to meet those housing targets. Using the area median income provides a means with which to relate income and expenses.
Governments and its partners can also better track whether markets are growing more or less affordable over time. Are incomes keeping pace with prices locally? Are lenders lending in the gap? All of these measures are more accurate if measured by percent of area median income rather than specific income or price levels (that is, are homes affordable to 80% of the area median income becoming more available? can the current housing stock provide an affordable option? ). South African housing markets are very different – from rural to urban, from Sandton to Soweto – but the area median income provides the ability to test whether the same segment of people – those earning 50% of the area median income, or 80% say – are moving ahead or not.
Of course, limitations must be taken into consideration when using this technique. The data is imperfect (all data is, and in particular data measuring human movement and momentum), averages are tough in an environment of such extremes, and new methods must be considered carefully. However, census data is consistent and freely available across every area in the country and thus can be calculated consistently in the same way everywhere; experts exist who can adjust for local conditions, and thus anyone can use it. This method, in existence for decades in other areas, provides the means for benchmarking goals, comparing opportunities and projecting costs in powerful ways previously unavailable in South Africa. Most importantly, it provides the means of finding agreement – on the goal, on the impact, on sharing the solutions. And that is the most powerful part of all.
This was prepared with the support of the National Housing Finance Corporation.