Is the reduction in housing affordability encouraging risk taking?

Asset prices can have a range of impacts on investor behaviour. Often these dynamics are represented as restricted to those associated with the price of the asset itself; i.e. momentum vs value seeking investors. Overlooked is the potential impact on investor dynamics of other factors outside the specific asset prices. A good example of this is home affordability and how that may impact on risk taking. To better understand these dynamics and potential impacts it is important to move away from the more traditional models of utility and look towards Desperation Threshold Models.

Evolution of Utility Models

To put the Desperation Threshold Model (‘DTM’) in a broader context it is worth reviewing at a very high level the evolution of some of the better known concepts of utility. Traditional, or classical, utility theory sets out a world where the relationship between utility and resources is one of diminishing marginal returns. In this world the more units of a resource someone has the greater the utility from an increase in the supply of the resource but the increase in utility occurs at a diminishing rate. The result is a relationship between resources and utility which can be described by a logarithmic function (see Figure 1).

Figure 1 : Traditional Utility Function

Later Prospect Theory shifted the emphasis from the absolute level of resources to the expected gain or loss in resources associated with a decision. This modification was made by Kahneman and Tversky to incorporate the real life observations that (a) individuals exhibit loss aversion and (b) tend to overweight small probabilities and underweight large ones. Utility was now evaluated by the expected change in resources relative to a reference point, often the status quo, with a loss in utility being more heavily weighted compared to the equivalent gain. The result was a shift towards a kinked distribution as individuals now exhibited risk aversion (see Figure 2).

Figure 2 : Prospect Theory ‘Risk Aversion’ Utility Function

DTM in turn added another element namely the relative position of the individual with respect to a reference level of resources. In contrast to Traditional Utility, which focussed just on total resources, DTM focussed on total resources relative to some ‘reference point’. Furthermore, unlike Prospect Theory the ‘reference point’ is not the individual’s current position, or ‘status quo’, but rather a level of resources associated with ‘basic needs’. The utility derived by an individual from a gain or loss in resources is now a function of the resources possessed by an individual relative to the level of resources required to satisfy ‘basic needs’. DTM is accordingly an extension of the ideas put forward by both Traditional Utility and Prospect Theories but adds further granularity to the decision making process.

What is the Desperation Threshold Model?

At its core DTM can be summarized as follows: humans have a strong preference for having at least some amount of resources that represent their ‘basic needs’. Above this level, they continue to derive utility from resources, but this is less important than keeping their ‘basic needs’ secured. The result is a utility function which features a steep region, representing that at some point resources are particularly valuable because they secure ‘basic needs’. Below the threshold, the utility function flattens out representing the intuition that as one moves further below ‘basic needs’ there is less to lose. Indeed at a point where the individual’s resources are so far below ‘basic needs’ the utility function becomes flat highlighting that at some point the individual has ‘nothing to lose’. Above the ‘basic needs’ threshold, utility increases in the traditional logarithmic fashion along with the level of resources; i.e traditional utility theory largely holds after ‘basic needs’ have been comfortably met (see Figure 3).

Figure 3 : DTM Utility Function

The highly kinked utility function results in two key predictions regarding individual behaviour. Namely :

1. Caution prediction. Individuals whose resources are only just adequate to meet their ‘basic needs’ will show greater avoidance of risk, compared to those whose resources are abundant. In Figure 3 this is denoted by the shaded area either side of the Desperation Threshold. Due to the ‘cliff’ in the utility function created by the Desperation Threshold small losses in resources have outsized impacts on utility as they increase the risk of falling below ‘basic needs’; i.e. focus of individuals in the region is on not falling below or remaining close to the desperation threshold.
2. Desperation prediction. Individuals who are unable to meet their ‘basic needs’ will take greater risk, compared to those who can meet their ‘basic needs’. The rationale is that for those individuals with less to lose the potential upside from taking additional risks more easily outweigh the potential losses as the key focus of individuals is on achieving ‘basic needs’; i.e. increase in utility from achieving ‘basic needs’ disproportionately outweighs decrease in utility from any losses. Accordingly, individuals further below the Desperation Threshold are more likely to take risks which could easily be avoided and indeed may appear irrational to those above the Desperation Threshold.

What Exactly is the Desperation Threshold?

As stated earlier the desperation threshold refers to the level of resources required to satisfy ‘basic needs’ : But what are ‘basic needs”? ‘Basic needs’ can be defined as “the minimum resources required for human survival and long-term physical well-being”. Often economists use this definition to establish a "poverty line"—the exact income required to purchase these necessary goods and services. If a household's income falls below this threshold, they are unable to afford ‘basic needs'.

It is now important to make the distinction between ‘basic needs’ and ‘wants’. ‘Wants’ can be viewed as any good or service that people desire or would like to have but is not essential for survival. Unlike ‘basic needs’ wants are subjective, unlimited, and change based on personal preferences, culture, and evolving trends. In contrast to the discretionary ‘wants’, ‘basic needs’ are universal, absolute necessities, primarily consisting of food, water, shelter, and clothing. The distinction is important as the utility derived from ‘wants’ is more likely to behave in accordance with traditional or prospect models whereas that from ‘basic needs’ is more likely to follow DTM.

Though ’basic needs’ is often utilised to define a ‘poverty line’ there is no reason why the concept should be limited to such a threshold. There is intrinsically no reason why the perception of ‘basic needs’ may not differ between groups and over time. Specifically a broader concept of ‘basic needs’ may exist for a large part of the population materially above the ‘poverty line’. For such individuals the definition of ‘basic needs’ may broaden to become the level of resources required to provide for a better than modest retirement. The perception of ‘basic needs’ are now extended to cover not just the level of resources needed for immediate survival but also for survival over the anticipated life cycle of the individual. The Desperation Threshold accordingly becomes the level of resources required to provide, at the very least, for a modest retirement.

Why is Housing Affordability so Important?

The impact of homeownership on utility and how housing affordability impacts on utility comes down to whether home ownership is considered a ‘basic need’ or a ‘want’. In a perfect world one could argue that ‘basic needs’ simply requires shelter and therefore the form of that shelter, whether it be via home ownership or home rental, is irrelevant; i.e. households should be indifferent to ownership or renting as long as they have shelter. Home affordability under this framework only becomes important if it feeds into the perception of the level of resources necessary to sustain ‘basic needs’ over the life cycle.

To identify any bias between home ownership versus rental as the source for shelter it is important to consider the institutional framework in which the individual is making decisions. Within Australian the institutional framework underpinning the retirement income system has long implicitly taken for granted the default assumption that the vast majority of retired people will have very low housing costs; i.e. presumption that retirees own their own home with little of no debt attached. Consistent with this view is that when the Association of Superannuation Funds of Australia (‘ASFA’) sets its retirement standards it assumes that retirees own their homes outright and hence make no provision for housing costs; i.e explicitly assumes no mortgage or rental costs. It has been argued that this presumption in turn has allowed successive Australian Governments to maintain aged pensions at lower levels than those in most other ‘advanced’ economies without resulting in higher levels of poverty among retirees. At an institutional level the entire concept of a safety net in retirement provided by the social security system is predicated on the retiree having access to debt free home ownership.

Debt free home ownership is therefore a significant component of the ‘three pillar’ framework to retirement planning built around government pensions, compulsory superannuation and voluntary private savings. One could even go further and consider that debt free home ownership is so important that it should be considered the fourth pillar of the retirement income system. To put the material benefits which home ownership provides in context it is useful to consider that retirees who have paid off their mortgage spend much less of their income on housing (on average 5%) than working homeowners or retired renters (25% to 30%). These benefits – which economists call imputed rents – are worth roughly as much again as the maximum pension. The end result is that individuals who retire without owning a home are at greater risk of living in poverty and experiencing housing stress. Analysis by the Grattan Institute shows about two-thirds of retired Australians who rent privately live in poverty, compared to only 11% of outright homeowners (see Figure 4)

Home ownership, within such an institutional framework, becomes less a ‘want’ and more a ‘basic need’ which is required to ensure the base level of resources to sustain the individual over their life cycle. The resulting dynamics associated with home affordability are accordingly more likely to follow those set out by the DTM *.

With home ownership being seen as a ‘basic need’ the impacts of housing affordability on individual risk taking can be quite material under the DTM framework. Those that are close to the threshold will be less risk averse. More importantly those well below the threshold will be willing to take materially more risks in an attempt to achieve home ownership. Accordingly, the ongoing declines in home affordability has the potential to increase the overall level of risk taking within the non home ownership cohort as they strive to achieve the ‘basic needs’ to sustain resources over the life cycle. The impact on risk taking may in turn have flow-on effects to other asset classes by increasing the investor appetite for higher risk/return assets and strategies.

* Worth noting that under a DTM framework the same dynamic may flow through to home owners with low superannuation balances. As ‘basic needs’ cover resources over a life cycle those with low superannuation balances may be encouraged to take excessive risk where they are viewed as being 'too far' below a DTM. With such low superannuation balances the perceived downside from potential losses, given the availability of a government pension as a safety net, may be more than offset by the potential upside available from higher risk investment opportunities.