RDP 2015-01: Stress Testing the Australian Household Sector Using the HILDA Survey 2. Literature Review

Stress testing typically attempts to quantify the impact of adverse scenarios, such as recessions and serious financial shocks, on financial institutions.[1] Private financial institutions use stress tests as part of their internal risk management. Prudential supervisors and other authorities also use stress tests to assess vulnerabilities facing individual financial institutions or financial systems as a whole.

The prominence of system-wide stress testing has increased since the onset of the global financial crisis, partly because authorities have wanted to make assessments of financial system resilience more forward-looking. The most common risk assessed in these stress tests is credit risk – the risk that borrowers will not repay their debts – given its central role in past episodes of financial instability. Stress tests are also increasingly used to assess other risks, such as liquidity risk.

Simulations based on cross-sections of household-level data (household micro-simulations) have become increasingly popular tools for stress testing household credit risk. One method for performing these is the ‘financial margin’ approach, where each household is assigned a financial margin, which is usually the difference between each household's income and estimated minimum expenses.[2] Under this approach, households with negative financial margins are assumed to default on their debts. An alternative method is the ‘threshold’ approach, where each household is assumed to default when a certain financial threshold is breached (for example, when total debt-servicing costs exceed 40 per cent of income).[3] Information on household balance sheets can be used to estimate loss given default and, when combined with information on which households are assumed to default, can be used to estimate ‘debt at risk’ (or expected loan losses).

Under either approach, shocks to macroeconomic variables – including asset prices, exchange rates, interest rates and the unemployment rate – can be applied. The impacts of these shocks can be estimated by comparing pre- and post-shock default rates and loan losses. Shocks are typically applied in a single-period framework, where defaults and loan losses occur instantaneously.

Both approaches have advantages and disadvantages. Financial margin-based approaches more closely match the processes lenders typically use to determine loan serviceability. Threshold-based approaches require fewer assumptions, but these assumptions may be overly simplistic; specifically, the assumption that all households with debt-servicing costs above a certain threshold will default may be unrealistic. Indeed, higher-income households should be better able to bear higher debt-servicing ratios than lower-income households.

Either approach can be modified to allow households to draw down on their assets or borrow against suitable collateral to avoid default. For example, Herrala and Kauko (2007) and Karasulu (2008) include a measure of assets directly into each household's financial margin. This approach reduces the number of wealthy households that are expected to default. The Bank of Canada (Djoudad 2012; Faruqui et al 2012) employs a multi-period model, where households default if they are unable to service their debts for a period of at least three consecutive months. In this model, households are able to draw down on their liquid asset holdings to help service their debts following an unexpected spell of unemployment. The model also allows unemployed households to return to employment in later periods.

As far as we are aware, there have been no published studies using household micro-simulations to stress test the Australian household sector. The model in this paper is based on a financial margin approach and shares many features with Albacete and Fessler's (2010) model for the Austrian household sector.


See Bilston and Rodgers (2013) for more detail on central banks' stress-testing frameworks. See APRA (2010) for more information on regulatory stress-testing practices in general. [1]

Financial margin-type approaches are also known as the household budget constraint method, financial surplus method or the residual income approach. For some examples of these approaches, see Johansson and Persson (2007) and Sveriges Riksbank (2009) for Sweden, Holló and Papp (2007) for Hungary, Herrala and Kauko (2007) for Finland, Andersen et al (2008) for Norway, Albacete and Fessler (2010) for Austria, and Sugawara and Zaluendo (2011) for Croatia. [2]

Threshold-type approaches have been used for the household sectors of Canada (Faruqui, Liu and Roberts 2012), Chile (Fuenzalida and Ruiz-Tagle 2009) and Korea (Karasulu 2008), among others. [3]