I believe it is just as important to understand the definition of something being measured as the measurement itself. The definition of a recession provides a fantastic and topical example. There have been numerous references in the press to Australia’s remarkable record of going 25 years without a recession. These assertions are based on a recession being defined as two consecutive quarters of negative growth in GDP.
This popular definition of a recession is attributed to US economist Julius Shiskin who first published this in a New York Times article titled The Changing Business Cycle in 1974. Having dug up the article from the New York Times archives I discovered that Shishkin’s definition was significantly more nuanced. He indicated a recession occurs when “there has been an extended, substantial and widespread decline in aggregate economic activity or depth, and diffusion”. He then helpfully translates this into the following six, easily understood quantitative definitions:
- Duration: Declines in real Gross National Product (G.N.P) for two consecutive quarters
- Duration: A decline in industrial production over a six-month period
- Depth: A 1.5 percent decline in real G.N.P.
- Depth: A 1.5 decline in non-agricultural employment
- Depth: A two-point rise in unemployment to a level of at least 6 percent
- Diffusion: A decline in non-agricultural employment in more than 75% f industries as measured over six-month spans, for 6 months or longer
Understanding that there are alternative definitions to measure a recession I wondered whether Australia’s 25-year record of no recessions would hold true if we considered some of the alternatives put forward by Shiskin.
At its core, the concept of a recession refers to a weak economic phase that deviates from historical trends. I wondered what an alternative RFi Analytics definition of a recession might look like. Rather than relying on a single measure like GDP or unemployment I decided to throw the wealth of ABS metrics, RFi Analytics proprietary data at the problem.
Taking into account the “three key dimensions of a recession, known as the ‘three Ds’: duration, depth and diffusion”...the ten measures included GDP and unemployment as well as consumer sentiment, credit delinquencies, consumer purchases, consumption expenditure and compensation of employees.
After some modelling, and taking into account the “three key dimensions of a recession, known as the ‘three Ds’: duration, depth and diffusion” (Fiedler 1990)1, I settled on ten measures. These included GDP and unemployment as well as consumer sentiment, secured and unsecured credit delinquencies, consumer purchases, consumption expenditure and compensation of employees.
As the five largest states in Australia tend to act as discrete economies I conducted this analysis for each state as well as at a national level. The result is the following “RFi Analytics, State of the States” report. The idea was to identify nationally and by state, where there has been a substantial decrease in economic activity affecting wide portions of the economy. If both the current and prior quarter are adverse it means that 7 or more of the indicators were negatively outside the historical bounds.
As this is more of a thought experiment than a formal paper I will forego getting into the detail of the methodology I used, including the bounds I used to define a “deviation from historical trends”. The ultimate goal here was to provide something that is easy to digest.
My key take outs when using this approach and definition:
- At a national level, we might be entering a period of weaker economic activity having been propped up to some extend by NSW and Victoria for the last two years.
- Western Australia could be construed as “being in recession”
- South Australia and Queensland would also be defined as being in a recession based on the last two quarters
- Only Victoria and New South Wales would be seen as in a positive cycle for the last two quarters – although the most recent quarter would suggest both may be experiencing a decrease in economic activity.
Now I am not an economist and agree that almost anything can be proven when playing with assumptions, definitions and statistics. However, this exercise will hopefully demonstrate two important things I have learnt in 15 years of conducting industry portfolio benchmarking.
- Understanding the definition of a metric is just as important as understanding the result
- Just because a metric, whether it is market share, a recession or delinquent accounts has “always” been defined in a particular way, do not assume this is the best / only definition. Often the best insights come from looking to redefine the inputs into any metric.
I do not think Australia is about to enter a recession based on the commonly used definition. I do however think it is important to understand that each of the major states act as a discrete economy.
Finally, I do not think Australia is about to enter a recession based on the commonly used definition. I do however think it is important to understand that each of the major states act as a discrete economy. If you rely on the traditional definition of a recession to make decisions and forecasts for your business or in your models, it is worth considering using a proxy value such as “Final Consumption Expenditure” which represents about 70% of GDP to get a quarterly snapshot of how each of the states are doing.
1 Fiedler, E. R. (1990), The Future Lies Ahead, in: Klein, P. A. (ed.), Analyzing Modern Business Cycles, Essays Honoring Geoffrey H. Moore, Armonk, NY: M. E. Sharpe.