Date Published: February 14, 2017
Publisher: Public Library of Science
Author(s): Shyama Ratnasiri, Jayatilleke Bandara, George-John Nychas.
The Australian per capita consumption of ruminant meat such as beef and lamb has declined over the last two decades. Over the same period, however, per capita consumption of non-ruminant meat such as chicken and pork has continued to increase. Furthermore, it is now observed that the human consumption of kangaroo meat is on the rise. This study investigates the implications of these changes in meat consumption patterns on Green House Gases (GHGs) emission mitigation in Australia using a Vector Auto Regression (VAR) forecasting approach. Our results suggest that the increase will continue in non-ruminant meat consumption and this will not only offset the decline in ruminant meat consumption, but will also raise the overall per capita meat consumption by approximately 1% annually. The per capita GHGs emissions will likely decrease by approximately 2.3% per annum, due to the inclusion of non-ruminant meat in Australian diets. The GHGs emissions can further be reduced if the average Australian consumer partially replaces ruminant meat with kangaroo meat.
Even though reducing Greenhouse gases (GHGs) has been discussed around the globe for more than four decades with international bodies establishing several mitigation protocols as an outcome, the GHGs emitted by humans are still increasing every year. What this suggests is that mitigation efforts so far have not been able to deliver the desired outcome of reducing GHGs. This, in turn poses the question of whether we should be looking to explore avenues that have not been explored before to mitigate GHGs emissions. What has been explored so far is central to supply-side or production systems, with components such as technological or biological improvements, and supply management strategies that includes emissions targets and taxes. Until recent years very little was explored about demand-side components such as changes in consumer behaviour, socio-cultural and economic values, risk taking attitudes and the ways in which they can affect the GHGs emissions mitigation. This study attempts to contribute to this strand of literature by exploring changes in consumer behaviour, essentially dietary changes over time and their potential implications for GHGs emission reduction.
We explore the option of dietary changes in emissions mitigation in the context of Australian meat consumers in this study. The motivation is due to a recent observation that there has been significant change taking place in meat consumption patterns among Australian consumers. It has been observed that the Australian consumption of ruminant meat, such as beef and lamb, has declined over the last two decades. For example, as depicted in Fig 1, per capita beef and lamb consumption fell by approximately 45% and 64% respectively between 1974 to 2014 . Meantime, per capita consumption of non-ruminant meat such as chicken and pork continued to increase over the same period . In particular, per capita chicken consumption has tripled while per capita pork consumption doubled during the same period . Furthermore, it has been observed that human consumption of kangaroo meat has increased tremendously, with many supermarkets now having a designated section for kangaroo meat products. Although in the past kangaroo meat has been used as a pet food, the demand for human consumption has increased following the recognition that kangaroo meat is as a lean red meat substitute. Taken together, it is apparent that the consumption patterns of different meat products have changed significantly in recent years in Australia.
As noted previously, the key objectives of this empirical study is to explore the future patterns of meat consumption in Australia based on past experience and subsequent changes in GHGs emissions. For this purpose a forecasting exercise will be carried out using a simple Vector Auto Regression (VAR) model that entails estimation of a system of equation in which both prices and quantities are determined endogenously. The choice of the VAR considers forecast performance out of the sample. This is despite there being a large volume of literature in relation to consumer demand estimation including the Linear Expenditure System , the Rotterdam model [46, 47] the Almost Ideal Demand System (AIDS)  and their variants such as the models in [49–53]. Wang and Bessler  present a comprehensive discussion on forecasting approaches for meat demand.
In order to estimate the above model, data on consumption (Cit) of different meat types, prices (Pit and Pjt) of different meat products, and per capita income (It) of the consumers were collected for the period 1968–2011. The data on per capita consumption (Cit) of beef, lamb, pork and chicken and their prices (Pit) were obtained from Agricultural commodity statistics (2012) published by ABARE. Per capita income (It) data was also obtained from time series spread sheets published by Australian Bureau of Statistics (ABS). In our analysis, for a meat type ‘i’ in consideration, we use the prices of remaining meat types (i.e when i≠j.) as cross prices (Pjt) in the models estimated.
The properties of data and empirical results obtained from the model are presented and evaluated in this section. In Table 1 we present the results of unit root tests used to test stationarity of data. In this table, all C variables are related to per capita consumption of different types of meat, all P variables are related to different prices and I is per capita income. It is evident from the results shown in Table 1, that all variables are non-stationary at their level forms, however, they are stationary (i.e. I(1)) after the first difference. Therefore the above the VAR model was estimated in the first difference forms of the variables.
This study has attempted to examine the link between changing patterns in meat consumption and GHGs emissions in Australia and the possibility of reducing GHGs further by encouraging substitution of kangaroo meat for beef and lamb. We observed that Australian consumers have been substituting a significant amount of ruminant meat with non-ruminant meat in the recent past. Our results suggest that, if such a trend persists, per capita ruminant meat consumption (that is beef and lamb consumption) will gradually decline over time. This will, in turn, translate into a significant reduction in per capita GHG emissions resulting from meat consumption.