The multimillion dollar consultant reports from the Murray-Darling Basin Authority are a far cry from being useful in determining the impacts of the basin plan.
Some of the key figures provided by the MDBA on Gross Regional Product (GRP), Gross Value of Irrigated Agricultural Production (GVIAP) and employment are the results of quantitative economic modelling that was commissioned by them for the draft basin plan. The basis for this economic modelling work lies in Computed General Equilibrium (CGE) models that are inherently complex, prone to misspecification and unreliable in estimating regional effects.
CGE models are based on a system of equations describing economic variables relating to household, production and government sectors. The equations used to estimate the economic impacts of the proposed basin plan are neo-classical in nature and linear in their structure. Here lie some, but not all, of the problems of the MDBA’s economic modelling.
A neo-classical framework creates a utopic image of the economy in which pareto optimality applies. This means that households are assumed to utility maximise based on rational decision making and producers cost minimise based on profit outcomes without a consideration of wider social issues.
To give a few examples of the assumptions imposed on the economic models, we need to look more closely at the structural specifications of the two core models used; AusRegion and TERM-H20. Both models are CGE models where labour and capital is assumed to be perfectly mobile, individuals have perfect foresight and make rational decisions about consumption and investment expenditure.
Based on these assumptions it is not surprising that the employment impacts of the proposed basin plan are minimal with estimates ranging from -1600 to +300. If workers and capital are assumed to move freely, then individual A who loses his job as a result of lower water availability in one region, he will relocate to another region to work; all under the assumption that adjustment costs are small. It should be mentioned here that all the employment figures from the MDBA’s economic modelling are quoted as net figures, hence the reduction of employment in one region is assumed to be offset by the gain in employment in another region.
It is quite evident that this assumption fails to capture realistic assumptions on an individual’s decision making. Even if we assume that labour is completely mobile, then something else has to adjust if water is taken from communities. A decrease in water available for consumptive use will likely affect output adversely. Lower output and associated lower labour demand will only lead to reasonable stable employment if workers accept lower nominal wages.
Furthermore, lower output is also likely to increase prices, which will act as another constraint on regional economies. If wages are taken as a proxy for welfare, then the employment figures quoted by the MDBA only show a partial picture of the impact of the proposed basin plan.
Turning our attention to the linearity assumption imposed on the economic models, we can see that this has very particular implications for GRP and GVIAP results. The outcomes of the economic modelling show that GRP and GVIAP are supposed to decrease by less than 1% and 10% respectively. Linearity is not the only assumption that drives such low output reduction figures. Two additional assumptions that influence these results are that the proceeds from water buybacks are assumed to remain within regional communities resulting in greater local stimulus and water buybacks occur over a long time horizon and hence communities are able to adapt and industries are able to redeploy labour and capital to other productive use.
Given the increased financial constraints that individuals and communities currently face due to the prolonged drought and the GFC, the assumption that additional capital remains within the communities is not realistic. Lower consumer confidence as a result of remaining uncertainty of future water availability and the additional recent developments in energy prices and energy related service charges is likely to result in a lower propensity to consume and hence the resulting local stimulus might be substantially lower than is assumed by the economic modelling.