The concept of productivity, including those for labour, capital and multi-factor productivity (MFP) are central to economic discussion about national economic performance, government policy and income distribution. There is common agreement that productivity growth has been, and remains, central to the long-run improvement in living standards. However, the orthodox or neoclassical, conceptual foundation of productivity, which informs both the work of the Australian Bureau of Statistics estimates of productivity growth, media analysis and economic policy is rarely subjected to critical analysis.
The issue of productivity, its measurement and the problems with these measures is a very large one. To assist important economic policy debates that revolve around the concept of productivity, David Richardson (The Australia Institute) and I have written this very short briefing paper, which is being published today on Progress in Political Economy (PPE). The paper sets out just a few of the long-standing questions that have been raised over many decades about the conceptual and empirical robustness of the concept and its role informing public policy.
The main findings include:
- Conventional productivity analysis can obscure the relation between productivity and profitability and, in relation to wage claims, obscure the ‘capacity’ of firms to pay.
- For some industries, such as mining, conventional measures of productivity can produce profoundly paradoxical results and undermine the usefulness of measured productivity as a basis for wage determination. This is because the ABS records the Australian mining industry as experiencing falling productivity for labour, capital and MFP, but also finds the industry experienced very rapid growth in net income, profit and profit per hour worked.
- The cause of the huge gap between productivity and profitability is that for mining the ‘quantity’ of output remained flat over the period but prices for mining output and the level of employment and capital investment increased substantially. Productivity calculations discount these commodity price increases with the result that the industry recorded no increase in the quantity of output but a large rise in the quantity of inputs over the period so that labour and capital productivity fell as did MFP. If the level of conventionally measured productivity was the only criteria for determining resource allocation it would be quite literally irrational for the large rise in investment in jobs and capital equipment to have occurred. But focusing solely on measured productivity obscures the fact that what actually determines resource allocation is profitability, which increased in both absolute and relative terms (rates of profitability) over the period.
- This perverse outcome reflects a broad range of deeper issues with conventional measures of productivity some of which are outlined below.
- Conventional measures of productivity require a perfectly competitive economy with many unrealistic and positively counter-factual assumptions such as perfect information, no scale economies, perfect factor mobility and firms as price takers.
- A key misunderstanding about ABS productivity estimates is that it provides an independent and objective measure of the contribution of labour and capital to production which reflect the ‘marginal product’ of these factors in the form of wages and profits, respectively. This follows from the fact that the neoclassical production function, used to estimate MFP and partial measures of productivity such as those labour and capital, is tautology. It is true by definition and does not, as its proponents suggest, provide an independent measurement of the relationship between inputs and outputs and provides no independent measure of the contribution of capital and labour to production. It thus provides no insight into the determination or measurement of distribution of national income. ABS estimates of the contribution of labour and capital are simply the result of an ‘adding-up’ exercise derived from ABS surveys of Australian economic units (firms, self-employed and public sector etc) of wages paid, expenses incurred and total output with profits being a residual (Gross Operating Surplus). This data is then ‘fed-into’ equations which result in estimates of the distribution derived from and closely matching this empirical data.
- Further, there are also major theoretical problem in defining capital, a key input to production, and also major practical difficulties converting theoretical constructs that underly MFP estimates into empirical measures.
Read the briefing paper here.
Comments