The demand interpretation of productivity outcomes (technical)
Most contributions to the debate on the productivity puzzle recognise both demand and supply must play a part. But any arguments about demand are normally set in a framework that isolates microeconomic factors in the specific context of the productivity outcomes, normally market failures such as rigidities (e.g. labour hoarding) and misallocations (e.g. finance allocated disproportionately to low productivity firms). Arguably they are not demand failings at all, but supply failings.
A genuine demand account must be set in a framework based on macroeconomic and monetary analysis of aggregate outcomes. As discussed in the main post, low productivity has followed the labour market accommodating demand-driven lower growth through price/wages rather than quantity/employment.
In the conventional theory, (real) wages are a function of productivity and cannot be viewed as acting exceptionally as part of any adjustment process. Indeed this account is not well suited to explaining any short-run or cyclical factors. Outcomes are determined by the inputs to or factors of production (notably capital and labour), and the efficiency with which those inputs are used is understood as productivity. (Be it simple labour productivity or a more involved notion of multi- or total-factor productivity.) Moreover there is no role for price or money wages here, everything is determined as quantities of the ‘real output’ of the economy, beyond inflation expectations. Cyclicality is not well explained; again seemingly it is down to rigidities in adjustment processes that are exacerbated by the cycle.
Under the demand approach, causality is very different, and the analysis is concerned with quantities in monetary terms.
Aggregate demand determines both aggregate output/ production of the economy and aggregate income. Income is then allocated to the factors of production (capital and income). The analysis might be done for levels of activity or growth; the latter relevant for the present context.
Reduced growth in government spending has led to reduced output growth and hence reduced income growth. Productivity outcomes follow from the allocation of income. There are two main processes: income is allocated between labour income and profits, and then within labour income (wages and salaries), between earnings and employment.
As seen, over the past five years in the UK, the burden of adjustment has been on wages not employment, on price rather than quantity. Low productivity is hence a residual or statistical artefact of this process. It is a consequence of how austerity-driven low growth has been borne by low wages. (Other factors may exacerbate the reduction in productivity, not least additional hours of work that are needed for some to earn a basic standard of living; in the UK VAT increases have increased inflation, which has also disadvantaged the productivity calculation.)
Plainly this wage adjustment is exceptional; separate work has shown it to be the largest and most prolonged in UK history. But the TUC analysis also shows that the majority of the adjustment was made through wages growth in several not dissimilar countries, notably Canada, Sweden, Switzerland and the US.
Why the adjustment took place in this way is obviously of interest but is not necessary to the present argument: it just did. The general case must follow from considerations of economic power and economic context. The obvious present context is the aftermath of the financial crisis; for example, the expansion of central bank balance sheets has supported firms (employment) that would normally have gone bankrupt (been made redundant) in the collapse of demand over the recession. From the point of view of power, a new factor in industrial relations is the growing importance of private equity and the changed and likely heavier pressures on workers, not least as ownership becomes more arm’s length and incentives are increasingly orientated to market value.
Note too that the demand approach normally preserves the relationship between productivity and wages, but as a matter of accounting rather than causality. Were employment growth lower, higher wages would be associated with higher productivity.
Viewed from this wider perspective, effectively in the context of national account identities, productivity is seen as a peculiar metric. In the case of a downturn in growth, the conventional calculation of productivity effectively favours employment bearing the adjustment. If profits lose so that wages and salaries gain (i.e. reduce less), then productivity is further advantaged. (The question is then, what is the benefit of high productivity on the basis of this measure? International competitiveness follows unit labour costs, but given the maintained relation between productivity and wages, these should be fairly stable, unless profit or price shifts significantly.) In the case of an upswing, matters are reversed, with wage gains more beneficial to productivity.
Fundamentally though – and to re-iterate – relations are turned back to front. Productivity is driven by demand and the allocation of income to the factors of production. Under normal conditions, the cyclicality of productivity follows from the cyclicality of demand overlaid (normally) with a greater pro-cyclicality in wages and profits than in employment. Arguably there may be some cyclical power relations at play here as well, but the logic of these reversed relations is complex to unravel!
The policy implications of the argument, on the other hand, are obvious. Policymakers are likely to continue to be disappointed if they simply wait for productivity and wages to be restored. Austerity has reduced government demand; policies are needed that will expand government demand and hence aggregate demand. Moreover, interpreting productivity outcomes as causal means a vicious circle in the context of the output-gap framework: demand-driven productivity failure is interpreted as a supply failure and more demand cuts are required. Capacity is written off as it is not used. This assessment also reads through to monetary policy, so that policy is always on the verge of tightening.
Productivity has been dominant in theoretical and policy discourse for too long. This dominance follows of course the contained role of demand since Freidman’s attacks and the emergence of monetarism over the 1960s and 70s. While the policy and theoretical position is now generally less extreme, and demand is permitted a role in short-run situations, supply considerations and hence productivity still define underlying outcomes. My belief is that the confusion around the productivity puzzle reflects the application of a conception that has little or no relevance on these time horizons and/or under current economic conditions. Productivity is perhaps an instance of the fallacy of composition: a microeconomic factor wrongly applied to macroeconomic outcomes and policy.
Understood as normally a function of demand, it is more difficult to identify when outcomes are genuinely driven by the supply side. Averaging over a number of years may report only the growth of demand modified by the income allocation. The highest productivity gains are likely to come in a cyclical upswing from a very severe recession. But this may tell us nothing about the supply side, though the tendency is for policymakers repeatedly to interpret demand-led growth in a cyclical upswing as structural. And then go to the other extreme when the upswing comes to an end; structural optimism gives way to supply pessimism. Nothing in this assessment precludes structural failings – most obviously the dysfunctional and still dominant financial system staring us in the face. Nonetheless, if correct, there is an urgent need to review capacity on the basis of a more demand-orientated perspective.