09 October 2018

Rebuilding Macroeconomics

Chair: Evan Davis, BBC
Speaker: Prof Wendy Carlin, University College London
Speaker: Prof David Vines, Balliol College, Oxford
Speaker: Prof Simon Wren-Lewis, Merton College,Oxford

The Resolution Foundation was the venue for a review for the latest thinking on whether academic macroeconomics is in need of a rethink.  A recent double edition of the Oxford Review of Economic Policy (OXREP Volume 34) had brought together a host of prominent thinkers to consider the question from all angles, and three of them presented their conclusions to the Society.

Professor David Vines of the University of Oxford and editor of the recent OXREP issue, discussed past paradigm shifts on the way to the current orthodoxy.  Keynes’ insight was that savings did not always equal investment at full employment; this ushered in general equilibrium thinking, ultimately via the IS-LM model. The next paradigm shift, in response to the great inflation, changed content in terms of a curved Phillips Curve and supply side research, and policy in terms of the primacy of monetary policy for stabilisation.

This revolution continued through the introduction of micro-foundations and rational expectations.  While the methodological upheaval was embraced, the policy implication of treating the economy as if in constant equilibrium was rejected, with various rigidities being empirically demonstrated, leading to the New Keynesian benchmark.  In the Great Moderation, the New Keynesian Dynamic Stochastic General Equilibrium (DSGE) model became the benchmark for policy making.

The crisis highlighted 5 problems with the dominant paradigm.  First, the supply side does not grow smoothly (the problem of hysteresis).  Second, our understanding of demand was too simplistic (John Muellbauer’s recent study of housing demand was cited as an example of how our thinking needs to become more advanced).  Third, endogenous risk spillovers and leveraged lenders are not to be seen in the benchmark model.  Fourth, replacing rational expectations with adaptive, bounded and extrapolation expectations is hard but necessary.  Fifth, multiple equilibria are not sufficiently provided for in the benchmark model.  Olivier Blanchard distinguishes usefully between formal models, structural models and toy models, and the criticism is that DSGE is really a kind of intricate toy model.  The consequence is that there may no longer be a true church: the profession is about to get messier.

Professor Wendy Carlin of the LSE sketched out a possible path forward in her observation of low inflation, productivity, interest rates and capital deepening, combined with low unemployment.  How can we develop a model which can generate this outcome (or high unemployment) in the presence of an inflation targeting central bank?  She described a 2 regime model of “normal” and “demand-led” regimes, and outlining the switching between them.  The micro foundations incorporate incomplete labour and credit markets (equilibrium unemployment and credit rationing), model-consistent expectations only at the central bank and FX markets, and productivity growth only through embodied technology where investment is drive on by high and low states of expected market growth.  This can generate an economy stuck at the zero lower bound, where a floor on nominal wages at zero prevents a deflationary spiral. A positive fiscal shock can allow escape from this regime only if it can get inflation above target. 

The modelling of investment as a dynamic game with different firms and their beliefs can generate low productivity via multiple equilibria.  The last stylised fact, the potential for low unemployment, can be modelled by a changing labour market which weakens the reservation position of workers.  Springing the weak-equilibrium trap then requires creating positive expectations about the future to raise investment - but not all policy responses will achieve this.  Moreover, revived productivity growth cannot be guaranteed to raise future incomes: the last 40 years in the US labour market testifies to this.  The key to this account is the refined understanding of investment behaviour. Investment behaviour is significantly under-researched in academic macro-economics.

Professor Simon Wren-Lewis of the University of Oxford criticised the adoption of the requirement for micro-foundations.  This was born of the New Classical counter revolution of the 1970s, epitomised by Lucas and Sargent’s criticism of model-inconsistent expectations.  Methodologically, this critique was a complete success despite its limited policy influence.  The Lucas critique appealed to a new generation of economists who valued the quality of internal inconsistency - but it made the criteria for admitting equations to models (and models to journals) about fitting the theory, not fitting the data.  The consequence has been a ponderous pace of progress.

The UK continued to maintain more data-founded approaches until around 2000, making it a relative holdout.  Many of these models were richer theoretically than DSGE models (eg time-varying credit constraints) but were discarded because they lacked the necessary micro- foundations.  It seems reasonable to assume that a universe of models which contained credit constraints would have found a way to endogenise them and understand this aspect of the financial crisis better.  The end result of Lucas and Sargent then, has been to significantly slow the pace of innovation in macro economics research.

Carlin added that the current orthodoxy should only ever have been one strand of the wider research programme, but instead academic research became captured by real business cycle theorists who in turn came to staff policymaking research teams.  The discussion which followed suggested that the appetite for ecumenism among practicing economists is large and yet to be satisfied.  Many thanks to Evan Davis for hosting an inherently technical discussion with the wit and energy that kept things accessible to all.

Sunil Krishnan, SPE Council Member