25 July 2022

Raising Keynes

A Twenty‑First-Century General Theory

Stephen A. Marglin
2021, Harvard University Press, 928 pages,
ISBN 9780674971028

Reviewer: Bridget Rosewell

This is an important book, but not an easy read even without the mathematical appendices.  Indeed,  Marglin says as much in the Introduction.  It is clearly his magnum opus and his attempt to set out his entire position on macroeconomics. It tries to set out a real synthesis of previous approaches and create a different approach to macroeconomics.  It will reward further study.  So I think my task in this review is to try and outline what the book is about and what it is saying to guide you, dear reader, in accessing it.  The book is over 700 pages long and I am sure I will have missed something.  

First, the objective.  This is to rehabilitate Keynes’ key proposition that aggregate demand may not be sufficient to support potential supply and full employment either in the short or long runs. Markets may clear and a stable equilibrium may exist below full employment. Moreover, an important part of the analysis is that Marglin proposes to show that a potential shortage of aggregate demand is not just the result of rigidities or imperfections in markets – the argument for deregulation and letting markets free – but built into the capitalist system, potentially any system, itself.  As a result, there is a continuing role for government and activist fiscal and monetary policy.

Second, the method. The first part of his discussion of method is to set out two propositions: one that Keynes’ simplest model, which he calls the “first pass model”, was restricted by having fixed wage rates in nominal term; the second that comparing equilibria in comparative statics is fundamentally misguided as it does not identify any path for moving from one equilibrium to another.  It is as if we are comparing one planet with another.  This doesn’t tell us whether a shock permits adjustment to a preferred state or not.  In illustrating this, Marglin uses a familiar – at least to whose of us trained some while ago – graphical illustration of how a shock percolates through investment, saving, and liquidity preference to derive an output outcome.  

In the “second pass” model, where wage rates can adjust, the path to a new equilibrium will depend on the relative speed of adjustment of wages, prices and output – in other words the slopes of the schedules.  There is no presumption ex ante that these will be conducive to achieving a full employment outcome and multiple equilibria are possible as adjustment proceeds.

This is set out clearly after p160. It is not a question of whether a full employment equilibrium exists but whether it can be reached.  Thus, he rejects the criticisms of Modigliani, Haberler et al. which ask only whether it exists. In the real world, it is all about the dynamics and a world in which knowledge cannot be sufficiently complete to guarantee that full employment equilibrium can be reached.  In particular, in a monetary economy views about the future must affect the present and humans have insufficient time to ‘learn’ the correct model.  This is a short form summary of his rejection of the Lucas critique.  Incidentally, I do agree with him here.  In models I have built, complete information would imply that firms would live a lot longer than they do.

Following on from this, it is no surprise that Marglin attacks the permanent income hypothesis, not only for its information requirements but, also, on the basis that for most people uncertainty (in the Knightian sense) means that the ability even to think about permanent income is risible.  Only tenured professors and other middle class professionals have any chance of living this way!

This brings us to a core part of the book which addresses the monetary economy, both interest rates and money supply. From relaxing assumptions about fixed wages, Marglin now moves to relaxing the assumption of fixed money supply.  Along the way, he argues that liquidity preference tells us nothing about the balance between investment and savings, but rather about interest rate spreads in financial markets.  Grounding interest rates requires a central bank or a limit to money supply (no fractional reserve banking).  This again drives a wedge into market clearing mechanisms which might generate a path to full employment.

And then money, reached on p497, which Marglin rightly argues is not well dealt with in the General Theory.  Properly considering money brings in debt and credit, and the consequences of inflation and deflation.  While falling prices might build credit balances, in the meantime debtors are going bankrupt as their debts can no longer be financed out of their revenues – witness the agricultural depression in the US Great Depression.  As a result, there is little willingness to invest in spite of the opportunities.  In other words there is once again no automatic mechanism to ensure equilibrium at a full employment solution.

A fixed money supply can potentially ground interest rates as can a central bank in a world of fractional reserve banking. He argues that once money is a store of value as well as a medium of exchange, the real interest rate becomes indeterminate. Not only might the system not deliver an equilibrium with lower than full employment, but the model cannot determine how much employment the system will deliver at all. This is because there is a gap between how spreads are determined and how investment and saving are determined.

What then to do?  Marglin brings in Lerner’s functional finance – the proposition that government spending fills the gap between private demand and full employment requirements.  This has in practice raised the concern of ever increasing government spending and debt, although Marglin relies more on automatic stabilisation mechanisms to manage the deficit over the medium term.  Reinhart and Rogoff get short shrift, as you might expect, although it is hard to pin down exactly what the policy prescription is, and when or at what point debt levels would become a concern.

Finally, the last section of the book looks at the long term, in much the same way as standard texts.  I always felt it odd that growth theory was an add-on – indeed in my day only reached as a graduate student. In a framework which is concentrating on dynamics it seems rather odd that growth is a separate topic to stabilisation. Marglin rescues himself by treating growth as a relaxation of assumptions about technological change and population growth.  This does seem weak and it would be helpful to integrate such changes more explicitly into the approach from the beginning.

I have tried, in order to write this review, to summarise the bones of Marglin’s arguments.  I hope I have done so fairly. Do I think he succeeds in creating a new macroeconomics and a new basis for macroeconomic policy?  Not really.  I agree that focusing on dynamics and the adjustment paths the economy takes should be central.  And I agree that thinking about how real consumers, investors and savers are able to behave, rather than positing an impossible form of rationality, should inform those adjustment paths.

Does this mean that there is a more activist role for government than that espoused in the dictum that if something is worth doing it will already have been done by the private sector?  Well, it might but I’m not sure the case is made convincingly here.  For example, government sets the parameters for tax breaks and may invest over longer horizons than private actors can.  This is mostly about supporting growth and productivity.  Consequently integrating productivity into the case for government debt seems to me to be crucial.  But like Keynes, Marglin concentrates on stabilisation and cyclical management.  That is very old-fashioned. I think.

On the one hand, he does a great job of setting out the bones of a new way of thinking about the evolution of the macro-economy.  He does a less good job of setting out how and why governments should make a difference to the outcome of that evolution.