In chapter 14 of Debunking Economics, Steve Keen walks us through the macroeconomic model he has developed in recent years, and discusses the implications its conclusions might have for policy. The model is in its early stages, and Keen himself says there are “many aspects of the model of which [he is] critical.” Nonetheless, it is a promising start to developing an alternative to the dominant DSGE method.
Keen’s is a model of a pure credit economy, with three aggregated agents: workers, firms, and bankers. Instead of focusing on preferences, individuals and market clearing, it focuses on the flow of funds between different sectors. Bankers create their own money in the form of loans*, which at this stage they are only allowed to lend to firms. The firms pay the workers and the interest on the loans, whilst the bankers and workers consume the output of the firms.
The crucial sector here, is, of course, banking, which few neoclassical models include explicitly, as they believe finance plays the role of intermediation between savers and borrowers. However, in Keen’s model the banks are central. He disaggregates them into several accounts: a vault in which to store notes; a safe into which interest is paid and out of which bankers are paid; a loan ledger; firm deposits and worker deposits. The flows between the various agents and accounts are then determined by some arbitrary coefficients, which Keen uses simply to determine whether the model will ‘work’ (i.e. not break down). Each flow (e.g. wages; consumption) is determined by a constant times a stock (e.g. firm’s deposits; worker’s deposits).
This is the point at which economists might scream ‘Lucas Critique,’ but Keen’s comment from an earlier chapter, that is absurd to suggest that any change in policy will have the effect of neutralising arbitrary parameters, applies. Furthermore, there is no theory that is policy-independent, so whilst we must examine the relationship between policy and reality, we cannot render our models immune to it by micro founding them. In any case, the model is in its early stages, and there is plenty of room for adding complexity.
Keen uses this basic model to explore what effect bank bailouts will have. Quelle surprise, bank bailouts have the effect of increasing loans slightly, and benefitting bankers, but don’t do much for the real economy. Conversely, bailing out firms and workers creates a better result for everyone except…the bankers! A small data point can be found in support of this in Australia, where the government bailed out everyone over 18 with $1000, and the economy has performed better than those where the banks have been bailed out. Obviously there are a multitude of conflicting factors, but Keen’s hypothesis does not seem at all unreasonable when taking recent events as whole.
The interesting thing about Keen’s model is that a ‘Great Moderation’ and a ‘Great Recession’ are simply two parts of the same debt-driven process, rather than a ‘black swan’ or some other such event. Debt to GDP rises exponentially in a period of relative tranquility, and this is followed by a huge crash and mass unemployment. A substantial part of this difference from the core DSGE models is created simply by adding banks as explicit agents.
Keen has since developed his model further – he has included ‘Ponzi’ lending, sticky wages/price (which actually stabilise the economy) and a variety of other factors. There is plenty of scope for adding more to the model, such as an exogenously set interest rate with a central bank, but for now the core alone seems to be able to generate behaviour that closely resembles that of a capitalist economy: one prone to cyclical breakdown and intermittent financial crises, not due to any particular ‘friction,’ but due to the inherent characteristics of the system. This should be enough to get any empirically driven economists to pay attention, whether they agree or disagree with the mechanics of the model.