How Not to Do Macroeconomics, Part II
Simon Wren-Lewis has written a response to my critique of Eggertson & Mehrotra's (EM) model of secular stagnation, as well as my subsequent jabs at macroeconomics as a whole. He charges that my comments "reflect a misunderstanding, often shared by non-economists, about what much academic macromodelling is designed to do". He believes that simple models like this help us understand more complex models and reality. In principle, I actually agree with this view: building models starts off simple and abstract, and, where necessary, you add complexity as you go. However, this doesn't mean that every simple, abstract model is automatically relevant, and in the case of EM there are good reasons to reject their framework from the start.
The problem with the model is quite simple: it is unable to explain its purported object of study. Secular stagnation is loosely defined as a period where there is insufficient 'real' investment to increase aggregate demand and employment, meaning the economy has to rely on financial bubbles to do so. However, EM contains none of the features of the economy that are integral to the this process. They have abstracted from investment, financial markets and banks, so by the necessity of their own simplifications, they are forced to impose arbitrary assumptions - such as young peoples' exogenous limit to borrowing - in order to get the kind of behaviour they want. This is not a model of how capitalism, the financial sector, or certain institutions or policies 'cause' secular stagnation. It's a model that has the outward appearance of secular stagnation forced onto it through careful choice of assumptions.
Wren-Lewis (WL) defends the model on the grounds that it is supposed to isolate particular mechanisms, so we can understand these mechanisms before transposing them into more complex models. However, if the mechanism is not relevant in the real world, then why would we want to isolate it in this first place? EM give us no reason to believe intergenerational lending is the driving force behind secular stagnation, and do not even make any falsifiable predictions concerning whether that could be the case.
We may, as WL suggests, be able to make the model more complex by adding things like hysteresis, sticky wages and financial frictions. However, this is a double-edged sword: on the one hand, if the underlying mechanism driving secular stagnation in EM is wrong, it will remain wrong despite adding complications. On the other hand, if the underlying mechanism does not survive the transition to the more complex model, then what was the point in the simple model, and how is it still a model of secular stagnation? All models may start simple, but they don't all start on the right path, and they do not always simplify in a relevant manner.
How to do macroeconomics
Perhaps a positive outlook is in order. In the comments on his blog, WL asked me "if this paper had been written by heterodox economists, how would it have been different?" First, I would argue that this is the wrong way of framing the issue: though heterodox economics is fragmented, one unifying theme is the search for 'rule-based', generally applicable theories; as opposed to the 'case-based' approach economists use. Hence heterodox economists (and I stress that such a blanket statement should be taken with a pinch of salt) would not look to create a model of secular stagnation per se, but would hope that their existing frameworks could already account for it. And they'd be right.
There are two major heterodox theories which help to understand both the 2008 crisis and the so-called period of 'secular stagnation' before and after it happened: Karl Marx's Tendency of the Rate of Profit to Fall (TRPF), and Hyman Minsky's Financial Instability Hypothesis (FIH). I expect that neither of these would qualify as 'precise' or 'rigorous' enough for mainstream economists - and I've no doubt the mere mention of Marx will have some reaching for the Black Book of Communism - but the models are relatively simple, offer an understanding of key mechanisms and also make empirically testable predictions. What's more, they do not merely isolate abstract mechanisms, but form a general explanation of the trends in the global economy over the past few decades (both individually, but even moreso when combined). Marx's declining RoP serves as a material underpinning for why secular stagnation and financialisation get started, while Minsky's FIH offers an excellent description of how they evolve.
Minsky's FIH, first formulated in 1974, argued that "stability is destabilising": tranquil macroeconomic periods will sow the seeds of their own demise. This is because financiers who invest prudently in stable times will likely see their investments pay off, and therefore will consider it sensible to take more risk with their next investments. If these investments pay off too, then over time the financial sector will become progressively riskier and more speculative, until it becomes so speculative it collapses. Minsky divided this process into 3 stages: hedge finance, where business has sufficient cash flow to cover its loan repayments; speculative finance, where business only has enough finance to cover interest payments; and ponzi finance, where business must sell assets/deplete its reserves in order to pay interest.
As many have noted, this process describes the past few decades in capitalist economies eerily well. Following the stability of the post-WW2 boom, financial institutions began to ratchet up leverage, debt-to-income ratios rose rapidly, spawning what Minsky christened 'money manager capitalism'. Subsequently, the global economy experienced a series of increasingly dangerous crises, including the S & L crisis, the dot-com bubble, and the 2008 financial crisis. Regulatory barriers initially limited this process, rendering the former crisis less severe, but their gradual elimination - accompanied by increasingly risky market 'benchmarks' against which managers measured their performance - had the precise effects predicted by Minsky, and the 2008 crisis itself was a classic example of Minsky's Ponzi finance. Clearly, it was not only financiers who were subject to this process: regulators and politicians began to believe in the new era of financial stability, and economists themselves famously christened the era the Great Moderation. Minsky's hypothesis about myopic expectations causing excessive risk taking was perhaps even more encompassing than he imagined.
Marx's Labour Theory of Value (LTV) and his subsequent theory of crises may be much maligned - generally based on fruitless misconceptions and disagreements over the word 'value' - but we can easily eschew such 'metaphysical' concepts and just focus on the empirical content of the theory. In essence, Marx's theory boils down to the proposition that labour-saving technology will lower prices and hence reduce the economy-wide Rate of Profit (RoP) - offset, perhaps, by depreciation, demographic change or whatever else - but demonstrating a clear, persistent fall in the long run. The Marxist economist Andrew Kliman has demonstrated that this tendency is consistent with evidence from the US:
Blogger Michael Roberts has also demonstrated that this tendency is robust to key choices of how exactly to measure 'the' RoP.
According to this theory, the post-WW2 boom was characterised by a high RoP following the massive destruction of capital value after the Great Depression and WW2 (capital is a cost: the less there is, the higher the RoP). However, following Marx's theory, a declining ratio of labour to capital input has caused the RoP to decline over time. Since firms' investment closely follows their realised RoP, a lower RoP means a lower rate of investment, causing overall growth and employment to lag. Furthermore, faced with this declining RoP, firms have been less willing to partake in 'normal' investment, instead relying on asset bubbles and other speculative lending to make money, resulting in financial crises. I don't know about you, but to me this seems like a pretty impressive explanation of secular stagnation - especially considering Kliman's book was written well before Larry Summers made his speech.
I'm not going to suggest mainstream macroeconomics is incapable of offering a plausible explanation of secular stagnation. However, there are a vast number of macroeconomic models, each offering ex post rationalisations of one thing or another, but none seeming to get any closer to a general framework that can understand and predict these processes ex ante. The framework is sufficiently flexible to account for any eventuality with a few clever tweaks, and is therefore unfalsifiable in practice. Conversely, models like the ones I presented above, though less precise and mathematical, offer ways to identify key mechanisms to make ex ante judgements about the economy. Aside from an a priori commitment to a certain brand of mathematics, and an ideological aversion to certain thinkers, I don't see why these kinds of models should not at least have a seat at the macroeconomic table.