Recent Posts

Wednesday, June 15, 2016

Macro Theory Strands

Brad DeLong wrote an interesting piece on the strands of macro economic thinking that one encounters. His article was a response to Noah Smith's earlier Bloomberg view article. As someone familiar with what Noah Smith calls "finance macro," I think that DeLong's description is a better fit to reality, although I would sub-divide macro slightly differently.

DeLong breaks down macro into five categories:
  1. Grifter macro
  2. Pointless (academic DSGE) macro
  3. Useful (Fed-technocratic, finance-forecasting) macro
  4. Coffee-house macro
  5. Policy macro
What a couple of these categories represent is not obvious from their names -- "grifter macro," and "coffee-house macro."

Noah Smith describes "coffee-house macro" as:
"coffee-house macro," and it’s what you hear in a lot of casual discussions. It often revolves around the ideas of dead sages–Friedrich Hayek, Hyman Minsky and John Maynard Keynes. It doesn’t involve formal models, but it does usually contain a hefty dose of political ideology.
This description is unfairly dismissive. The correct name for this category is "political economy," and it is an area of enquiry that predates macroeconomics. There may not be mathematical models, but to say that there are no "formal models" is a case of mathematical reductionism. Political economy uses the same style of argumentation as the other branches of the humanities, and has a long history within academia. Furthermore, without an underlying political philosophy, we have no way of justifying why we are studying macroeconomics in the first place.

"Grifter macro" comes from Brad DeLong:
One of its pieces–call that the second of our four components of macro–is grifter-finance macro: essentially affinity fraud to terrify rich people and get them to let you overcharge them to manage your money or simply to sell some snake oil you have to offer.
A less dismissive version of this is that "grifter macro" is political economy that you happen to disagree with, and so I do not think it deserves its own category.

Macro Theory Proper

Although I understand the thinking behind the distinction between "academic (DSGE)" macro and "useful" macro, I think we need to break the theory down by subject topic, then look at who is producing the research.

I would subdivide macroeconomics into two categories: whole economy models, and partial economy models.

Prominent examples of whole economy models include:
  • representative agent Dynamic Stochastic General Equilibrium (DSGE) models;
  • old Keynesian IS/LM models;
  • post-Keynesian Stock-Flow Consistent models.
What I call "partial economy models" is analysis of one of more aggregated sectors of the economy, without attempting to model the whole economy. To give an example of recent interest, take the U.S. Labor Market Conditions Indicator. There is a great number of research topics suggested by this indicator.
  • How do we build such an indicator?
  • How sensitive is the indicator level to the method of construction?
  • Does the indicator have predictive powers for the behaviour of other macroeconomic indicators (e.g., wage inflation, etc.)?
We need to have some whole economy model in the back of our mind when we look at these partial models -- why do we care about the conditions in the labour market in the first place, and what do we think will happen if the labour market is "overheating"? However, we do not have to agree on the whole economy model in order to discuss the research around the partial economy models.

As Brad DeLong and Noah Smith suggest, whole economy models are a dead end. Even if we think we have the correct model, there is a massive leap to relate the model to the real world. For example, as I discuss at length in Interest Rate Cycles, even if we accept mainstream models as being correct, they offer very little guidance for economic forecasting, as we do not have access to reliable real-time estimates of key variables (the output gap, the natural rate of interest, NAIRU, etc.).

The fact that whole economy macro is a dead end is why it is ignored in market research. If we want to make forecasts, we have cobble together a bunch of partial models and hope that they work. Although DSGE macro is prestigious, central banks have figured out that they are in the same boat as financial market participants. Meanwhile, many academics have a residual amount of common sense, and so they also develop partial models.

As a result, "useful macro" is mainly partial models, and the research is done using a wide variety of techniques. Academic and central bank research relies heavily on regurgitating econometric significance tests, while a lot of market (and blogger!) macro is seat of the pants chart analysis. But we are on a continuum; market research may use exactly the same econometric tools as academic research, it is just that anyone who wants their market research to be read avoids stuffing their reports with irrelevant statistics dumps.

Policy Macro

The most interesting part of DeLong's piece is the discussion of policy macro:
Policy macro is the intellectual framework that underpins the policies that the North Atlantic has followed since the start off 2010.
It is not in close dialogue with any of the others.
It has been, on the fiscal side, a complete disaster. It has been, on the regulatory side, a mixed bag. And it has been, on the monetary side, a very partial success. 
Monetary policy mainly revolves around setting the level of interest rates (and discussions of the magical powers of changing the size of the monetary base). Since it revolves around changing one numeric quantity (the policy rate), and seeing how it affects other economic variables, it is amenable to mathematical modelling. Unfortunately, the effectiveness of monetary policy was oversold, and there is not a whole lot of evidence that low interest rates are doing economies any good.

The other parts of policy are pretty much a disaster area, and I see little hope that this will change.

Fiscal policy should be somewhat amenable to analysis, but the economic mainstream turned its back on realistic analysis for political reasons. Whether or not the policy prescriptions of the Old Keynesians were correct, Functional Finance is the starting point for the analysis of fiscal policy. The politically-motivated attack on Functional Finance meant that policymakers are stuck with only magical thinking about fiscal policy: we have to avoid increasing government debt levels, because "bond vigilantes" might punish us. This is as useful a theoretical framework as is blaming diseases on witchcraft.

Other aspects of policy (such as regulatory, tax, and trade policy) are in even worse shape. The most charitable description of economic policy in the developed countries in these areas is "smokestack chasing" -- making the economic environment favourable for large corporations, so as to attract them to your country. Although it is relatively easy to argue that such policies are not in the national interest, they are unlikely to change as a result of realpolitik considerations. Modern corporations are extremely focused on delivering short-term results, and so they fit perfectly with the chronic short-termism of policymaking elites.

(c) Brian Romanchuk 2016


  1. The funny thing is that there are a sufficient number of developed countries that there are naturally lots of experiments going on. Of course, there are also an absurd number of variables. One post I read recently was making an example of the policies of Venezuela. One commenter responded something like: "The problem isn't with the policies, exactly. The problem is that Venezuela is filled with Venezuelans." Yes.

    I think we spend way too much time thinking/writing/theorizing about issues when we should be attempting to solve obvious problems in reasonable and straightforward ways. Do we KNOW that a change will improve things? No. But it's just policy. If it doesn't work well, change it again. This is how progress is made.

    1. We have a good idea of the broad implications of policies. (The only real question marks are things like a Jobs Guarantee or Income Guarantee, as we have only limited large scale tests.) The problem is that voters keep electing politicians who want to maintain the status quo.

      However, our understanding of the current system is hardly perfect. You can argue we know roughly what will happen, but we still have a hard time doing things like forecasting inflation.

    2. I'm not sure I agree. I wonder what kind of policies you're referring to. Even something as simple as a minimum wage hike is mired in controversy and finger-pointing.

    3. My view is that *sensible* people have a rough idea what will happen; the details are unclear. We know that some jobs would disappear if the minimum wage goes up by more than a tiny amount; that's the whole point. However, the net effect on jobs is less clear; the higher paid workers will generate greater demand, which pays for the higher wages. In any event, the net job changes for reasonable changes to the minimum wage are dwarfed by other cyclical effects. We can use the exact same argument that the mainstream uses for the ineffectiveness of fiscal policy - stabilisation policies (monetary, automatic stabilisers) will bring the economy back to "normal" employment, regardless of what happens to the minimum wage. (One might argue that if the minimum wage increase is inflationary, then unemployment will be kept higher by the central bank, but this is offset by higher productivity.)

      Of course, none of this common sense is reflected in the political debate. There is either massive screaming about details that we do not know about, or ideological hacks just make stuff up. Since said hacks are paid very well to be hacks, no advances in our economic knowledge will make them go away. I was excluding them from the "we" when I wrote that we have a broad (I perhaps shuld have used "rough") idea of the effect of policies.


Note: Posts are manually moderated, with a varying delay. Some disappear.

The comment section here is largely dead. My Substack or Twitter are better places to have a conversation.

Given that this is largely a backup way to reach me, I am going to reject posts that annoy me. Please post lengthy essays elsewhere.