Recent Posts

Saturday, April 4, 2020

Macro Models Aren't Useful Now, And That Is Perfectly Fine

There has been a small flurry of publications by neoclassical economists attempting to fit a pandemic into standard frameworks. This is what to be expected, as neoclassical models are frameworks designed to maximise the amount of publications over time. However, aggregated models are not particularly useful right now. This is true both for neoclassical as well as traditional heterodox macro models. The reason is that they do not offer much insight into either forecasting, nor are they useful for policymakers. That will change, but we are not there yet.

I will quickly discuss the two main justifications for looking at macro models.


Why do we care about forecasting the macro economy?
  1. We are trading macroeconomic outcomes, like breakeven inflation, or the policy rate. (The policy rate is a second level case, as we are predicting how central bankers will react to aggregates.)
  2. We are interested in forecasting what variables related to macro variables will do.
The first is straightforward, but it is very niche. Meanwhile, if we want to predict what will happen to inflation, we will need to do a bottom up analysis (link). Each component of CPI will tear off in a different direction, and there are a lot of judgement calls where each will go.

The second is more important. We do not normally care what happens to GDP, rather we want to infer what will happen to what we are actually interested in based upon a scenario for GDP. Since many variables have a relatively stable relationship with aggregates, this works. 

The other macro worry will be things like defaults. However, default risks are going to be hard to infer from aggregate data; we would need very granular data to get a handle on distributions. Meanwhile, the risks depend upon the length of lockdowns, as well as the effectiveness of aid programmes. We can only guess about that.

Aggregate information will be more useful once the emergency situation has stabilised. Even then, much will depend upon what industries can re-start safely, and which will not. That is going to have to be determined by bottom up analysis.

Policy Guidance

Countries like Canada and the United States were caught flat-footed by the virus, and are in a emergency mode. The prospect of excessive casualties and a collapsed health care system was politically unacceptable, and soft lockdowns were put in place.

Even if we accept there is very large uncertainty about medical outcomes, we are not dealing with magical creatures that can move at will -- the virus is mainly transmitted from person-to-person. (Animal transmission is presumably possible, but animal travel speed is relatively slow.) For larger countries like Canada and the United States, we can look at virus transmission on a region-by-region basis

Once the initial confusion subsides, the authorities will be in a position to restrict regional travel, and then deal with the problems on a regional basis. The Province of Quebec has already started putting in place travel barriers within regions, and U.S. states have been also implementing border controls. Lockdowns will be relaxed on a regional basis to allow critical infrastructure and goods production, with agriculture and food processing being extremely pressing.

This relaxation process is entirely a question of the medical situation, geography, and an analysis of the importance of goods production. The more regions with few cases there are, the more aggregate production drops back. The mix of production will depend on which regions are up and running. The main policy decisions will revolve around where to allocate floating resources, and that would depend upon strategic priorities. For example, it does not matter that re-opening the film production industry might add more to GDP than picking fruit; the food must flow.

Otherwise, the question is how to keep non-essential workers and firms in stasis. Since nobody is worried about financial constraints, the policy questions revolve almost entirely around the implementation of aid policies. That is, the dollar amounts are not important, rather how those dollars are distributed. 

Once a vaccine is widely available, we will likely face the macro issue of getting unemployed people back to work, which is a traditional macro story. However, given the uncertainty around the future state of the economy, macro modeling work now is offering no actionable information.

(c) Brian Romanchuk 2020

No comments:

Post a Comment

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.