Hysteresis vs Full Recovery by creating a Boom?

Simon Wren-Lewis has an important post. Read it first. If he's right then what looks like hysteresis from the recession is due to deficient Aggregate Demand. But is he right? This post is my attempt to figure it out.

I want to start with a very simple model with no hysteresis -- where recessions have no permanent effects. The two inputs are labour and land, with a fixed stock of land, and constant technical knowledge. A fall in AD causes a recession with unemployed labour. Then wages fall, but AD remains too low. (Whether the fall in wages would cure or worsen the AD deficiency depends on the monetary policy regime, and your intuition on this question is very probably wrong since very few people can think intuitively about general disequilibrium with monetary exchange, and it's off-topic for this post.) The fall in wages causes firms to switch to more labour-intensive methods of production, but since AD remains too low, output remains too low, so the only effect of lower wages is to replace unemployed labour with unemployed land.

I could imagine something like this happening in an economy where tenant farmers ("farms" = firms) rent labour and land from workers and landlords. If wages are stickier than rents, then deficient AD causes unemployed labour. If rents are stickier than wages, then deficient AD causes unemployed land. It is not enough to look for unemployed labour to tell whether there's deficient AD; you need to look for unemployed land as well. If both labour and land are at "full employment" (whatever that means in practice) then there is no deficiency of AD, the economy has fully recovered from the recession, with output back to the original level.

(But it is hard to imagine this happening in an economy where farmers own their land, and only hire labour. Would you leave some of your land unemployed because you are charging yourself too high a rent?)

Now let's go Dutch. Change the model by assuming that labour and land can produce new land. The "land" is now capital, and production of new land is investment. And we can also assume that land depreciates over time.

Now read Simon: "Central bankers might say that they would still know there was inadequate demand because surveys would tell them that firms had excess capacity. That would undoubtedly be true in the immediate aftermath of the recession, but as time went on capital would depreciate and investment would remain low because firms were using more labour intensive techniques. The surveys would become as poor an indicator of deficient aggregate demand as the unemployment data."

The temporary recession causes a fall in investment in producing new land, so even when the unemployment of labour and land ends, the stock of land is lower than it would have been otherwise, and output is lower than it would have been otherwise. There's hysteresis.

Does this mean that AD is still "deficient"?

Consider a thought-experiment. There are two parallel growing economies. They are identical in all respects, except that the Alpha economy figured out how to produce new land and started growing 10 years earlier than the Beta economy did, so in any given year the Alpha economy has a larger stock of land, and larger output, than the Beta economy. Then the Alpha economy has a recession, investment falls, which allows the Beta economy to catch up with the Alpha economy. When the Alpha economy is back to full employment of labour and land, both economies are identical. If we say there is an AD deficiency in the Alpha economy, shouldn't we also say there is an AD deficiency in the Beta economy? Both, or neither? Aren't bygones bygones? What is it that makes the "memory" of the recent recession in the Alpha economy different from the "memory" of the later development of the Beta economy?

The answer to this question might be "inflation". If both economies had been targeting (say) 2% inflation, then it is possible that inflation would be lower than 2% in the Alpha economy, because of the recent recession, but still at 2% in the Beta economy which had no recession. And if you believe that 2% inflation is the right target (or even if you don't believe that, but nevertheless believe that an announced target is a target and should be hit), then it makes sense to say that the Alpha economy has deficient AD while the Beta economy does not. And if that is what Simon is saying it's an OK thing to say. A target is a target, dammit, and should be hit!

But you can also read Simon's "The ultimate arbiter of whether there is demand deficiency is inflation. If demand is deficient, inflation will be below target." as saying something more than this. And if so, I'm not getting it.

Comments

[Brett Reynolds]

I think the spelling is hysteresis.

[Nick Rowe]

Brett: Damn! Thanks! Will edit.

[Roger Sparks]

Hmmmm.

People all eat about the same amount if measured in calories. On the other hand, between persons, there is a tremendous difference in the quality of food consumed.

People need at least some shelter. Between persons, there is a tremendous difference in the quality of the shelter.

People certainly enjoy transportation. Between persons, the quality varies from walking (on your own two feet) to chauffeured cars and better.

If you have more things, you need more room to store them.

Now we come to taxes. In general, the better the quality and the bigger the quantity, the more the tax. Let the rich pay for the poor.

With this general scheme in mind, when we have "deficient demand", what are we really talking about? Has there been a collective decrease in expectations? Has there been a genuine oversupply of quality and quantity, causing over-migration of people into unaffordable tax realms?

If either of these two maladies are the cause of deficient demand, we could see lingering effects that we might define as hysteresis.

[Nick Rowe]

Roger: an AD deficiency is none of those things. It means that the stock of money is too small, or is circulating too slowly, to buy and sell all the things that people want to buy and sell. A good metaphor would be a shortage of plastic shopping bags, where everyone is hoarding shopping bags, and you can't sell anything without giving the buyer a bag to put it in.

[Thaomas]

Is it relevant that Alpha was "targeting" 2% inflation? Would it not be important if over the cycle an average 2% inflation had been achieved?

[John Monz]

Nick - thanks for a thought-provoking post. Part of the problem is that "deficient" implies a shortfall relative to some desired state, but it is not obvious what that desired state is. Some "optimal growth rate" of the economy, presumably, but how do we define that? Maybe the 2% inflation is serving as a kind of proxy for the desired state in Simon's post. Whether that is a useful proxy or not is a different question.

[Nick Rowe]

John: thanks. I think that's right. Though I would maybe add "feasible/sustainable" (or something) to your "desired state" (though that's presumably implied).

[Nick Rowe]

Thaomas: I found your comment in the spam filter, which explains why it was delayed. Sorry about that.

Well, I assumed that both Alpha and Beta had been targeting 2% inflation to try to make the model relevant to what Simon says in that sentence I quoted near the end of the post. And I expect it depends on whether they are targeting 2% for every year, or only over the cycle.

[lawrence e stirtz]

Interesting but useless since you omit factors to simplify. As usual we do not include all of the events history so the prediction is flawed. The good news is you cannot know them all anyway so I put it in category of sort of a hobby.

[Majromax]

@Nick Rowe:

> What is it that makes the "memory" of the recent recession in the Alpha economy different from the "memory" of the later development of the Beta economy?

Nothing, but we don't expect the Beta economy to remain behind forever. Developing countries are supposed to experience catch-up growth precisely because they're ostensibly more fertile places for investment.

This is also more than just a pure thought experiment. Since the last recession, the US has seen stagnant and declining capital intensity per hour worked, whereas the previous trend was notably positive.

The deficiency alleged by Wren-Lewis (as applied to the US, I haven't looked at UK statistics) is an ongoing deficiency, characterized by continuing low investment rates. Even if an economically-leading country can't "make up" for a period of foregone investment, that doesn't explain why it would continue to have low investment rates even after the aggregate demand deficiency is allegedly fixed.

[Nick Rowe]

Majro: if we embed my "model" within a Solow Growth model, which has catch-up, then yes; without the recession in Alpha, Beta would still have caught up (asymptotically) with Alpha eventually. But with the recession in Alpha, it catches up immediately.

It's puzzling why investment would be lower in Alpha than in the counterfactual conditional Beta after Alpha is back to "full employment" (just like Beta). One possible reason is AD uncertainty. Investors in Alpha, having just lived through a recession, are less confident that it won't happen again. (I think I did a post on something like this once). But I'm not sure that creating an AD boom would be the best way to restore their confidence.

[louis]

A helicopter drops bales of money down people's chimneys at night. This should cause inflation, right? What's the mechanism? Everyone who received extra cash feels wealthier and plans to buy more at current prices. If agg demand equaled potential supply, the increased demand for goods should force prices to rise so quantity demanded equals potential supply again. What if we conducted a helicopter operation and prices failed to rise? Wouldn't that mean real GDP is rising and therefore we were/are suffering from deficient AD?

[Nick Rowe]

louis: suppose an increase in AD did increase RGDP in the Alpha economy, and we therefore say that AD was deficient. Wouldn't it also have the same effect in the Beta economy, and wouldn't we therefore say that AD was deficient in the Beta economy too, even though (by assumption) it never had a recession?

[louis]

I think the alpha and beta setup is flawed, because it presupposes hysteresis. A's economy is smaller than it could have been by virtue of having invested less in land. If instead of a recession, an asteroid hit A and destroyed some land, we would have the same outcome. There's no reason to suppose an increase in NDGP would mean an increase in RGDP in that scenario. It would be a wonder if it did.
Better to use an example that better captures the dynamic Wren-Lewis set out, perhaps with differences in relative use of the factors of production.

[Nick Rowe]

louis: I thought my example *was* capturing the dynamic that Simon set out? The fall in wages during the recession results in a more labour-intensive production method, compared to what would have happened otherwise.

[Nick Edmonds]

If you extend the model slightly to include public debt, you may get the memory needed.

Households in the Alpha economy may have accumulated greater savings (consisting of either land or public debt). When the recession strikes, the level of land comes into line with that of the Beta economy, so with higher savings the public debt is higher. An increase in AD in Alpha may then have more scope to divert savings from public debt to land.

[Nick Rowe]

Nick E: I need to think about that one. It's plausible (I think) that debt could be higher in Alpha. Not sure if the rest follows.

[louis]

Maybe I'm not getting it, but Simon seems to be talking about a suboptimal switch to a more labor intensive mode. Ie economy could produce more output with a more land intensive mix. In your example there's no reason to believe that's the case... And if it is the case, the condition preceded the recession and is not a result of it.

[Nick Rowe]

louis: the economy can't switch to a more land-intensive mix unless it increases investment to produce more land. And if the Alpha economy can do that (as a response to increased AD) why couldn't the Beta economy do the same?

[louis]

Reread both posts and I see more of where you are coming from. The assumption Simon made was that labor was utilized but capital was slack. You would see that as part of the recession and a sign AD is still under utilized. He goes further and says measures of utilizstion will be less accurate as time goes on, including for reason that we could be underinvesting in capital due to weak demand. Your example would be closer to his case if after the recession A and B had same GDP but B was investing more in capital than A. I see though that calling this deficient AD feels a little uncomfortable.

[Nick Rowe]

louis: yep. (Though it's not obvious why investment would be lower in Alpha than in counterfactual Beta, unless it's the increased uncertainty about future AD from the experience of the recent recession.)

[Majromax]

> Though it's not obvious why investment would be lower in Alpha than in counterfactual Beta, unless it's the increased uncertainty about future AD from the experience of the recent recession

Investment would be expected to be lower in absolute terms from a less capital-intensive economy. If population growth is exogenous and constant in the two economies (say 1%), if Alpha has a stock of $100 land per person-year of labour and Beta has a stock of $200 land per person-year of labour, then a constant and equal growth rate for the two economies requires that Alpha invest $1/person in land while Beta invests $2/person.

If the central bank of Alpha credibly promised that aggregate demand would "make up" for the recession and return to the pre-recession expected trajectory, then Alpha would invest something more than $1/person in land and would at least partially recover its capital intensity, with the exact rates depending on the price-flexibility of capital (with inflexible investment resulting in more inflation and less real-terms capital formation).

Wren-Lewis's argument is then that capital is underutilized and is effectively very elastic, so an increase in aggregate demand will result in increases to real investment and output in much the same way that AD increases will result in more labour utilization if unemployment is high.

[Majromax]

As an addendum to the previous, the difference between post-recession Alpha and Beta is that during the recession Alpha left some easy investment "on the table," whereas Beta did not. For some time, Alpha can make use of these missed opportunities.

Back to the land example, imagine if land investment required expensively raising land from the sea and inexpensively tilling that land. Immediately post-recession (by prescription) Alpha and Beta had the same level of employment and the same amount of tilled land, but Alpha would have had more raised but untilled land.