Hacker News new | past | comments | ask | show | jobs | submit login
Can negative supply shocks cause demand shortages? [pdf] (nber.org)
45 points by wallflower 30 days ago | hide | past | favorite | 40 comments



I wonder how this plays out in the real world. Restaurants have suddenly become unavailable, my demand for them has dropped precipitously, they become available again, but I don't see myself reverting to my prior demand for years, if ever.

Abstract:

We present a theory of Keynesian supply shocks: supply shocks that trigger changes in aggregate demand larger than the shocks themselves. We argue that the economic shocks associated to the COVID-19 epidemic—shutdowns, layoffs, and firm exits—may have this feature. In one-sector economies supply shocks are never Keynesian. We show that this is a general result that extend to economies with incomplete markets and liquidity constrained consumers. In economies with multiple sectors Keynesian supply shocks are possible, under some conditions. A 50% shock that hits all sectors is not the same as a 100% shock that hits half the economy. Incomplete markets make the conditions for Keynesian supply shocks more likely to be met. Firm exit and job destruction can amplify the initial effect, aggravating the recession. We discuss the effects of various policies. Standard fiscal stimulus can be less effective than usual because the fact that some sectors are shut down mutes the Keynesian multiplier feedback. Monetary policy, as long as it is unimpeded by the zero lower bound, can have magnified effects, by preventing firm exits. Turning to optimal policy, closing down contact-intensive sectors and providing full insurance payments to affected workers can achieve the first-best allocation, despite the lower per-dollar potency of fiscal policy.


They call this demand destruction.

It also happens when oil prices get high and companies invest in oil saving technology that permanently reduces their demand.


> Restaurants have suddenly become unavailable, my demand for them has dropped precipitously, they become available again, but I don't see myself reverting to my prior demand for years, if ever.

That's not considered a demand shortage in this context, because you're not necessarily holding on to more cash (the medium of exchange) simply because you've stopped going to restaurants. You might be dispensing of your cash in other ways.[0] It's a negative supply shock that's confined to the restaurant sector.

[0] (Yes, there is a multiplier effect but it goes both ways. The multiplier effect is a result of agents trying to keep to their preferred level of cash holdings while facing an outside change in the amount of monetary spending. It's really a "hot potato"/"frozen potato" effect: if agents are suddenly giving you more cash you'll want to spend the excess, and if no one is giving you any money you'll seek to restore your balance either by seeking cash from others or by curtailing your own spending.)


I'm not an economist, but I actually am holding onto the cash (putting it in savings). Yes, I'll probably ultimately spend it on something else but not for decades.


Unless you're literally buying government bonds (that are yielding very low and sometimes negative returns) your savings will still translate to increased spending elsewhere.


Well, maybe. There isn't a hard requirement that a bank will invest the money; excess reserves are currently over 3 trillion.

https://fred.stlouisfed.org/series/EXCSRESNS

The Fed even pays a small amount of interest. IOER is currently 0.1%

https://fred.stlouisfed.org/series/IOER


I googled for a bit and wasn't able to find a definition for "demand shortage" specifically, so I can only guess at what it means, versus the intuitive idea that "restaurants expected a certain level of demand, and those expectations have been dashed".

Does anyone understand the exact point being made?


I think the established terminology is more like "demand shortfall". In a general (non-sector-specific) economic context, this almost always refers to a fall in demand for all goods or services that introduces severe frictions into the workings of the market mechanism itself, and that can thus be "solved" in a comparatively straightforward way by addressing these frictions.

The "liquidity constraints" discussed in the abstract may be one such friction, but it's quite common for agents to be liquidity-constrained or unwilling to incur debt even irrespective of any negative supply-shock, so it's not quite clear to me what points the paper is trying to make.


Whelp, I guess that's why I call myself an amateur. Thank you for your attempt. :-)


How so?


This is a quick sum of the study's basic intuition:

> When workers lose their income, due to the shock, they reduce their spending, causing a contraction in demand. The question is whether this mechanism is strong enough to cause an overall shortfall in demand.

Answer:

> Positive, demand may indeed overreact to the supply shock and lead to a demand-deficient recession.

Suggested solution:

> The optimal policy to face a pandemic in our model combines a loosening of monetary policy as well as abundant social insurance.


> loosening of monetary policy

The federal funds rate is zero. So, direct injection of money seems the only way forward. Direct injection failed massively in the last recession where money was given to the markets but failed to reach the consumers. There is a lesson to learn there.

Europe approach in the current situation is to pay companies to keep citizens employed in combination with it’s already existing social protections. USA approach has been directed to assure companies capital to keep them for failing. The pandemic is an awful situation but a very interesting economic experiment. Time will tell the pros and cons of the different approaches.


Monetary loosening did not fail post-2008 recession, except to the extent that it wasn't really tried in the first place. Whenever aggressive monetary loosening was done, even by unconventional means, the market responses in the forex, stock and even in the bond markets were very clear. (Contrary to common misconceptions, loose monetary policy should not be expected to cause low rates in e.g. the bond market; low market rates occur when money has been too tight, and interest-rate based monetary policy works like stabilizing an unstable system, so pushing market rates higher requires setting the policy rate even lower!)


That I think was his point: we did see large responses in asset prices, but this excess money going into the forex/stock/bond markets did not trickle down into the labor market like conventional economic theory assumes it would. Instead of stimulating the real economy, it led to asset bubbles and widening income inequality.


> Instead of stimulating the real economy, it led to asset bubbles and widening income inequality.

Yes. I guess that I was not clear enough.

“ Wages in California have grown fastest for earners at the top of the income distribution. Wages for those in the bottom- and middle-income earners remained stagnant up until the recession and dropped until about 2014. As in the national data, low- and median-wage workers in California finally experienced wage growth that outpaced inflation starting in 2014. This was in part due to labor market tightening, but also due to state and local increases of the minimum wage.” https://irle.berkeley.edu/the-post-recession-labor-market-an...

Until 2014 there were no wage growth for bottom-middle income workers (California). That is a long fall. Meanwhile the markets recovered way faster.

We see this pattern repeating in this current crisis that just is starting. There has been a fast fall and fast asset valuation. But, employment has fallen fast but is recovering slower. It’s too soon to get to conclusions as is going to take years to see how the economy evolves. But, there is always lessons to be learned on how to improve the economy to all social strata from past mistakes.


Excess money does not go "into" markets, it flows through markets and into private-sector balances. And excess money is not what you think it is anyway: when bond-market returns are low due to tight monetary policy in the past, the demand for cash and cash-like instruments is sky-high and the monetary authorities need to create even more excess money to satisfy it. This is pretty much non-negotiable: if you want to avoid this kind of conundrum, you should avoid tight policy in the first place. (High asset prices are simply the flip side of low returns for bond-like instruments.)


So that's all technically correct, but the point I'm making is that as that excess money is flowing through asset markets, it's ending up on the balance sheet of private-sector actors who have little need to spend it on a day-to-day basis anyway. In Keynesian terms, money is pooling at precisely those firms who have a low MPC (which is not coincidental: it pools at those firms because they have a low MPC, because pooling indicates a high MPS). Thus the government needs to inject much more money through these markets than it expects, because the Keynesian multiplier is low. In the process, it inflates the paper wealth of these firms significantly more than it actually stimulates the real economy, something we saw from 2010-2020.

I think the OP is alluding to the strategy of giving money directly to consumers, and particularly poor consumers (who tend to have a high MPC). This limits the pooling effect: it still happens, but it can't happen until the money has circulated throughout the economy a few times, which increases the stimulus effect for a given amount of inequality generated. As an added benefit, the effects are more immediate, which helps the central bank understand the consequences of its actions more rapidly and not have to act 18 months or so in advance of when the stimulus trickles down to real production.


If you want to do this kind of policy by directing money flows in a non-neutral way, the highest-impact policy by far is subsidizing investment. Investment spending reacts very aggressively to any change in forecasted returns, far more so than would be implied by differential MPC. (A policy of giving more money to the poor is highly desirable for other reasons; e.g. it directly addresses liquidity constraints. It's sensible to regard these as distinct issues, however.) At the moment, the Fed is paying interest on excess reserves, which is quite bad if you think there's too little demand and we should have more investment. That's one way of saying that policy might have been too tight even post-2008.


Yeah it a lot of math to propose a link between supply shocks (where everyone to get fired) and demand (via employment).

Kind of like in 2008 when the macro guys woke up to the idea of a financial system as important and outside the existing model.


The pandemic is not primarily a supply shock. It is primarily a demand shock. People aren't going outside.


It is an oscillation between a supply shock and a demand shock, and back again. Multiple so-called "bullwhip effects" stacked on top of each other, all across the various sectors in the economy.

It's worth remembering that from the Western POV, the pandemic absolutely started with a supply shock - a big chunk of Chinese manufacturing essentially shut down for two months, and a lot of the usual cargo thus never sailed. This not only caused raising prices on all sorts of imported goods two months down the line (it takes as long for the container ships to reach Europe and the US), but also caused export problems - less incoming containers = less containers available to ship things back out. This coincided with the western governments instituting lockdowns and stay-at-home orders, resulting in reduced demand, which further forced the shippers to park their containers, cancel sailings, or even consider scrapping some capacity. Ocean shipping will take years to recover.

Supply and demand shocks are essentially two sides of the same coin, each 180 degrees in front of the other. They're oscillations in the system. We'll be facing repeated demand and supply shocks in the future, hopefully each with smaller magnitude than the previous one, until the whole system stabilizes again.


It's both, but operating on different timescales. Friends of mine who operate in old-line businesses (packaging and hot tubs) tell me that supply chains are in shambles right now. They can sell things, but there's nothing to sell, because critical parts or ingredients just aren't available. This is all masked right now because few people are buying anything.

We saw the demand shock first, and because there's a demand shock, we don't see the supply shock. But if the reopening goes well, we're going to see massive shortages and inflation as all that pent-up demand chases goods that are sitting offshore on container ships or were never produced at all.


It's pretty hard to establish causality, but there is one way I know of establishing that A did not cause B. In particular, if A came after B, then A did not cause B. If sending signals back in time was possible, the hedge funds would have figured it out. If the demand shock happened first, then it wasn't caused by the supply shock.


> The pandemic is not primarily a supply shock. It is primarily a demand shock. People aren't going outside.

That it's not primarily a supply shock (all kinds of products and raw materials are just as available as they were before the pandemic) is yet another indicator of the impact increased automation and labor efficiency in manufacturing has had. It's remarkable - and a bit frightening - how little human labor is required to provide a sufficient supply of goods to a huge number of people.


> It's remarkable - and a bit frightening - how little human labor is required to provide a sufficient supply of goods to a huge number of people.

And, yet, we still insist that a "full time" employee has to work 40 hours a week, typically in the form of "butt in seat" time.


We aren't providing sufficient goods. We're consuming a backlog that was already produced and waiting, and that won't last.


If you're suggesting there were several months worth of backlog prior to COVID-19, that would imply pretty massive inefficiencies...


There weren't; if there were, we wouldn't be in half as much of a bad shape as we're now. "Inefficiencies" is just a greedy optimizer's word for "safety margins", "redundancies" and "buffers". But we're living in a word run by greedy optimization.

We're not consuming buffers (those were all cut long ago as "fat"), we're just lucky that the people responsible for logistics - who are right now forced to fly by the seat of their pants - are still managing to route goods around well enough for us to not notice, beyond an occasional shortage or price hike.


[Citation needed]


> It's remarkable - and a bit frightening - how little human labor is required

A huge number of jobs are engaged in verification/audit/fraud detection/related activities. "Paper pushers", the cliche goes.

An OK book that looks at this (and goes farther than I do, I think making some eventually unsupported claims) is _Bullshit Jobs_ by David Graeber. Worth a read - some of it is quite good.

Audit functions will always be needed. That said, I'm pretty sure there will be a lot of job functions that looked "normal" pre-C19, that people will decide no longer need doing, or will be done substantially differently after.


> A huge number of jobs are engaged in verification/audit/fraud detection/related activities. "Paper pushers", the cliche goes.

Agreed, and a lot of the rote work of auditing and fraud detection has been automated. But at some point, this devolves into the problem staying ahead of human adversaries, and that is hard to fully automate. Also, sometimes the slowness and inconsistency of human labor in these situations is itself a barrier to fraud.


Of course, humanity didn't get C19 and suddenly become honest.

I'm thinking of it more as a reset - there are a lot of job functions where things are done the way they are because that's how we do things. After pauses like this, people look at those with fresh eyes.


I think the public narrative not being kept congruent with actual supply levels is the main factor at play here. If everyone's talking about a shortage of toilet paper, you initially rush out to buy extra toilet paper, but can't find any. You continue to see dialogue of "toilet paper shortage", and figure it's not even worth the effort to walk down that aisle, but for all you know your local store is fully stocked.


The grocery store I go to has had limits on paper towels, toilet paper, flour, etc. to one per customer per day for I think months and that seems to have solved the problem.

The choices of flour were briefly very limited, but now there's pretty much anything I'd want.


If you want to see Ivan Werning's presentation: https://www.youtube.com/watch?v=-PkK9IlTHVM


Quick take on a first skim and thinking:

1. Looks interesting. Worthwhile type of thinking.

2. Control+F "savings" / "saving" ... no results. You'd think that the amount and distribution of savings in a population would greatly effect the outcomes.

That said, "elasticity" showed up 14 times in a control+f of the paper so it might be implicitly covered or named differently.

Edit: "reserve" / "reserves" had no relevant mentions (1x "rights reserved", 2x in citations of Federal Reserve), capital 0 mentions, "insurance" 14 mentions, "ability" 0 mentions, "willing" / "unwilling" / "willingness" 1 mention total, "substitute" / "substitution" 19 mentions, "bankrupt" / "bankruptcy" 0 mentions, "unemployment" 11 mentions, "funds" 0 mention, "cash" 0 mentions, "money" only 1 mention (!)... "monetary" shows up 20 times though.

Edit2: Lots of "utility function" going on. I personally don't really love the theoretical grounding of utility functions but it might actually produce interesting insights in this case.

Edit3: "Each agent is endowed with n > 0 units of labor which are supplied inelastically" - inelasticity of units of labor in this case seems like a pretty aggressive assumption. They carry that assumption through - "As before, agents inelastically supply labor n to their respective sector in the flexible price equilibrium; they may supply less than n in the equilibrium with wage rigidities. For now, we assume that workers are perfectly specialized in their sector."

Quite theoretical with a fair amount of simplifying assumptions to make it simpler to mathematically model. The conclusions are rather tentative and not strongly stated, in any event. Interesting line of thinking, I think you'd have to look closely at the assumptions and premises and how variables are defined to see if it corresponds well to real life.


Savings is implied through extensive discussion of marginal propensities to consume (MPC) - the fraction of additional income that a person will NOT save.


That's right, yeah.

But here's what they're investigating:

> ”In this paper, we introduce a concept that might be accurately portrayed as “supply creates its own excess demand”. Namely, a negative supply shock can trigger a demand shortage that leads to a contraction in output and employment larger than the supply shock itself."

They've got some time variables; I think an explicit savings variable (and maybe credit, too) would be important in determining whether that'd happen.

I might be mistaken, I tossed some quick thoughts up when this had zero comments, looking at methodology and assumptions moreso than conclusions. But you'd think regions with high-savings, low-savings but lots of credit, and low-savings combined with low credit would behave very differently towards "a contraction in output and employment larger than the supply shock itself."

Math gets way more complicated as you addd variables like that, though, and maybe it wouldn't effect the conclusions too much though, so... ¯\_(ツ)_/¯


Also see section 5.4. They use the word liquidity rather than saving - it's not being obtuse, just highly technical vocabulary.

" What happens if firms are liquidity constrained? If firms have some finite amount of liquidity at their disposal, say, because they cannot borrow nor issue equity and have limited past accumulated profits at their disposal, then they no longer maximize the present value of profits in an unconstrained fashion. This distorts firm decisions towards laying workers off, since the current period loss cannot be financed."


Determining information about Marginal Propensity to Consume directly implies information about Marginal Propensity to Save, as MPC + MPS = 1.




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: