To what extent are business cycles just random?

Let’s ask Guido Menzio, who perhaps knows something about random.  Here is one of his papers, with Mikhail Golosov (pdf):

We propose a new business cycle theory. Firms need to randomize over firing or keeping workers who have performed poorly in the past, in order to give them an ex-ante incentive to exert effort. Firms have an incentive to coordinate the outcome of their randomizations, as coordination allows them to load the firing probability on states of the world in which it is costlier for workers to become unemployed and, hence, allows them to reduce overall agency costs. In the unique equilibrium, firms use a sunspot to coordinate the randomization outcomes and the economy experiences endogenous and stochastic aggregate fluctuations.

In other words, by coordinating with each other, if only implicitly, employers make the firing threat more fearful.  You don’t have to interpret this paper literally as an entire explanation for cyclical unemployment, only that it may have something to do with the story.

And here is his about to appear JPE piece with Greg Kaplan (pdf):

We propose a novel theory of self-fulfilling unemployment fluctuations. When a firm increases its workforce, it increases the demand facing other firms—as employed workers spend more than unemployed workers—and decreases the extent of competition facing other firms—as employed workers have less time to search for low prices than unemployed workers. In turn, the increase in demand and the decline in competition induces other firms to hire more labor in order to scale-up their presence in the product market. The feedback between employment and product market conditions generates multiple equilibria—and the possibility of self-fulfilling fluctuations—if the differences in the shopping behavior of employed and unemployed workers are large enough. Empirical evidence on spending, shopping and prices paid suggests that this is the case.

In general, not enough popular macro discourse asks the question of how much of the cycle results from self-fulfilling prophecies.  Furthermore what does that imply for policy? yes, “confidence” can be important, but confidence in what exactly?


'who perhaps knows something about random'

There was nothing random about a freedom loving American ensuring the safety of her fellow citizens from someone who clearly did not belong in that airplane, what with his inability to engage in small talk while writing symbols.

And let us be honest - she was right, he was not an American, and thus worthy of all the scrutiny required to prove him to be safe to fly on an airplane. After all, the U.S. has invested a massive amount of time and effort in security in this area, for example fingerprinting non-Americans entering the U.S., just to give one concrete example.

To stir your pot:

In addition to producing biometric passports for German citizens, the German government has put in place new requirements for visitors to apply for visas within the country. "Only applicants for long-term visas, which allow more than three months' residence, will be affected by the planned biometric registration program. The new work visas will also include fingerprinting, iris scanning, and digital photos".[

So, Europeans want fingerprints and eyescans.

Governments do lots of things that citizens don't want.

I'd give pretty close to 0% odds that it would get a majority approval in a referendum.

Some mad scientists are even advocating for complete brain scans as means of identification. I can't imagine how things could possibly go wrong if the state were in possession of a complete copy of our neural functioning. I encountered a very "pro" argument to this end on the site of an organization which purports to be activist against "remote neural monitoring".

Oh please. As if a week doesn't go by with people down your end of the political spectrum freaking out over much more trivial issues. People have been fired for using the word "niggardly" - and it wasn't Trump's supporters who fired him. Every second college in the US seems to melt down the minute a White guy orders tacos.

Seriously, on a scale of one to ten, how much more dumb is it to assault a White student because he has dreadlocks than object to someone who might be a terrorist on a plane?

In the meantime, would it be too much to ask you to stick to the topic? What relevance does this post have? Of course we know all about self-fulfilling confidence boosters. We have seen Its a Wonderful Life

Wow. I never heard of any white guy getting assaulted BECAUSE he has dread locks. Surely some guy in dreadlocks has been assaulted, as have many bald guys and guys with crew cuts, but are there actually cases where the dreadlocks were the cause? (anti-hippy incidents excepted...)

Or in the words of George Bailey: "Can't you understand what's happening here? Don't you see what's happening? Potter isn't selling. Potter's buying! And why? Because we're panicking and he's not. That's why. He's picking up some bargains. Now, we can get through this thing all right. We've got to stick together, though. We've got to have faith in each other."

"The feedback between employment and product market conditions generates multiple equilibria—and the possibility of self-fulfilling fluctuations"

So in the words of a president trying to exhort consumers and businesses out of the Great Depression, "we have nothing to fear but fear itself".

Which probably also applies to some airline passengers and crew, such as the ones on Menzio's flight. And there might be a similar spreading of attitudes throughout society. I.e. paranoia might be contagious, under these Menzio-type models.

"To what extent are business cycles just random?" - not at all in a fiat currency regime - the CB causes them by over-tightening.

Of course as Keynes's pointed out way back in the 1930's there are positive and negative feedbacks in any economy that can reinforce trends. But the job of the Central Bank is to steer monetary to ensure that the positive feedbacks are dominant. Too often though CB's like the Fed act pro-cyclically not anti-cyclically.



Maybe we need to have more models on self-fulfilling sunspot equilibria, but in fact there has been quite a bit since Azariadis, Cass and Shell, Farmer, and of course Keynes. The problem may be more that too few economists have taken it seriouslyi.

As for the remark by prior_test2, I am not sure if you are being wittily sarcastic or serious. If the latter, you are truly pathetic.,


Based on comments by you on other threads here, perhaps your remark is the first, sarcastic. But you should realize given some of the people who comment here it is hard to know without some further indication. There are a lot of people here who make Donald Trump look like the Dalai Lama.

How is the second model different from Keynes? It looks identical to the model we teach in introductory macro. Good equilibrium: people buy, so people are employed, so people buy etc. Bad equilibrium: people don't buy, so people don't work, so people don't buy etc.

What implications does this have for an economy like Australia that hasn't had a recession in decades?

And, relatedly, does this model work in the context of stable NGDP growth?

In that case : the longer period without recession, the probability of having a recession in the present year converges to 1.

Look this ship in unsinkable......famous last words.

But predicting a recession every year is also not helpful.

Self-fulfilling? People, being people, expect tomorrow to be just like today, only more so. Contrarians, being oddballs, expect tomorrow to be different from today. Economists, being not people, are expected to know if tomorrow will be just like today or different. Faced with only two choices, same or different, economists have roughly a 50% chance of being right. Somehow economists are almost always wrong about tomorrow. But being the not people that they are, economists come up with creative explanations for why they are almost always wrong about tomorrow, blaming people (self-fulfilling, herd, irrational, phools, etc.) for the economists' wrongness being the accepted custom.

"Random" usually just means "we don't know why". How about "all the different factors are variant in length, magnitude, etc., leading to fairly unpredictable business cycles"?

I do believe that "animal spirits" can often be relevant, although perhaps they do more to bring on the change in the business cycle sooner, and exaggerate the fluctuations around potential. But I sort of doubt that they explain the fact of the business cycle.

Random can not only mean "we don't know why," but also "dependent on initial conditions to an extent that we couldn't possibly diagnose."

Don't preditor-prey ratios exhibit boom and bust? If things like that happen in simple natural systems, why not in the more complex?

The answer that we are too smart for that is easy to discard

The market is not a "predator-prey" system.

Think about it a little more. A preditor-prey system is both real and a standard teaching story for complex systems. Preditor-prey models easily produce unpredictable but self-similar behavior. They show that agents with very simple behavior and self-interest produce periodic catastrophic results.

And why would there not be a parallel to market agents competing with simple behavior and self-interest?

I just googled. There are a load of academic papers applying predator-prey models to markets and economics.

Economists' understanding of mathematical models associated with natural systems rarely exceeds that learned at the level of 1st or 2nd year biology/ecology, in the rare cases where it is even that high. But that does not stop them from making use of these concepts and applying them in the way that they seem to see fit for their specific theoretical explorations in economics.

However, I was unaware of all that research you refer to from the Google search. With infinite time I'd like to know lots about what they're doing with that ...

Isn't Shiller's Irrational Exuberance popular economics right up this alley?

But yes, I agree. There is often a "Real" component to Business Cycles, but there is never not randomness and emotion.

I think if economics PhDs spent a year running a small company, they would opine more intelligently on such matters.

Business ebbs, and it flows. This is the nature of the beast. Things get busy and you scramble, things slow down and you consolidate. Over and over. It's a constant fine tuning process. Are we overstaffed? Understaffed? What's the pipeline look like? Do we need to focus on sales, execution, internal management, cutting costs, expanding operations?

And it's all interconnected in the larger economy. Sometimes the "ebb" rhythm of several big companies or economic sectors coincide.

People in government and academia have no conception of this constant churn.

Don't they? How do you know all they know?

I don't think following another person (herd behavior) is truly "random", because it depends on the time constant (natural resonance) of the system. Simple explanation of the above engineer-speak: if A follows B follows C, etc, then how long before the pendulum switches and they stop following each other? That's the 'time constant' in a system.

Personally, I think "Kondratiev waves" of 40-60 year cycles explain more than this study, as do "Kuznets swing" waves. Here's a simple explanation of how to reconcile this study with those cycles: a pioneer invention is appropriated by a businessman (it's always that way) and he becomes a "first mover". But though there's always a "first mover advantage" (MySpace), there's always profit in being a "me-too" follower (Facebook), and others who come up with small improvements. More and more people follow this fashion, until it burns out. How long before it burns out? It depends on the time constant as I referenced above, but historically, the 'big cycles' are 40-60 years and the 'smaller cycles' are 15-25 years. Various explanations can account for the big and smaller cycles, such as obsolesce of machinery (60 years max) or land use value changing, which depends on the life cycle of buildings and demographic trends (one generation = 25 years, grandma moving from NYC to Miami, etc). You see the same cycle in music (Swing, Jazz, Bebop, Rock, Pop, Disco, Grunge, Dance, Rap (should be dying out about now according to my calculations), etc)

When I grew up Kondratiev waves were a punch line. Something you heard on UHF channels from stock hucksters. Like head and shoulder formations.

I see that seemingly bright people now talk about Kuznets waves, and I grasp for how it can be less crazy.

The world is big. A lot of things are going on. Medical and energy and information and economic and political transformations are not aligned. Anything that reduces literally a world of change to a wave must be a reduction too far by a mind that needs order too much.

With the exception of efforts in physics to find a "unified theory", I think reductionist thinking is dangerous according to the principle "a little knowledge is a dangerous thing". Perhaps not at the level of people who are progressively learning about stuff, or don't have time to get advanced knowledge, but definitely so when in the hands of people making/influencing hugely important policies.

Regarding the headline: none. Business cycles happen for a reason.

If you want to argue recessions are a choice, then there should be some rationale about frequency, depth, and length.

"Random" can mean

- a chaotic system without any stable rules or central tendency, or
- a system informed by the momentary emotions of its participants, with some momentum built in, or
- a system whose dynamics we don't understand well

If a system lacks rules, then business cycle theory is not within the realm of economics at all, as knowledge and analysis cannot provide insight by definition. If this were the case, however, I would expect that the unemployment rate would tend towards 50% and would see no self-correcting tendencies in the market. Bubbles would not pop and recessions would never end. Employment and output would be a random walk, without meaningful trend lines. But that's not what we see.

The emotion-based model is an 'influenza' approach to economics. Sometimes the economy just gets sick for no apparent reason. Rising unemployment and falling output are not signs of market imbalances, but simply an unwell economy. This leads to a kind of Keynesian interpretation, in that, deficit spending should revive the patient. Give the patient Sudafed (or more likely, Prozac), and all will be well. This is not terribly convincing either in many cases, as booms often have visible bubbles and countries who borrowed heavily during the downturn don't appear to have done much better than those who didn't (indeed, they performed more poorly). Further, this model is hard to sustain over a multi-year period. It's OK to be depressed for a week or a month, but seven years seems like a long time to be in a bad mood.

The we-don't-understand-it model is simple intellectual capitulation. No need to publish on that.

I think we do need to contemplate that a recession--a downturn in output and employment--may be more than one thing. It may be a simple real business cycle, with a demand shock leading to liquidity increasing faster than fixed assets (due to a construction lag), and thereby creating asset bubbles which burst when supply finally catches up with demand.

Financial shocks, of the sort seen after 1929 and, I would argue, 2007, tend to have persistent effects and involve deleveraging, ie, they are balances sheet, rather than income statement, events.

And then there are 'mini-recessions' which last, say, three months, involve a brief deviation from trend, but then disappear. These might conceivably be linked to short-term changes in emotions or beliefs, which will be discarded as new incoming data refutes market sentiment.

Comments for this post are closed