Kotlikoff on PSST

Laurence J. Kotlikoff writes,

Economics has many theories of economies rapidly flipping from good to bad. They go under the headings multiple equilibrium, contagion, self-fulfilling prophecy, panics, coordination failures, strategic complementarities, sun spot equilibria, collective action, social learning, and herding.

Pointer from John Cochrane, who offers extensive comments. Read his whole post.

Kotlikoff argues that the mere perception that the economy was in trouble was enough to cause trouble.

Employers laid off their workers in droves to lower their payrolls before their customers stopped arriving. This was the worst of the many types of multiple equilibria associated with the GR.

…The slow recovery is hard to explain except as the result of everyone expecting a slow recovery.

I see this as a PSST story. Patterns of specialization and trade depend on business managers’ confidence that those patterns will continue.

I, too, have been thinking a lot about the contingent nature of economic outcomes. I am mulling an essay that will strongly criticize the view that the market acts like a set of equations providing a deterministic solution for given tastes, technology, and resource endowments. Instead, there are multiple equilibria that depend on people’s perceptions and beliefs.

Part of my argument is that hardly anyone in the economy has a measurable marginal product. Most of us are producing intangible output, even if we work in goods industries. For example, relatively few of the employees at a pharmaceutical company are actually bottling pills.

Businessmen operate in a world of high fixed costs, with mostly overhead labor. If they doubt that revenue is going to remain at current levels, one response is to cut costs by reducing overhead labor. If enough firms do this at once, their fears of recession becomes self-fulfilling. This sort of self-fulfilling recession is particularly easy to fall into when there is a financial crisis.

Kotlikoff’s main point, which Cochrane emphasizes and expresses support for, is that a financial crisis of 2008 was itself a sudden shift in perceptions that took place in the context of an inherently fragile financial system.

15 thoughts on “Kotlikoff on PSST

  1. That still leaves the question of where those overhead worker wages come from.

    Currently the labor markets are pretty tight and work is still expanding in many places. What is the point of view of a typical private organization with regards to hiring?

    Well, it starts to notice things picking up, and then stresses and bottlenecks and delays emerging from certain areas becoming overwhelmed. If there’s some slack in the other complementary components of the firm – which there often is as expansion burden doesn’t scale linearly for all firm specialties – then the target for more hiring is in that particular area.

    Then one tries to hire from the market at a certain level of skill and quality, and discovers what the going rate is for someone like that, such that one gets enough acceptable candidates responding to advertisements and showing up to interviews, etc. And also, one might have to start paying the existing workers more if they might be poached by other firms.

    Then the firm has to ask itself, “to handle x% more business, where the average profit is y, is another worker at going-rate wage w worth it?” If so, hire. If still so after one hire, keep hiring. If not, stop.

    (This is a static analysis, but it’s possible to make it dynamic and plug in expectations and so forth.)

    And what one ends up with after it all shakes out is a kind of rough and informal “partial differential” for marginal productivity.

    My impression is that the same thing happens in reverse during downturns, that is, firing is not uniform across specialties and subdivisions and facilities, with the average firm firing until removing suspected fat starts to cut into muscle and bone.

    As to the perceptions-cycle (or Minsky cycle, or Animal-Spirits cycle), it raises the question on what, if anything, can or should be done to buffer and smooth out the social-psychological cycle, beyond making the system less fragile, with more counter-cyclical shock-absorbers, and fewer “piss your pants” moments.

    My view is that this is one of the major advantages of an NGDPLT approach. It provides a sure pole star of a rooted expectation and around which everyone else in the economy can coordinate their actions and plans, further ensuring that unexpected burdens or windfalls of debt instruments from real shocks are softened, which dampens psychological manias or collapses.

    • I can’t tell if it’s embedded in your description, but I would add the following: “the firm” is, at least in large organizations, a collection of budget centers with some ability to add staff independent of their contribution to marginal product.

      So adding staff will not be an efficient process. The stated rationale will be to accomplish valuable work. Whether the work of that budget center is actually marginally valuable is much murkier. And unstated drivers like empire building will also play a role.

      On the opposite side, budget centers will tend to shed staff only when economic imperatives require it. And which budget centers shed staff will be partly economically efficient and partly political.

  2. “Employers laid off their workers in droves to lower their payrolls before their customers stopped arriving. This was the worst of the many types of multiple equilibria associated with the GR.”

    I don’t see this as an accurate representation of what happened. Layoffs were fairly normal into September of 2008, the rise in the UE rate was due to lower hiring rates not high layoff rates.

    Layoffs here https://fred.stlouisfed.org/series/JTSLDL

    Total vehicle sales here https://fred.stlouisfed.org/series/TOTALSA

    We can see that they had fallen off by about half of what would be their total drop before layoffs became a significant contributor.

    Total (seasonally adjusted) retail sales here https://fred.stlouisfed.org/series/RSXFS

    Sales had stopped growing early in the recession and had already started to decline before the large layoffs of October/November/December.

    Layoffs were among the last things to shift from the norm unless you decide that 2006 was the norm and 2000 through 2005 and 2007 into 2008 were abnormal.

  3. “Instead, there are multiple equilibria that depend on people’s perceptions and beliefs.”
    I don’t believe there are any equilibria — each day is different and includes changes from the day before. Instead, there are always multiple forces at play, some pushing for more hiring, some for less hiring.

    Some seen in the statistics, like legal workers, and some not seen. Most analyses will fail to include the millions of illegals who were working so productively building houses and McMansions in the ’96-2006 house construction boom, where everybody who bought saw almost immediate price increases; until house buying credit was fully maxed out.
    The 5 million (?) illegals who were working in construction were not counted in many worker statistics (reported wages), but were partially seen in other statistics (retail purchases, rents).

    The real economy started having problems in 2006 — these problems were mostly not seen until 2008 mostly because a huge amount of the pain was absorbed by the illegals, until the rocket-scientist Mortgage Securities folks ran out of suckers to give them more cash to throw at housing, which started dropping and the leveraged “investors” (=speculators) saw their options profits turn into huge losses.

    Here’s a question I haven’t seen asked or answered.
    Assuming huge numbers of illegals in the US left between 2006-2009. How many would have to leave in order to get a Great Recession sized effect from emigration?
    Were I to still be studying economics, rather than reading blog posts, this is the question I’d be asking and trying to answer, because it’s such a huge influence which is mostly unstudied.

  4. “Part of my argument is that hardly anyone in the economy has a measurable marginal product.”

    This is undoubtedly true at the macroeconomic level when attempting to convert say administrative work into bushels of wheat units.

    I would be leery of taking it too far though. Within firms, management information systems are nearly ubiquitous. How many people knock off for the weekend without having had to file a weekly time report? For decades I had to account for my time by entering data into various automated systems charging time in various increments from 5 minutes to an hour to be charged to a client or to some project or other designated categories. Sure, a lot of it is garbage in and garbage out, but I’ve also had to justify budgets and personnel requests using these same management information systems.

    I am highly confident that production managers, business planners, executives, and others rely upon information about their various production factors and make decisions based upon projections how each additional employee contributes to the bottom line.

    As you noted yesterday, much of employment is a fixed cost and unrelated to changes in production levels. Most plants have to have an safety officer or an environmental engineer to file EPA reports regardless of whether the plant is producing 1o units or 100 units. But the firm will still have a confident notion of what each additional plant will be able to produce if they decide to build it.

    A minor point but still, it seems necessary to distinguish between what happens within a firm, and what happens when you aggregate a bunch of firms or industry sectors.

  5. Strongly criticize the view that the market acts like a set of equations providing a deterministic solution for given tastes, technology, and resource endowments. Instead, there are multiple equilibria that depend on people’s perceptions and beliefs.

    I am not sure what huge differences are here. How modifying the market providing a multiple equilibria for given tastes, technology, resource endowments and people;s perceptions and beliefs. I would agree that there is no defined deterministic solution but people make decisions based on perceptions, realities, tastes, technology, and resource endowments. I tend to side with soft Keynesian versus defined AD-AS curves.

    And the market is moving and economic actors see a different equilibria at different times. Look at the 2008 Financial Crisis in which Dean Baker & Peter Schiff both saw a Housing bust/crisis in 2006 compared Paul Krugman seeing it in Spring 2007. (Or that some homeowners saw this storm in 2006.) Or most Financial institutions internally knew January 2008 things are going South already. The thing about The Big Short is people started shorting the market in 2005 and it took 3 years to hit Crisis point.

  6. Something telling from the Kotlikoff piece, he quotes Andrew Lo

    “… (there is) significant disagreement as to what the underlying causes of the crisis were and
    even less agreement as to what to do about it. But what may be more disconcerting for
    most economists is the fact that we can’t even agree on all the facts.”

    And then he says

    “To be sure, my take is an outlier relative to the standard diagnosis of the problem and its cure.
    The standard view is that the housing market experienced a bubble, that regulators were asleep
    at the wheel, letting households and banks overleverage, that Wall Street manufactured and
    overvalued dangerous, complex derivatives, that conventional and shadow banks issued and
    then sold fraudulent subprime mortgages, that rating companies were routinely bribed, that
    financial traders traded too much, that shadow banks operated outside of regulatory scrutiny,
    that Congress forced Fannie Mae and Freddie Mac to encourage subprimes, and that there was
    too much risk taking. ”

    How can there be a standard case when there is little agreement among economists? This is the approach that a lot of people take, they put up the “standard” argument which is really bits and pieces of a great many arguments and present them as a whole.

  7. I see this as a PSST story. Patterns of specialization and trade depend on business managers’ confidence that those patterns will continue.

    If that was true, then the late 19th century economy should have been relatively immune to booms and busts because the patterns of specialization and trade were much less complex. In actuality, that period was characterized by large swings.

    If anything, patterns of specialization and trade probably add some stability – if management knows that it’s employees and equipment are specialized and hard to replace, it may make sense to try to ride out a downturn (accepting a loss today in order to maintain PSST that are expected to be profitable in the future).

    On the other hand, wheat is a very simple product that humans have produced for thousands of years with minimal specialization. If farmers anticipated at planting time that wheat prices would be low at harvest time, would that affect their planting decisions (including the demand for farmhands at planting time)? Probably. So this might be a systems-control, feedback-loop sort of story.

    • Under PSST the economy in the 19th century would have been more boom and bust prone, PSST is not about riding out a downturn and things returning to normal, it says that the downturn is when normal ends and the upswing is when a different normal emerges.

      • Developing PSST takes investment, and people don’t easily walk away from investments. That tends to be a stabilizing influence.

        Personal example – I have an Ivy League Ph.D. in a physical science field with applications in electro-optics. It took about 9 years to get, including undergrad. I graduated just about the time most of the jobs in the field went to China. This isn’t uncommon; only about half of hard-science Ph.D.s work in their fields.

        I tried to make a living in the field for about 12 years with intermittent success and extended periods of unemployment (fun fact: McDonald’s doesn’t hire Ph.D.s! I applied!). Eventually I gave up and became an accountant.

        Hypothetically, if I had been less specialized (i.e. had invested less in my initial career choice), do you think I would have given up sooner or later? I think later, which is why I think PSST is probably a stabilizing factor. Specialization is investment, and it tends to lock people into pre-planned behavior.

        • People walk away from unprofitable circumstances, it might take time but they always do. This isn’t a stabilizing factor, it is actually more stabilizing to walk away sooner when the losses are the least and the investment can be swapped rather than later.

    • if management knows that it’s employees and equipment are specialized and hard to replace, it may make sense to try to ride out a downturn (accepting a loss today in order to maintain PSST that are expected to be profitable in the future).

      This would seem to be a case of making no small changes when things are a bit bad, and only make the big changes when the downturn for your firm becomes more clearly a downturn for the whole industry.

      Recently, in my own life, we’ve seen the “mini-computers” get replaced by smaller PCs; and all non-IBM mainframe computer makers (remember the BUNCH?) stop making mainframes. I’d guess IBM’s push into cloud & security will lead them to replace their own mainframes with dedicated server farms, too, but more slowly (as there are some transaction processing jobs really done better by a single mainframe. Maybe.)

      Firms, niches, segments, industries. When is a “temporary” downturn actually permanent? I recall my own study / career choices of Systems Engineering and Nuclear Power — but after Three Mile Island, I understood that there was a lot more future in computers. Not so much in FORTRAN or BASIC, tho.

      Jay, I’m surprised you didn’t drift into programming, as so many hard-science guys did when they didn’t want to be professors after getting a PhD. Yet accounting / finance is becoming more tech savvy every year, too.

      I’m also getting some “specialization fatigue”, as I work at one company getting deep knowledge of that company and its particular set of programs, definitions, and ways of working. The general SQL/excel (incl. VBA) skills can easily transfer to another company, like any junior, but the deeper firm-based knowledge is much less useful.

      Some manager might be very “productive” at one company, making excellent strategic decisions and meeting comments, yet if at another company, very similar decisions might not be so effective. Especially at the lower levels of where the work get done, and managers need to understand what the worker/analysts are actually doing.

      At higher VP levels, where budgets get allocated to various departments, the “manager/people skills” often seem quite transferable. And a successful VP at one firm can easily become a successful VP at another firm.

      Finally, on the issue of PSST being a “stabilizer”. It is, insofar as it helps maintain higher max capacity ready to instantly increase if customers start buying more, again. Like car making going up or down.

      But this means if the downturn is so bad that the factory is closed, it’s because both that industry is not coming back (soon) AND because the general economy is bad. But then the closure of the factory makes the general economy worse and contributes to an economic downward spiral. Making more questionable factories close, making the spiral go further down. The classic Ag Demand getting weaker and weaker. Many plants, due to their specialized investment, were delaying the difficult adjustment decisions, so go bankrupt waiting for the upturn that doesn’t come. Prior Specialization investment leading to wrong decision of staying with the old too long.

      The successful entrepreneurs & investors make big bets on the future new Patterns, and are right.

      • Sure. When I say that PSST is a stabilizer, I mean that it tends to work against the “rapid flips” that the main post was talking about. I eventually abandoned years worth of specialization, but I tried to maintain its value as long as I could. That sort of behavior seems typical to me.

        Speaking to one of your other points, a startup company that I worked for once hired a manager who’d been very successful in the Federal Department of Energy. It probably won’t surprise you to learn that his skills at managing within a huge bureaucracy did not translate well to the ten-guys-in-a-garage milieu.

  8. ” If they doubt that revenue is going to remain at current levels, one response is to cut costs by reducing overhead labor.”

    Does Sumner’s NGDP target attempt to keep revenue at an expected level across the economy and thereby forestall this doubt?

Comments are closed.