Matt Bai's Incompetence is a Reason to Shut the New York Times Down Now and Let the Good Journalists Get Picked Up by Other Outlets
Liveblogging World War II: August 12, 1940

First Draft of September 1 Principles of Economics Lecture: Depression Economics

2.0. What We Hope You Will Learn

2.1. The Need for Depression Economics

2.1.1. The Coming of the Great Recession

In December 2006 63.4% of American adults of working age had jobs. By December 2009 only 58.2% had jobs. Over those same three years the unemployment rate--which looks at a narrower group, not all American adults but only those who say that they are actively looking for work and would take a job if offered one--jumped from 4.4% to 10.0%.

Total production in the economy had stood at a level of $13.06 trillion per each year at the end of 2006 (measured in the prices as they stood in 2005). It had then been growing at an average rate of a hair above 3% per year. People expected it to stand at $14.3 trillion per year as of the end of 2009. But it did not. The flow of production at the end of 2009 was a mere $13.1 trillion per year--fully 8.5% lower than what three years before we had all expected it to reach. More than 8% of the flow of production of useful goods and services that we ought to have been producing and could have been producing at the end of 2009 was not there. It had vanished completely--into thin air.

Why was there this sudden shift, this "Great Recession" in the pace of the flow of production and demand and and the level of employment?

The answer is that the "Great Recession" of the late 2000s was a virulent attack of a disease that has periodically but irregularly struck industrial market economies since at least 1825. We call this disease the demand-driven industrial business cycle. Large numbers of people are unemployed in 2009-2010 because aggregate demand is low. People are not working because they and their potential employers don’t think they could sell the things they make and do if they were. The expectation of sellers that they can, generally and on average, find willing buyers at more-or-less the prices that they had expected, has gone wrong. And there is, in general and economy-wide, there is excess supply: too many goods (and services) chasing too little spending.

In such a recession--we generally do not use the word "depression" for anything after World War II, largely because the word sounds too scary--sellers all across the economy find that buyers do not show up in the numbers they had been expecting. Inventories of unsold goods pile up on the shelves. This wave of extra unexpected inventories works its way back through the production chain. Producers respond as you would expect when they find they are making goods they cannot sell. Firms lay off workers, cut back production, and cut prices.

Normally, whenever there is deficient demand for some commodity--and hence a glut of it on the market--there is excess demand for and hence a shortage of another one. Thus one firm or industry will be hiring workers, increasing production, and raising prices when another is firing, cutting, and lowering. Workers will move more-or-less smoothly from one set of declining occupations and industries to another set of rising occupations and industries.

In a recession that is not the case. A recession is a general glut: a shortage of aggregate demand, not a shortage of demand for only one or only a few commodities and a surplus for others. And in a recession the things that producers do to handle a single-commodity glut--firing, cutting back, and lowering--do not help repair the situation but instead work to make matters worse.

One (partial) reason there is low aggregate demand is that so many people are unemployed--and so have reduced incomes, and so can spend less. The feedback loop from lessened aggregate demand to reduced employment to reduced incomes to even further reduced aggregate demand is a vicious circle. It makes recessions and depressions worse than they would otherwise be.

But a feedback loop that magnifies the effects of a decilne in spending is not by itself a complete explanation. Where does the recession start, and why?

2.1.2. Economists and the Possibility of a “General Glut”

You cannot have a downward vicious spiral without an initial push. Where does the initial push that got us into this situation come fromo? The answer is that the initial push can come from a number of places, and take a number of forms, but that once the recession begins the process by which deficient aggregate demand is generated and propagates itself is very similar. But in order to understand how it is that situations of excess aggregate demand can emerge, we need first to go back to understand why they are not omnipresent—why excess aggregate demand is a sometime and not an all the time thing.

French economist Jean-Baptiste Say (1767-1832) was the first person to write down his thoughts on this issue. Say had been special assistant to Tom Paine’s friend and France’s Girondist Secretary of the Treasury Etienne Claviere--who was then purged, arrested, imprisoned, and executed by Maximilien Robespierre’s Mountain Party. Somehow Say escaped the purge of the Girondists with his liberty and, more important, his life. Say decided, perhaps wisely, to retire from politics and government. He become a theoretical economist. Ten years later he published his first economics book, his Treatise on Political Economy. And thereafter he churned out more and more volumes

In his 1803 Treatise Say dealt with the possibility of a "general glut," of deficient aggregate demand. He concluded that there could be no such thing. Aggregate demand had to match supply, he wrote, because the only thing that could generate demand was supply:

it is production which opens a demand for products.... Yonder farmer... can buy none at all [of your woollens] if his crops fail] altogether. Neither can you buy his wool nor his [wheat] yourself, unless you contrive to [first sell] woollens or some other article.... The silver coin you will have received for the sale of your own products and then use to buy those of other people will in the next moment do the same thing for other contracting parties, and so from one to another to infinity.... You will have bought, and everybody will have bought, what you want or desire, each doing so with the value of his own respective products sold and transformed into money for that instant only. Otherwise, how could it be possible that there should now be bought and sold in France five or six times as many commodities as in the miserable reign of King Charles VI? Is it not obvious that five or six times as many commodities must be produced now [as then]. And that they must have served to purchase each other?

You should recognize this as the circular flow principle of the previous lecture. Households earn money--and they then spend it: it doesn't do them any good if they don't spend it on anything, and "spending" includes buying a bond or putting it into a bank. Businesses receive what households spend, and they then use that money to (a) hire workers, (b) buy things, or (c) distribute to their shareholders as profits: it doesn't do them any good if they in turn do not spend or distribute it. But the spending of businesses hiring workers and the distribution of profits are the incomes of households.

Thus Say argued in 1803 that we didn’t have to worry about a lack of aggregate demand. Consider a simple toy model of a three-sector economy--agriculture, industry, and service sectors, and since this is Berkeley let's talk about baristas, potters, and yoga instructors. We thus have baristas who make lattes, potters who make ceramics, and yoga instructors who teach lessons. Can there be a situation in which baristas have brewed more cups of coffee than potters wish to buy who have made more ceramics than yoga instructors want to buy who are offering more yoga lessons than baristas want to take? Say in 1803 said no. And others have picked up the argument ever since.

2.1.3. Does Excess Supply Here Mean Excess Demand There?

As we saw last time, Say’s argument has at its core the truth that is the circular flow of economic activity. Everybody’s expenditure is somebody else’s income, and everybody’s income is somebody else’s expenditure. You cannot earn the money that you will yourself then spend unless you can sell what you are making. And they cannot buy what you have to sell unless you have bought what they are selling. That circular flow seems at first glance to rule out any possibility of a “general glut”--of a general economy-wide excess of supply. Say in 1803 certainly thought that it did so.

But by the end of his career, in his last book, his 1829*Cours Complet d’Economie Politique Pratique,* Say was singing a very different tune. Describing the British economy’s crash and depression of 1825-6 he admitted not only the possibility but the reality of such a “general glut”:

As [the price of] every type of merchandise had sunk below its costs of production, a multitude of workers were without work. Many bankruptcies were declared among merchants and among bankers, who having placed more bills in circulation than their personal wealth could cover, could no longer find guarantees to cover their issues beyond the undertakings of individuals, many of whom had themselves become bankrupt...

And Say’s 1829 analysis of how the British economy had then gotten itself so wedged sounds remarkably modern:

The Bank [of England]... obliged to buy gold back... [t]o limit its losses... forced the return of its banknotes... ceased to put new notes into circulation... was then obliged to cease to discount commercial bills. Provincial banks were in consequence obliged to follow the same course, and commerce found itself deprived at a stroke of the advances... to create new businesses, or to give a lease of life to the old.... [B]usinessmen... obliged to meet [the bills they had issued]... each was forced to use up all the resources at his disposal. They sold goods for half what they had cost. Business assets could not be sold at any price...

But how can it be that the price of everything “had sunk below its cost of production” if everyone’s expenditure is somebody else’s income and thus everybody’s cost is somebody else’s purchasing power? The circular flow principle seems to rule it out.

We understand how the price of coffee can fall below the cost of making lattes with two sugars and an extra shot if people decide that they have enough lattes and wish instead to purchase inner peace via yoga. But then people have rechanneled their spending to yoga lessons. And then the price of yoga lessons is greater than the cost of giving them. Excess supply in one industry means excess demand in another one. If baristas are making losses (and coffee shops are firing workers and closing locations), then yoga instructors are making huge profits (and hiring assistants and training new colleagues).

2.1.4. Disrupting the Circular Flow

It was an economist a generation younger than Jean-Baptiste Say who put his finger on the reason: moral philosopher, libertarian, colonial bureaucrat, feminist, public intellectual, and economist John Stuart Mill put his finger on the answer in a piece he wrote in 1829:

[T]hose who have... affirmed that there was an excess of all commodities, never pretended... money was one of these.... [P]ersons... at that particular time... [fearing] being called upon to meet sudden demands [for payment], liked better to possess money than any other commodity. Money, consequently, was in request, and all other commodities were in comparative disrepute... the result is, that all [other] commodities fall in price, or become unsaleable...

We don’t just buy those goods and services that are then currently being produced. We don’t at just sell the current flow of services from the labor, the machines and buildings, and the natural resources we own. We add to the current flow of our incomes by selling our assets. We spend our purchasing power not just on the goods and services currently being produced but on financial assets as well.

Thus it is perfectly possible for there to be an excess supply of goods and services--for the current flow of aggregate demand for goods and services to be less than the cost of the goods and services currently being produced--if there is an excess demand for assets. Depressions come, and we need depression economics to analyze, when there is an excess demand for one or more of three particular kinds of financial assets:

  1. "Liquid" assets, assets that can be readily and easily used to pay for things, which assets we usually call “money”.

  2. High-quality assets, assets that are generally regarded as safe ways to store up purchasing power so that it will still be there intact to be used later on--like U.S. Treasury bonds.

  3. Long-duration assets, assets that allow us to take some of the money we are earning now and move it back in time away from us into the future.

Whenever there is full employment and yet the population as a whole wants to hold more of any of these types of assets than exist, people try to switch their spending away from spending on currently-produced goods and services and towards accumulating these assets. And that puts downward pressure on employment and production.

That is the important insight. let us see if we can make it--the mechanism of how the economy falls into a depression--clearer:

Consider, first, a normal shift in demand: Berkeleyites decide that they want to spend somewhat less on lattes that make them jumpy, irritable, and stressed and somewhat more on yoga lessons in order to seek inner peace. Baristas find that they have brewed more lattes than they can sell. Some cut their prices and see their incomes fall, some cut back on hours, some find themselves unable to buy the shade-grown beans for their next round of production and are unemployed. Yoga instructors find demand booming. They schedule extra classes. They work late into the night chanting “om mani padme hum” to satisfy demand. They raise their prices. They take on extra apprentices to help them carry the load. Prices fall in the coffee industry. Prices rise in the fitness industry. Excess supply of coffee and baristas comes with excess demand for yoga lessons and yoga instructors. In a short time the economy adjusts. Labor exits the coffee industry and enters the yoga industry. And in a short while the economy has rebalanced with fewer baristas and more yoga instructors, the structure of production has shifted to accommodate the shift in demand, and there is no more excess unemployment.

But now consider what Jean-Baptiste Say and John Stuart Mill were talking about. Consumers decide that they want to spend somewhat less on lattes purchased from baristas and to hold more cash in their wallets instead. Instead of spending normally, everybody decides to keep at least one $20 in reserve at all times. Those with less than $20 simply stop spending on clothes—until somebody buys some of what they have made and they have more than $20 in their pockets. What happens?

Well, what happens in the coffee industry is the same thing that happened when there was a shift in demand from caffeine to inner peace. Baristas find that they have brewed more lattes than they can sell. Some cut their prices and see their incomes fall, some cut back on hours, some find themselves unable to buy the beans for their next round of production and are unemployed. Inventories of unsold beans and cold coffee pile up. Entrepreneurs looking at their growing piles of unsold inventory cut back on hours and production even more.

But there is no countervailing increase in spending, employment, and hours for yoga instructors.

Things then snowball. The unemployed baristas now have no incomes. They cannot afford to buy as many pots or as many yoga lessons or, indeed, as much of the coffee made by other baristas. Inventories of unsold goods keep rising, and so employers cut back production and employment even more. Thus there is a second-round fall in demand which renders even more people unemployed—and not just weavers this time. And then there is a third round. And so on.

Moreover, everybody sees rising unemployment and falling incomes around them. Can you imagine a better signal to make you decide to try to hold onto more cash? Instead of cutting back on spending on coffee when you have less than $20 in your pocket, people start cutting back on all spending when they have less than $40 in their pocket. And the more the prices at which you can sell your goods falls and the higher unemployment climbs, the more desperate people are to pile up more cash in their wallets.

In a normal market adjustment--a fall in the demand for lattes and a rise in the demand for inner peace--the workers fired from the coffee industry would rapidly be hired into the yoga instructor industry. But this is not a normal market adjustment: this is depression economics.

How far down does production and employment decline when the economy gets itself into a depression economics state? How high does unemployment rise? Well, employers keep cutting back employment—and thus keep cutting back their workers’ incomes—until they are no longer producing more than they can sell and inventories are stable rather than rising. And households keep trying to build up their cash balances until their incomes have fallen so low that they do not think that they dare economize any further to try to boost their cash. How far is that? To determine how far that is, we need to build another, different economic model—a macroeconomic model.

2.2. The Income-Expenditure Framework

2.2.1. Downward-Sticky Wages

To understand depressions, we need to build an economic model in which the market system does not work well. If the market economy was working well, we would not have a depression and mass unemployment. And so a model that premises that the market system works well cannot help us.

So let us start, instead, with the assumption that prices and wages are, at the level of the economy as a whole, “sticky” downwards. When total spending falls—as it did from 2007-2009—average wages and prices will not. Businesses respond to falls in demand first by firing workers and shutting down their production lines, and not by cutting wages. And if businesses do not cut wages on a large scale, they cannot afford to cut prices. Losing money on each item sold and trying to make it up on volume is not a profitable business strategy.

Why are wages sticky? Here are four possible reasons:

  1. Managers and workers find that renegotiating wage levels downward is a costly and disruptive exercise as people make all kinds of threats about how they will behave if the other party doesn’t knuckle under that they do not mean but then feel forced to carry out. Hence cutting wage levels best delayed as long as it possibly can be, and then it is best delayed a little longer than that.
  2. Managers and workers lack information and so confuse changes in total economy-wide spending with changes in demand for their specific products: if it is demand for your particular product that has fallen, you won’t be able to cut wages and still keep your same-quality workforce—better to get ahead of the game by shrinking your operations.
  3. The level of wages is as much a sociological as well as an economic variable—determined as much by what values people think is "fair" as by the balance of supply and demand. Workers take a cut in their wages as an indication that their employer does not value them—hence managers avoid wage cuts because they fear the consequences for worker morale and worker effort.
  4. Managers and workers suffer from simple "money illusion"; they overlook the effect of price-level changes when assessing the impact of changes in wages or prices on their real incomes or sales, and so don’t notice that other prices and wages are falling all around them when they consider whether to cut wages.

All of these reasons are operating.

People do wish to stabilize commercial relationships by long-term contracts. Customers do find frequent price changes annoying. When other firms are not changing their prices and wages, you attract attention you may not want when you change yours. Hence managers and workers do prefer to keep their prices and wages stable as long as the shocks that affect the economy are relatively small—or as long as they think that they will quickly pass. People do lack full information, and so they are unsure whether a change in the flow of spending on their products reflects a change in overall demand or a change in demand for their product in particular. Managers who are uncertain which the change is will split the difference. Workers and managers are really not the flinty-eyed rational maximizers of economc theories. Work effort depends mightily on whether workers believe they are being treated fairly, and cutting your wages is almost universally perceived as unfair.

Which of these is the most important factor?

The best thing to say is that economists do not really know. But we do know that total spending in the American economy in mid-2010 was 10% below what the pre-2008 trend had led everybody to expect it to be, and that this fall in spending was unaccompanied by any noticeable decline relative to trend in either wages or prices. All of the decline in spending was, instead, a decline in production and employment.

2.2.2. Consequences of Downward-Sticky Wages

Thus any economist who wants to describe the real world will note that price and wage levels—not individual prices and wages, but economy-wide average levels—are sticky downward. Prices and wages remain fixed at predetermined levels as businesses expand or contract production and employment in response to changes in demand and costs.

If wages and prices are sticky downward, then the consequences of a sudden rise in household or business desire to hold cash are clear: as businesses see spending on their products begin to fall and inventories, they will cut production and employment. They want to avoid accumulating unsold and unsellable inventory, so they will cut production and employment until their level of production is no greater than total economy-wide spending, and so inventories are no longer growing. And by the circular flow principle, as they cut production total economy-wide incomes will fall as well, for the flow of production is nothing other than the flow of incomes. Thus to determine how much they will cut production, we need to figure out what total economy-wide expenditure will be.

2.2.3. Consumption Spending

Above we saw that total spending was divided into four components:

  1. Consumption spending (C),
  2. Investment spending (I),
  3. Government purchases (G), and
  4. Net exports, the balancing item—which in the United States, for my lifetime, have almost always been negative.

Add up these four components and call their sum E, for total expenditure.

C + I + G + NX = E

Now this equation is needlessly complicated for this part of the lecture. Let us add up I, G, and NX and call their sum “O” for "other," non-consumption spendin

I + G + NX = O

And write total spending E as:

O + C = E

Now look at consumption spending. It will be higher the higher are households’ incomes. And it will depend on the confidence that households have in the economy—which itself depends on how much of their incomes they expect to be taxed away by the government (with higher expected taxes leading them to curb spending), on whether they think that they need to boost their cash balances or not due to uncertainty about the future, on whether they have confidence that they will be able to borrow money if they need to or can afford to pay off the debts they currently owe, and other factors. So to start thinking about this, let us write down a very simple arithmetic rule for consumption spending:

C = C0 + Cy x Y

Consumption spending is going to be some number c0 times some other number cy times the level of total economy-wide incomes Y. The “cy x Y” captures the dependence of consumption spending on current incomes, and the “c0” captures all the confidence, tax, desire to boost cash-on-hand, and other factors. In the United States in 2010, the proper value to pick for cy is roughly 0.5: as a rule, if total economy-wide incomes fall by one dollar, consumption spending is likely to fall by fifty cents; and in the United States in 2010, the proper value to pick for C0 is roughly $3.5 trillion/year. Were economy-wide incomes to be $15.5 trillion/year, consumption spending would be $11.25 trillion/year.

So if we take our equation:

C = c0 + cy x Y

Substitute in $3.5 trillion/year and 0.5 for C0 and Cy x Y:

C = $3.5T/y + 0.5 x Y

And then substitute in $15 trillion/year for Y, we see that we get:

$3.5T/y + 0.5 x $15T/y = $3.5T/y + $7.75T/y = $11.25T/y = C

So why do we write these symbols “c0 and Ccsub>y”? Why not simply write “$3.5T/y” and “0.5”? Because as the economy changes over time those values will change. And those values do not apply to other countries. And those values can shift—especially C0 should consumer confidence collapse.

Notice that in writing this particular equation—this particular consumption function—we have once again followed economists' principle (or vice) of ruthless simplification.

In this complicated world, consumption spending does not depend on disposable income and confidence alone. It depends on a host of other factors—including the interest rates at which households can borrow, the values of people’s houses, the values of their 401(k) retirement accounts, the distribution of income across the economy, expected future income growth, risk tolerance, and a host of other factors. We hope that confidence and income are the most important one—but if e come across a situation in which other factors are the most important, there is no reason not to ditch this equation for another one that more accurately models reality.

2.2.4. Production, Spending, Output, and Income

Recall our equation for total spending E (E for Expenditure):

E = O + C

We can replace the “C” with our consumption function:

C = c0 + cy x Y

To get:

E = O + c0 + cy x Y

What happens in this model of the economy if expenditure E is greater than income Y? Well, by the circular flow principle income is the same as production, so if E is greater than Y then spending is greater than production—and inventories are falling. If inventories are falling, then businesses are hiring workers and expanding production, so Y is rising. What happens if expenditure E is less than income Y? Well, if E is less than Y then spending is less than production—and inventories are rising. If inventories are rising then businesses are firing workers and cutting back on production, so Y is falling. The only situation in which things are in balance and Y is not quickly changing is if:

E = Y

Then inventories will be balanced, and firms will be neither hiring and expanding nor firing and contracting. Thus the economy will very quickly spiral down in production and employment until it reaches a state where E=Y.

2.2.5. Where the Economy Settles: Equilibrium

Where will that be? We can see where the economy will settle, where its stable level of production and income will be, by doing some algebra. If we substitute Y in for E:

E = O + c0 + cy x Y

Since we would like to figure out what Y is, we should subtract Cy x Y from both sides to get it all by itself on the left:

Y - cy x Y = O + c0 + cy x Y - cy x Y

We can cancel terms on the right:

Y - cy x Y = O + c0 + cy x Y - cy x Y

We can gather terms on the left:

Y x (1 – cy) = O + c0

We can then divide both sides by (1 – cy):

Y x (1 – cy)/(1 – cy) = (O + csub>0)/(1 – cy)

We can cancel terms on the left:

Y x (1 – cy)/(1 – cy) = (O + c0)/(1 – cy)

And so arrive at our destination: our formula for what the economywide level of production and spending will be:

Y = (O + c0)/(1 – cy)

Thus to determine the level of economy-wide production (and income, and economy-wide spending) under conditions of depression economics, you follow a three-step plan:

  1. Add up “other” spending O—the sum of net exports, investment spending, and government purchases—and the “confidence” component C0 of consumer spending.
  2. Divide that sum by one minus the marginal propensity to consume—the number Cy that tells you how much consumption spending typically changes when economy-wide incomes change.
  3. You are done: PROFIT!!

2.2.6. How Well Does This Work?

How well does this work? Quite well, actually--impressively well for such a simple and crude model. At the depeest part of the recession--the third quarter of 2009--total other spending O--NX + I + G--was $487 billion/year less than its pre-2008 trend pace, and total spending E was $1,015 billion/year less than its pre-2008 trend pace. The fall in consumption spending set in train by the fact that households with unemployed workers and lowered incomes spend less had doubled the magnitude of the spending shortfall. This multiplier process had doubled the size of the recession over what it might have been otherwise.

[20100627 Keynesian #2707455.xls]Sheet1 Chart 1

There are trained professionals who do this for a living--and have high-paying jobs doing exactly this at a much more complex and sophisticated level.

2.3 What We Hope You Have Learned