Monetary policy with excess capacity

Here is some simple evidence of excess capacity and its relevance for unemployment.  I’ll give it more debate another time, and don’t too rapidly infer causal linkage from that graph behind the link, but at the very least it is not a crazy hypothesis.  Now that the gdp numbers have been revised downwards, and we see the U.S. economy has not reattained pre-crash output levels, the hypothesis that excess capacity is driving some of current unemployment is more plausible.

In one standard model, excess capacity renders nominal wage flexibility moot.  Workers could lower their reservation wages without it helping much.  Firms don’t want to produce any more, or they could produce more by working their current capital assets more heavily, rather than by hiring another worker.

Now let’s say there is a burst of inflation.  It might lower the real reservation wages of the unemployed, but employers still don’t bite, at least not until the excess capacity is worked off.  That can take a long time, especially if aggregate demand is low and the rate of innovation is sluggish.

There might be a stimulative effect if the nominal shock induces employers to expand output and, sooner or later, expand employment.  In other words, monetary policy must rely on the employers having money illusion.  Nonetheless employers are more likely to read financial news than are workers and arguably employers are less likely to be tricked by monetary policy.  Even if the employers are tricked, that just means they expand output, using spare capital, and work their way through the excess capacity more quickly; it doesn’t mean they hire more workers now.  Much will depend on capital-labor substitutability.  There is also the risk that employers will cut back on output once they see they were tricked by monetary policy; in contrast, workers who took a job under money illusion are not so likely to quit just because they pick up the WSJ and see that the nation has been suffering from hyperinflation or something like that.

Excess capacity is one reason why monetary policy isn’t always so effective, even when labor resources are unemployed.

Fiscal policy with excess capacity deserves a post of its own.  For now I’ll note there is a difference between hiring people to work directly for the government, contracting with firms to use some of their excess capacity, and contracting with the highest quality firms which perhaps do not have much excess capacity at all.

I again stress that microfoundations matter.  And please don’t read this post the wrong way.  I still favor looser monetary policy, but I view myself as lacking in real influence on the world and thus I am “working through the available variations” rather than propagadizing for my favorite policy.  If I were President of the United States, my blog posts would read somewhat differently.


What am I missing here? Why does working through excess capacity mean employers won't hire more workers? In my mind, along with underutilised capital, unemployed workers ARE a major component of this excess capacity. To me, excess capacity means that if there was an upsurge in demand, firms could expand output rather than raising prices because they don't need to buy more capital and they will have no difficulty finding workers willing to work at subdued wages.

What you just said, no price inflation, right?

Agree w/ Britonomist. I find this post very puzzling, since most accounts of "excess capacity" aiming for consistent microfoundations rely on monetary disequilibrium (price stickiness, excess demand for money). As such, monetary easing would relieve these conditions and restore stable/sustainable utilization of resources (which is the goal of macro-stabilization policy).

I think this old Scott Sumner post is relevant:

It would seem that we have excess capacity in housing and automotive. The human capital of workers in these industries is worth a premium in those industries but the wage adjustment is huge for these workers outside of their field. Automotive is correcting but housing is taking much longer.

Employers can easily substitute unskilled (or increasingly skilled) international workers.

Employers think they are very smart during boom periods. Smarter then they actually are. Employers think they know less during bad times. But they know more then they realize. So employers take more risks during booms and take fewer risks during busts.

Money illusion seems stronger in the housing sector (the illusion of increased wealth for owners and the lower interest rates lowering monthly payments i.e. the expected long term return increases if you think current interest rates (and home prices) are low compared to the future.)

When it comes to the employment picture, we do not have full employment because the cost of employment is above the current value. Because we have a fixed floor, (hourly wage, benefits, etc) we cannot have full employement. The simple answer to full employment is to lower the cost of it. Why is this so hard?


You missed the implication of these two sentences: "Much will depend on capital-labor substitutability ... I again stress that microfoundations matter."

Tyler's point, if I am reading him correctly, is this:

Consider the standard theory of the firm approaching a cost-minimisation problem for a given quantity of desired output. The quantities of capital and labour demanded depend on their relative prices and the degree of substitutability between the two in production.

Now consider an increase in the quantity to be produced where (a) there is at least some substitutability between capital and labour; and (b) the marginal cost of capital is zero while the marginal cost of labour is strictly positive.

That's not what excess capacity means. Excess capacity implies that firms are unable to turn their capital over quickly due to depressed demand (especially given imperfect competition, etc.). The marginal product of capital is exceptionally low, so its cost (adjusted for productivity, as in cost minimizing approaches) is higher.

"Capacity", as measured by the Federal Reserve when surveying industrial (i.e. mostly manufacturing, but also utility) companies is:

"the maximum level of production that this establishment could reasonably expect to attain under normal and realistic operating conditions, fully utilizing the machinery and equipment in place."

Capacity utilisation is then expressed as a ratio of actual output to the answer given.

In other words, it is a measure of the rate of use of *currently installed physical capital*, although it is confounded by the responding firm's beliefs regarding the substitutability of (physical) capital and labour.

Excess capacity (defined simply as 1 - Capacity Utilisation) is then the fraction of potential output as a function of currently installed physical capital that is surplus to demand.

Your phrase "unable to turn their capital over quickly" seems to imply some sort of financial capital?

At lower quantities of output, the marginal product of capital is typically HIGHER, not lower. You are perhaps confusing it with the REVENUE VALUE of the marginal product of capital?

"Capital-labor substitutability" is really only a concern for the long run, where the firm can adjust installed capital at will. Saying that the "marginal cost of capital is zero"--given less than 100% utilization--is quite a vacuous statement.

When utilisation rates are low, "quantity of capital deployed" is not fixed in the short run, by definition.

There is no monetary policy "trick".

Let's say there is a burst of inflation. Nominal wages increase, while nominal mortgage payments stay the same. Discretionary income expands by the difference. 5% is saved, 95% is spend. Aggregate demand increases. Capacity utilization rises. More capital, and at least SOME more labor, is employed.

Oh and by the way, the higher aggregate demand is PERMANENT, as long as future deflation is avoided.

You have only looked at half of the balance sheet. If real mortgage payments drop due to inflation then bondholders face a real drop in income and their AD drops to compensate for lower discretionary income.

All broadly true as a first approximation, but you're forgetting that the effects on households will differ markedly across the liquid-wealth distribution.

Since low net wealth is highly positively correlated with unemployment, the effect that you describe is at best only a secondary effect (after a primary one of an increasing number of jobs) for people at the bottom of the distribution, and as Tyler is (I believe) arguing, an increase in aggregate demand will not be met with a proportional increase in demand for labour.

To make matters worse, at the top of the distribution, households have a far lower marginal propensity to consumer and nobody is underwater on the mortgage anyway, so an increase in nominal income will not necessarily lead to a large increase in aggregate demand.

Yes, see also Tom Cullis' comment. But there are other transmission mechanisms which cause monetary easing to increase AD, such as the effect of expected future prices on money balances.

If the US had excess capacity in energy production, the US would not be importing energy with borrowed money.

If the US had excess capacity in people moving, the US would not have any lost time from congestion and grid lock and after storms.

If the US has excess capacity in storm response, the US would not have had losses from flooding or losses from drought.

Capitalism was the response to risk and loss - collective investment spread the costs of frequent individual losses of ships among all trading ventures to increase the collective capacity to trade, providing increased wealth to all. As well as increased use of available resources.

Out current US economy looks at the ten thousand deficient bridges and other structures as producing at 100% of capacity while a capitalist like Warren Buffett would weigh the wealth value of many of those structure at near zero - any of those bridges could fall in an instant, either structural failure or flood.

I think the US is like the 80 year old retiree who is letting their house become increasingly rundown. For them, the house is producing at 100%, or at least at 98%, while the objective view of the house is its value is less than zero because of the cost of removing it to begin again.

The US lacks the capacity to grow the food people eat (as opposed to the ag input to factories) so we borrow from other nations to import the basic foods which are inferior to the foods grown in abundance in the US a century ago.

The US lacks the capacity to turn the iron ore and coal we have into the steel we need, so we export both to other nations, and then borrow from other nations to import the steel we need.

The 80 year old owns the rundown house and 5 acres, and is supplementing her meager income by borrowing against the land she owns - she is decapitalizing, reversing the sacrifice of consumption early in life to buy the land, to build the house that productively serves her today, but soon she will be dead and their will be no capital left, as the house will have no value to anyone, and the land will be owned by her lender.

In my view, Reagan changed the perspective of the American economy to that of an old person who worked hard and sacrificed for decades, and now turns to relaxing and enjoyment and living off the accumulated capital from those decades of sacrifice.

In my youth, the leaders, Republican and Democrat, laid out great projects that could not be done by individuals, or even a bunch of corporations. Hoover had spent decades on developing the West, and in response to the high unemployment sped up the plans for the Hoover Dam, and that unleashed many other projects. Around the same time, the advocates of auto touring had been pushing for better roads, so from the 20s through the 60s, road construction had been pressed by multiple presidents, with Eisenhower most famously calling for the Interstate highways. Eisenhower and JFK sought to make war efforts more productive and relevant with nuclear power and NASA. Nixon made environmentalism a major project for the nation, followed by energy independence that was most clearly laid out by Carter. One can criticize the specifics of those initiatives, but they were forward looking to the future.

When Hoover proposed the Hoover Dam, it was going to produce huge excess capacity, especially in electric power, the means used to finance the water management.

Eisenhower proposed a highway system that would produce great excess capacity.

I remember the economists telling everyone that the steel mill construction driven by Asian government policies would create huge excess capacity in steel production, so much capacity that half the output would not be needed. Wall Street dismantled US steel production in response rather than invest in US steel production, so the US now imports instead of exports. Yet the world steel consumption today is twice what it was when the US was cutting back because productive capacity exceeded any conceivable demand.

The US has excess capacity only when looking back, and not looking forward.

Monetary policy can cause inflation which focuses minds to the future higher prices and that forward thinking drives buying and stocking up today to save money in the future.

"In one standard model, excess capacity renders nominal wage flexibility moot. Workers could lower their reservation wages without it helping much. Firms don’t want to produce any more, or they could produce more by working their current capital assets more heavily, rather than by hiring another worker."

Wow, it would be the exact opposite for me.

Let's put it this way. Let's say Ford has a Variable Cost of $15,000 per car, of which $10,000 is the direct factory labor going into the car. In 2007, the factory is running at full capacity. The equilibrium price for Ford cars is enough above $15,000 to keep Ford's factories running at full capacity.

In 2008, demand plummets for cars. How does Ford respond? Well, the other $5,000 of VC goes to commodities like steel and aluminum. The prices for commodities would go down and Ford drops the price of cars accordingly. But the $10,000 of direct labor does not drop in cost because firms never reduce hourly wages.

No matter the demand, at a low enough price Ford could keep its factories running at 24/7. Why doesn't it drop the price to that equilibrium, market-clearing level? Because it can't reduce hourly wages and Ford will not sell cars below VC. VC is not a sunk cost. Therefore, Ford will drop its prices but not enough to keep its factory running at full capacity. And what happens when its cars are overpriced? Volume goes down and Ford has to cut the workers hours, i.e. unemployment.

If a model says excess capacity has nothing to do with nominal wage flexibility, then the model is wrong. Excess capacity is cyclical unemployment and cyclical unemployment is excess capacity. The contortions libertarians (including myself a few years ago) to not admit the huge market failures in the labor market vs. other markets is astounding.

To put it simply, the market for labor is not the market for oil. If the market for oil ran like the market for labor today, then 10% of the world's oil would be sitting in tankers not being used, often for months at a time. Of course oil companies do not leave oil sitting in tankers for months. They sell the oil, no matter the price. The market for labor, by contrast, is not auctioned continuously like all other goods. It can only reset its hourly price very slowly, usually on an annual basis (vs. daily or sooner basis for oil). That creates a lack of wage flexibility which then feeds into prices above market-clearing equilibrium for 100% capacity, which then creates unemployment due to lack of demand.

Have a nice day.

Also, shifting more of the costs to capital does not help get the car's price down to market-clearing levels.

To continue the previous example, in 2007 let's say the market-clearing price is $20,000. Due to the huge demand shock in 2008, the market-clearing price (i.e. price to get enough demand to run the factory at full capacity) goes down to $10,000. While in theory Ford could get down to that price by firing most of its workers and replacing them with capital, in practice the depreciation on that capital would overwhelm the $10,000 per car.

Furthermore, whoever is selling the machinery that makes the capital faces the same issue with sticky wages. The machinery makers also cannot drop their costs below VC and so capital does not become cheap enough to make $10,000 cars at a profit. Therefore, the factory has to run at less than full capacity because of prices above equilibrium. The capital vs. labor mix at Ford will not change the fact that its cars are only selling for $10,000.

You are wrong about this

"To put it simply, the market for labor is not the market for oil. If the market for oil ran like the market for labor today, then 10% of the world’s oil would be sitting in tankers not being used, often for months at a time. Of course oil companies do not leave oil sitting in tankers for months. They sell the oil, no matter the price."

The majority of the world's oil does sit idle for months, years or even decades. Oil companies don't often sit on oil they have already extracted but the frequently change the amount that they produce in response to changing prices.

While labor is not exactly analogous to commodities it is a lot closer than you are making it out to be. Your oversimplified model ignores the possibilities of using machines to replace labor which the auto industry has done to a vast extent. They can use steel and electricity to replace (some) human labor. If the prices of robotic arms falls enough they can maintain production while cutting workers.

It's true that most of the world's oil sits idle for years, but most of that oil either has high extraction costs or its oil owned by cartels which have an incentive to artificially reduce the supply of oil. The idle labor of the unemployed does not have similar "extraction" costs, if you will, and have more similarity to oil sitting in tankers for oil companies. Oil companies have to keep some oil in inventories to handle fluctuating demand, but if they want to sell the oil, they're able to sell it for some price.

The unemployed, however, often cannot find work for any wage in their field. In my Ford example, let's say the equilibrium wage for plant electricians goes down by 10%. Ideally Ford would decrease hourly wage by 10% across the board. Indeed, if it was oil instead of electricians, ALL oil would sell for 10% less. But instead of reducing all wages by 10%, Ford is likely to reduce 10% of workers' wages 100% (or cut 10% of hours, working less hours is still unemployment).

It could also keep 80% of electricians working for the same wage and cut 20% of electricians wages by 50%. But this is why reservation wages do not really matter. If demand has gone down by 10%, it's FAR more likely that 10% of electricians become unemployed than for 20% of electricians to take a 50% wage cut. In the first case, it does not matter what those 10% of unemployed electricians will work for. Even if they work for one penny an hour, there will not be demand.

Those 10% of electricians could, theoretically, somehow go straight to the customer or find another company with lower wages. But all car companies are facing the same shortfall in demand and all car companies face the same inflexibility with their electricians' wages. Nor will a new car company start up overnight that pays electricians minimum wage. Therefore the electricians could not find anywhere to work, as electricians at least, no matter their reservation wages.

So, therefore, the only thing that could adjust for the demand shortfall for electricians is for the entire field of electricians to somehow turn over and all electricians' wages to go down by 10%. For things like car factories, which have big economies of scale and take a long-time to start up, it will take many years until new car factories arbitrage the difference in the employed wages vs. the unemployed's reservation wages, if it ever happens at all. In the long run, after all, we're all dead, and those unemployed electricians will most likely leave the labor force instead of waiting for a new car company to get off the ground.

Finally, I addressed labor vs. capital above. If the market-clearing price for Ford's cars goes down precipitously, no amount of capital investment will increase the new equilibrium price for Ford's cars after an AD shock. Also, the makers of robotic arms face the same issues with wage rigidity. While the prices of steel and electricity fall, nearly all sources of labor in big companies will not similarly fall. Capital replacement does happen somewhat, as shown by huge investment in software and automation, but capital just does not fall enough to get costs down to the new equilibrium prices for cars after a demand shock.

"Excess capacity is cyclical unemployment and cyclical unemployment is excess capacity."

Is it, really? Yes, if sticky prices in the market for goods are mostly due to wage stickiness. But one can imagine a possible world in which wages fall and labor markets clear (The workers all bought gold and T-Bills in good times as insurance against macro shocks, so they're not bothered by falling wages). But the price of Ford cars does not fall (Because the marketing dept. cares about that commitment, and printing new menus is too costly). So the excess capacity is still there, and there's still a recession of sorts.

Well, New Keynesian price stickiness may be somewhat of a factor, but I have a much harder time buying it then wage stickiness. Ford can increase incentives much easier than it can get the UAW to go for hourly wage cuts. If the UAW did go for hourly wage cuts, Ford would not have an issue increasing incentives or decreasing prices.

In fact, prices DID fall precipitously after the demand shock in 2008. As commodities such as steel and electricity became cheaper, Ford did pass on those cost savings to customers to remain competitive. But their factories still did not run at full capacity despite the price drops. That's because the wages and benefits inherent in every car could not be reduced to the point where prices at market-clearing level made sense. Instead it was much easier to cut workers than to renogotiate with the UAW. Same for, say, teachers in Wisconsin, who were fine with FIFO layoffs but protested for weeks when wage cuts (through higher benefit and pension contributions) became a possibility. It may be easier to cut the wages of non-union employees, but such wage cuts for even non-union employees are extremely rare and only happen if the company is pretty much near bankruptcy due to underutilized capital.

Every time I hear about government directly employing excess capacity, I laugh.

If liberals REALLY were worried about the unemployed, w'ed just pay current public employees less and hire the unemployed with the savings.

See Nick Rowe:

"Now let’s say there is a burst of inflation. It might lower the real reservation wages of the unemployed, but employers still don’t bite, at least not until the excess capacity is worked off. That can take a long time, especially if aggregate demand is low and the rate of innovation is sluggish."

Never reason from a price change! Why would there be inflation if AD is low and there is substantial excess capacity? Massive supply shocks? If that's your model, say so.

The economy has become so globalized that the Fed's measure of capacity utilization is often misleading: Walmart or Sears are much more likely to be ordering from manufacturers in China or Mexico than from the USA. It's much more difficult for the Fed to trick a Chinese manufacturer with inflation and even if he is tricked he will be hiring Chinese workers rather than the Americans measured by BLS.

Another short-term problem with capacity utilization numbers is that electric power plants have become steadily more important as manufacturing and mining declined. Severe winters or summers can influence demand electric power capacity enough to swamp the underlying economic trends for a while.

I agree with Mulp, that much of our supposedly excess manufacturing capacity is no longer viable because cheaper capacity exists abroad. The USA is in a bad situation in a globalized economy, where we are no longer competitive at the margin for too many products, and we have allowed ourselves to be trapped in a trade system that will cause massive trade deficits until the USD collapses BigTime.

From the point of view of the CEO of a mid-sized firm that both manufactures and imports capital equipment, Tyler is on to something. We're in a depression - revenues are off 40% because we don't have customers for the machines we make because our customers don't have customers for the products they make. Since we have CNC machine tools that are idle and the core staff needed to capture our institutional knowledge, we could increase production by 20% tomorrow without any new hires. When we did advertise for one position last month we got over 400 resumes, some extremely overqualified. Our experience is of course just one of the millions of microeconomic contributors to the economic disaster. BTW, several CEOs I have spoken to who have never voted for a Democrat have told me we need a new WPA.

I think this point by Job Creator is very important, where long supply chains amplify small changes in end user demand. For example, take an industry which makes products directly used by consumers (say washing machines). Normal demand is growing at 2% per year. As a result a certain amount of new capacity is added per year. A recession happens, consumers cut back a little, and demand growth falls to 1% a year. So the industry cuts back on adding new capacity. If you are in the industry which supplies this new capacity, you will see demand fall by 50%. And of course this industry also has its own supplier industry. Not only does this cause significant reduction in labor during slowdowns, it also causes problems with forecasting demand in the supplier industries. For these supplier industries, there literally is no price that they can sell their excess capacity at, their customers have all the capacity they need for at least two years (in this example). So even if money illusion exists and people allow their labor to be massively repriced, there will be no recovery in this industry until ultimate recovery occurs in the end user industry. The idle resources in the supplier industries could move to other growing industries - but which ones? Given the difficulty of forecasting supplier industry demand rates, the calculation by capital and labor as to where to invest is really hard and takes time (similar to the PSST story), especially if the slowdown is thought to be temporary.

If you have ever played the MIT Beer Game (google it),you will get a vivid example of the limits of being able to forecast as a supplier. Its amazing how, even given small or no perturbations in ultimate demand, the supplier industries whips around, desperately trying to make sense of the signals from the end user. The rate of change in expected demand is huge over fairly short periods.

In this environment, monetary expansion just causes inflation, with no demand increase.

Your mention of microfoundations is entirely empty in the absence of a discussion of monetary policy transmission mechanisms. A "burst of inflation"? Where? How? How quickly and evenly?

Comments for this post are closed