Comments on economics, mystery fiction, drama, and art.

Friday, December 28, 2012

The Meaning of Efficiency In Economics

I just saw a link to a blog post, and the title of the blog post is "No, Christmas Is Not an 'Economically Efficient' Holiday (and That's Okay)."  The sub-head is "If you were to design, from scratch, a gift-giving holiday to warm the hearts of the coolest classical economists, it would look nothing like December 25. But gift-giving rituals have never been about the economics of efficiency ..."

So I'm an economist, and how do I react to such an argument?  Well, let's start with what the introductory economics textbook I just finished using (I did not choose it, so don't blame me), Macroeconimics, 4th Ed., written by Glenn Hubbard and Anthony O'Brien define economic efficiency (p. G-2):
"Economic efficiency   A market outcome in which the marginal benefit to consumers of the last unit produced is equal to the marginal cost of production and in which the sum of the consumer surplus and producer surplus is at a maximum."
Greg Mankiw, in the glossary of his intro econ text (Principles of Economics, 6th Ed., p. 834) defines efficiency as "the property of society getting the most it can from its resources."  On p. 145, he comes very close to Hubbard & O'Brien's definition:  "the property of a resource allocation of maximizing the total surplus received by members of society." 

I've always had a different take of efficiency.  My take is that an action, or a set of actions, is efficient if it achieves a desired--and desirable--outcome at the lowest possible cost.  (Ken Boulding, somewhere--I think in his remarkable book Economics as a Science--talked about "sub-optimizing," which he defined as "doing very well something you should not be doing at all.")  Note that the outcome, in my version of this, is key.  Unless the outcome is both desired and desirable, achieving it cannot be efficient.  I believe strongly that my interpretation is the correct one, even within economics.

Indeed, both the Hubbard & O'Brien version and the Mankiw version work--if the outcomes from mmarket processes are desired and desirable.  That is, if and only if markets are (1) competitive, (2) free from externalities, and (3) operating under conditions of complete information, markets will generate efficient outcomes.  As near as I can tell, that excludes any actually existing markets.

But my view is not the majority view in the profession, and it's easy to find examples.  Joel Waldfogel's article "The Deadweight Loss of Christmas" is one prominent (and fun) example (and was published in the profession's most prestgious journal, the American Economic Review):
"In the standard microeconomic framework of consumer choice, the best a gift-gver can do with, say, $10, is to duplicate the coice that the recipient would have made.  While it is possible for a giver to choose a gift which the recipient ultimately values above its price--for example, if the recipient is not perfectly informed--it is likely that the gift will leave the recipient worse off than if she had made her own consumption choice with an equal amount of cash.  In short, gift-giving is a potential source of deadweight loss."  (p. 1328)
The solution, of course, is to give cash.   But Waldfogel's analysis overlooks what the desired--and desirable--outcome is.  That outcome is for the giver to express his/her affection for the recipient, to demonstrate "caring" in a way that shows that the giver has thought about what the recipient will value, and for the recipient to receive the gift as an expression of that affection and caring.  The gift is not the thing being given, and to mistake the gift for the embodiment of the gift is a category error.  In fact, this explains why people are often--most often, when there is thought to be a close emotional relationship between giver and recipient--offended by gifts of cash.  A gift of cash is, in my terms, inefficient.

Efficiency in economics, then, is not about maximizing surplus (maximizing surplus is, under certain very restrictive conditions a consequence of efficiency), or about maximizing progits, or about maximizing Gross Domestic Product.  It is about achieving desired and desirable outcomes.

Thursday, December 20, 2012

The Intemperate Discourse Continues

I noted a couple of days ago that the folks on the Society for the History of Economics listserv were having something of an intemperate discussion about Keynes and the clasical economists.  We branch out now into a sideshow about Say's Law, as one of the participants writes:

First, I don't recognize what he calls Mill's "'temporary' deviations from Say's Law" as deviations at all.  The law says there cannot be an overproduction of all goods at the same time. [John Stuart]Mill is simply pointing out that money (cash or currency) is also a good (or commodity).  Thus, should there be an excess demand for money (or insufficient supply) there would be an excess supply of all other goods and a fall in the price level. 

Well, perhaps the potted version of Say's Law ("Supply creates its own demand") says nothing more than that the income generated in production is euqal to the value of the goods produced (which remains a truism of national income accounting).  But that is not what one takes away from a more, um, complete reading of Say's work (have a look at A Treatise on Political Economy, if you're interested). 
Elsewhere, Say argues that a surplus of even one commpdity, and a corresponding shortage of another commodity, cannot exist except very temporarily, because prices will adjust to eliminate both the surplus and the shortage.
Why is this important?  It suggests that all markets return to equilibrium quickly and with no frictions.
Let us return, then, to the argument that being made above--that Mill says that money is also a commodity.  Frankly, I think this is a mis-reading of Mill, as it is surely a mis-reading of Say, who argues that
Money performs but a momentary function in this double exchange; and when the transaction is finally closed, it will always be found, that one kind of commodity has been exchanged for another...(p. 134)
This is a fairly clear statemnt of the conclusion reached by classical economists that money is neutral, that the amount of money in an economy has no effect on the underlying, or "real," economy (money is neutral).  If the amount of money in the economy increases, then the general level of prices inreases--we have inflation.  If the amount of money in the economy decreases, we have deflation.  Because money is valued and used for its function as a medium of exchange, it does not make much sense to take about an "excess demand for money."
What is an "excess demand for money," anyway?  In general there is an excess demand for a good if, at the existing price, people are willing and able to buy more of it that is made available for sale.  So, if there is an "excess demand" for money, peple will be willing and able to buy more money than is available at the existing price.  What does it mean to "buy" money, in this context, and what is the "price" of money?  In a classical explanation of the world, people "buy" money by selling goods for money.  And the "price" of money is the quantity of goods it takes to acquire a given quantity of money.  In the aggregate, the "price" of money is the reciprocal of the general level of prices.
So how do we get an excess demand for money?  Assume that there is a central authority that controls the amount of money.  If that authority reduces the amount of money (the supply curve of money shifts to the left) and the price of money does not rise (which is the same thing as saying that the general level of prices does not fall), there will be an excess demand for money.
(In the diagram above, the excess demand for money is shown by the red line segment.)  Note that, for the excess demand for money to be resolved, the price of money must rise to its new equilibrium level--which is the same as saying that the general level of prices must fall.  Now, Say and Mill and the other classical economists concluded that this increase in the price of money (fall in the general level of prices) would occur quickly, that the excess demand for money would be quickly remedied.  Or, explaining it differently, and using the quantity theory of money as a framework, we would have:
M*v = P*Q

(Where M is the quantity of money supplied, v is the "velocity of money" (roughly, the number of times per time period the "average" unit of money is spent), P is the general level of prices, and Q is an index of the total output of goods.)  Classical economic theory concludes that v is (relatively) constant.  So, if a decrease in M leads immediately to a corresponding decrease in P, the quantity of goods produced will remain constant.  And, hence, money is neutral.  Say believed money is neutral.  Mill believed money is neutral.  That was one of the underlying tenets of the classical system.
But what if the general level of prices does not adjust quickly?  (What if prices are sticky, especially downward?)  Then, with a relatively constant velocity, Q must fall.  Oh, wait.  That's, essentially the argument Keynes made.
(More in a while.)


Wednesday, December 12, 2012

Keynes, the Classicals, and Capital

There’s been a fairly acrimonious conversation going on at the Society for the History of Economics listserv, revolving around what Keynes had to say about “classical” economics, whether what he said was fair or accurate, and what this means for how we should interpret Keynesian economics.  [Some of the participants suggest (a) Keynes’ interpretation of classical economics was almost wholly wrong and (b) as a result Keynesian economics should be almost wholly rejected.  Others reach pretty much the opposite conclusion.]  One specific area of contention is the meaning of Say’s Law within classical economics and resulting interpretation of Keynes’ theories of capital and of the rate of interest.  While I don’t expect to settle any part of this debate, I have a long-standing interest in classical economics (my understanding of it has been shaped mostly by Samuel Hollander’s Classical Economics and Thomas Sowell’s Classical Economics Reconsidered and in Say’s Law.  My conclusion is that the parties to this discussion have been talking past each other, for one primary reason:  The world changed from one in which one conception of capital (and therefore of interest) was more-or-less valid to one in which it was not.

Consider the classical conception of production.  It’s a world in which there is a discrete period of production, in which resources are combined to produce an output.  During that period of production, entrepreneurs have to pay for the resources they use.  At the end of that period the output is sold.  Then, entrepreneurs have to decide whether to undertake a subsequent period of production.  And periods of production are not overlapping.  The implication of this is that the entrepreneur must have on hand at the beginning of the period of production the means of paying for the resources used during production.  As a consequence, the classical conception of capital includes whatever it is that serves as a means of compensation for the resources used during production.  One way this shows up is in the concept of the wages-fund, which is a form of capital used to pay labor during the production period.  (It also shows up in the concept of circulating capital.)  Where in the world is this conception of the production process (and the cycle of production) valid?  In agriculture, and particular in the cultivation of annual crops, like wheat or oats or corn.  And in the late 17th and early 18th century, one could argue that this was a reasonable way to think about production. According to this study, for example, agriculture accounted for more than 70% of total employment in the US in 1800 and a similar share of employment in England.  So we have, in both countries, a production period that is (1) long—specifically, longer than the pay period for labor—and (2) sequential and non-overlapping.  Producers need to know how they will pay for (for example) labor (and other variable resources) before undertaking production.  Or, in other words, production is not a flow; it occurs at the end of the production period as a stock of output.  Meanwhile, variable resources are a flow and must be paid during their period of use.

Now let us fast-forward to the 1920s.  By 1930, agriculture accounted for less than 25% of US employment and less than 20% in England.  How does this change things?  First of all, we now have a situation in which periods of production are short—shorter than the pay period for labor.  Secondly, we have overlapping periods of production.  In automobile manufacturing, for example, a producer need not finish one production run of vehicles before beginning another.  So the cycle for producers has changed.  Payments for labor (and other variable resources) need not be funded in advance; rather, these payments can be made from an ongoing flow of revenue from the sales of the firm’s product.  In short, production does not, in the “modern” world, result in a stock of product at the end of a production period that is then sold.  Production is now a flow of product (the production period, again, is shorter than the pay period for variable resources) the sale of which can be used to pay for those variable resources.  Firms will tend to hold stocks (inventories) of finished goods in order to be able to make sales at any time; firms will also tend to hold stocks (inventories) of intermediate goods in order to facilitate the ongoing flow of production.  So inventories (and inventory finance) can—and is—still considered to be a part of capital, which labor is not.  The concept of a pre-existing fund used to pay labor during an extended production period is, in fact, no longer useful, or valid.

Keynes wrote in this later period.  And while he may well have mistakenly applied the (relevant) modern concept of capital to the period in which classical economists wrote, I don’t think he was wrong to introduce a new concept of capital, which required a new concept of interest, and a new concept of the position of labor in the production process.

Monday, December 03, 2012

The "Standard" View of the Production Function

I've been away for a while, settling into retirement and trying to get three home remodeling projects finished (which they--finally--are).  But I've been reading.

And one of the things I read recently was a blog post by Nick Rowe, which linked to a blog post by Steve Roth, in which Roth argued that the standard view of the (microeconomic) production function (specifically as presented in Mankiw's intro textbook), specifically the notion that marginal cost rises in the short-run as output increases (in the presence of one or more fixed resources)

"...1. is ridiculous on its face, 2. is completely contrary to how profit-maximizing producers think, and 3. is based on just-plain incorrect math.* It’s just one of many central pillars of “textbook” economics that are still being taught with a straight face, even though they been resoundingly disproved by the discipline’s own leading practitioners, on the discipline’s own terms, and using its own language, constructs, and methods."
Why does he reach this conclusion?

"The rising marginal cost theory is ridiculous on its face because it assumes that producers add one factor of production at a time — hire more workers, for instance, without renting more space for them to work. (This is exactly what Mankiw describes in his textbook; see “Thirsty Thelma’s.”) Voila! Each worker’s productivity declines. This is of course not what producers do (my emphasis), which is why only 11% of top-corporation execs say they face rising marginal costs of production. (I’m wondering if that 11% made the mistake of taking an intro econ class in college.) For a nice recap of that executive survey, and the faulty math of rising marginal costs, see here (PDF)."
In fact, this is exactly what producers do in many, many instances.  Examples of producers doing exactly what Roth says they do not do  are all too easy to find:
  •  a gas station owner observes that her late evening sales have been rising and decides to see what will happen if the station remains open until midnight instead of closing at 10 PM.
  • a fast-food franchise adds additional workers during the late afternoon hours when it notices an influx of students stopping by on their way home from work.
  • retail stores add additional sales staff during November and December.
  • an indoor tennis faility extends its hours to accomodate increasing demand (and reduces its hours when demand falls).
  • steel companies increase (and reduce) their output (and use of labor) from existing production facilities.
Furthermore, we can extend this to non-labor inputs, by looking at some of the classic studies of a production functions, which I first ran across in Edwin Mansfield's intro to micro textbook.

    • studies of the use of fertilizer in the production of corn, where fertilizer use per acre of land was the variable in question, found a declining marginal product of fertilizer.
    • a study of throughput of an oil pipeline found declining marginal product from the use of higher pumping pressure in an existing pipeline.
All of these are examples of firms altering output without adding new capital equipment, or expanding existing facilities.  Is it even remotely plausible to think that US Steel, faced with an immediate ability to sell additional steel, would refuse to expand output until it had enlarged its steel mills?  Or that USS would reduce its output only by removing some production capacity?

Asking "top corporation execs" about their marginal costs, frankly, seems to me to be a fool's errand.  Most corporations do not seem to analyze their costs in that way, and there's no particular reason why people not involved on a day-to-day basis in cost analysis or pricing decisions would think in those terms anyway.

In part the difficulty here is in using a textbook presentation of a fairly complicated subject (the nature of production and a production function) as one's exemplar of how economists think about production.  When I teach this, I teach it as an abstraction, telling students that we reduce the problem to two resources (labor and capital) so as to be able to draw two-dimensional diagrams.  We know--believe it or not--that virtually every production process uses more than two resources--that, for example, when a power company increases its output of electricity from its existing generating facilities, it uses more fuel as well as more labor (and more of a lot of other "variable resources").

Capital is not quickly or (in a lot of cases) easily added or subtracted.  And production is not a case of expanding all resource use in the short or long run in fixed proportions (as, for example, input-output analysis would have it).  Furthermore, substitution possibilities constantly present themselves to firms (especially in the long-run, but also in the short-run), as in McDonald's making a (long-run) change to automatic-shut-off soft-drink equipment behind the counter and substituting the (unpaid) labor of their customers for the (paid) labor of their employees by placing soft-drink dispensers in the dining areas.  Or gas stations moving to pay-at-the-pump, which allows them to operate with fewer cashiers (or even to remain "open" all night with no cashiers.  Short-run substitution choices include, for example, power plants with the capability of using alternative feuls substituting toward the one whose relative price has declined.

Neoclassical microeconomics has  its problems, but the concept of fixed resources (in the short-run) and rising marginal costs are not, it seems to me, among them.