potential gains from trade are captured by someone, whicb is to say, that its impossible to make one person better off without making someone else worse off. This requires, among other things, that the economy operate at a point along the production possibility curve. Thus, by definition if the economy were operating inside the frontier, it would be possible to produce more goods and services. If this additional output were distributed to some consumers (or even one consumer) without reducing that available to others, we could therefore make at least one consumer better off vnthout hurdng anyone. Consequendy, operadng inside the producdon possibility curve is inherently inefficient because resources which could be used to sadsfy human wants are wasted. Figure XVI below illustrates the point. Consider the output given by point A. Any point to the northeast of this point and all points along the line segments AB and AC contain more of both goods X and Y Then if we were to move society to any of these points, we could distribute the extra output to any number of people without decrease ing that available to others. In short, we could make some people better off without making anyone else worse off. While moving from point A to point D improves welfare, point D is still inefficient because we could sdll make some people better off without harming anyone else by moving further to the northeast. Thus, anytime were inside the production possibilides curve, whether were at point A, point D or anywhere else, we can always improve welfare. Naturally to reach the frontier we must use all our resources wdsely, wasting none of them. As such, we would want the total cost of producing any given level of output to be as small as possible, a topic we further explore at the close of this Secdon and in Secdon VII dealing with labor demand.

However, operadng along the producdon possibility frontier is not in itself sufficient for efficiency: To exhaust gains from trade the mixture of goods and services must be those that are the most desired. For example, if at the current output people would be willing to give up some X to have a litde more Y, then the current output must be inefficient.

In Figure XVII q is not only the profit maximizing output, it is also socially optimal (efficient) in the absence of externalities, i.e., costs and benefits incurred by outsiders to the transacdon, which we cover in Section IX. Remember from the consumer surplus discussion that the price of a good reflects the subjecdve value or marginal benefit (MB) to the consumer of consuming another unit of that good. If we

assume for now that no externaliues occur, then outside pardes will be unaffected by the firms acdvides. Therefore, the marginal benefit (MB) to the consumer also represents the marginal benefit to the whole society because the MB to every one else is zero. Moreover, according to the definidon of opportunity-cost, the marginal cost (MC) of a good rrieasures the highest valued alternadve use of the resources used to produce that unit. This means that when price or marginal benefit (MB) exceeds marginal cost (MC), the value to society of added resources in this particular industry exceeds their value in any other, so we would want more resources to flow into the industry. On the other hand, when P or MB is less than MC, we would want resources to flow into another industry where their value is higher. When MB = P = MC, the value of the resources in this particular industryjust equals that of their highest valued alternadve use, so resource allocation is optimal. In pursuing its own interests the price-taking firm automatically produces the output at which price equals marginal benefit equals marginal cosL Nonetheless, in doing so it umvitdngly achieves a social opdmimi as well. Well have more to say on this subject when we consider monopoly and externalities in Sections VI and IX.

THE LONG RUN COST CURVES OF THE FIRM

In the long run, all costs are variable by definition. As a result, we do not need to bother about fixed cost, making die analysis much cleaner The general shape of the long run average total cost ( ) and marginal cost (MC) curves appears in Figure XVIII. All curves have the characteristic "U" shape. As before, profit maximization occurs at the output at

Fig. XVI The lnefficient:y of Operation inside the Production Possibility Curve

GoodY

GoodX

which P = MC. However, should price He below the minimum , the firm will incur a loss and therefore leave the industry rather than produce. Consequently, the long run supply curve of the competidve firm consists of its marginal cost curve above the minimum of its averce total cost curve. For example, if the long run price of a product is $12, the compeddve firm would produce the quandty corresponding to $12 along its long run marginal cost curve. The ponion of the marginal cost curve lying below the long run average total cost curve is irrelevant because the firm would never produce there in the long run. In our example, if $12 was less than the firms long run average total cost, the firm would leave the industry.

One notable feature of the long run average total cost curve is its reladon to various possible short run curves. Remember we generated the short run average total cost curve by starting with a fixed amount of capital and then adding labor undl there was enough to produce a given quantity q. The total cost obtained under this procedure was then divided by q to yield at that output. Repeating the procedure for every other output generated the rest of the short run average total cost curve.

In the long run we are free to capital as well as labor in order to produce a given quantitx- q. For example, in the short run we may have had onh one machine and had to reach our producdon goal by using it intensively, by adding, say, workers. In die long run we can perhaps reach the identical output by using two or more machines and less labor. Presumably, a different average total cost ( ) will exist for this different mixture of inputs. The average total cost associated with one machine, which we will

call capital and abbreviate as = 1, is given in Figure XIX. When we increase to 2, we get the short rtm average lotal cost curve that corresponds to ihis greater capital stock. Similarly, with = 3, we get yet another average total cost curve.

Now consider the output q*. This output can be produced ou any one of three different average total cost curves corresponding to the use of 1, 2, or 3 different machines. Which one should the firm use? Clearly, the one which results in the lowest average total cost. If the firm should produce q*, it will surely choose to do so with the smallest possible total cost in order to maximize its profits. Since total cost at q* is equal lo (ATC)q*, the firm will pick the smallest average total cost that can generate q*. This is associated with two machines in the graph. If we eliminate the irrele\-ant pordons of the short run curves in Figure XIX, we get the long run bumpy average total cost curve of Figure XX. If fracdons of inputs and outputs are available, the graph would be smoother, approaching the "envelope" sketched underneath. We work widi the smooth curve because it is so much

more analydcally convenient.

The falhng pordon of the long run average

-total cost curve is known as the region of economies of

scale. Thereafter diseconomies of scale exist. In

Figure XX economies of scale-exist up to the output q.

Some economists assert that economies of scale are

present when doubling inputs more than doubles out-

put. These two definidons are equivalent We can eas-j

ily see in Table III that a doubling of inputs whichj

more than doubles output will cause the long run]

average total cost to fall. For example, suppose that!

the firms total cost of producing 100 units is $500,

I Then its average total cost will be $5 per unit ($500/1 OOq). Now let the firm double its purchases of iinputs - capital, labor, raw materials, and so forth. Then clearly the new total cost will be $1000. If doubling-inputs more than doubles output, the new quan-jj.uty must exceed 200. Say it is 250. Then the average iJtotal cost will be $4 per unit ($1000/250q). This implies that average total cost is falling, here from $5 jto $4, which we set out to prove. Nonetheless, for rea-jSons lying outside the scope of this book, this way of inking about long run average total cost is contro-jVersiak In any case, we would expect lower average 3tal cost in the long run since the firm has more ppdons for reducing its cost.

USTRY SUPPLY SHORT RUN AND THE NGRUN

The differences between the short and long in industry responses to changes in demand remain [or consideradon. The important point is that in the lort run, but not the long run, both the number of ms in the industry and their capital are fixed, msequendy, as we show below, the price response to change in demand is greater in the short run, where-the output response is greater in the long run. In ther words, the elasticity of suppK is greater in the ng run because existing firms ha\e greater opportu-.nides for making adjustments and firms can enter or tve to further affect supply.

Consider Figure XXI, showing both the equihbrium of the representadve firm and that of the dustry along the inidal demand and supply curves of and Sj. As you vrill recall from Secdon III, the equi-um industry output, Qj, represents the horizontal

summation of the outputs of many similar firms responding to the industry demand of D]. Each such firm produces an equilibrium output of qi where price = marginal cost = minimum long run average total cost (P = MC = ). Clearly, both the firm and die indus-ury are simultaneously in short and long run equilibrium.

Now what happens if demand increases from Di to Dg? In the short run, the ability of the industry to adjust to the new condidons is limited by its fixed stock of capital. This means first of-ali-that new entrants would not be able to acquire capital for producdon. Hence, the number of firms is essendally fixed. But it also implies that existing firms expand output solely by hiring more labor. Thus, a movement occurs along the short run industry supply curve, Si in Figure XXI repictured below. (You will recall that supply is the horizontal sum of the existing firms short run marginal cost curves, which we emphasize by writing the Greek letter L, meaning sum.) The ne\v short run market equilibrium quantit) and price are Q and P respectively. For the representative firm, output is q2, where its short run marginal cost (SRMC) equals price.

At , however, firms are making super-normal

Table III Doubling Inputs and Economies of Scale

Tntfll Cost (TC) $500 $1000

Qugntity (q) 100 250

Average Total Cost (TC/o) $5.00 $4.00