CHAPTER

TWO
GENERAL EQUILIBRIUM

We turn now to positrve economic theory. Its main purpose is to provide a framework for predicting the effects of particular causes and for detecting the causes ofparticular effects. We might, for example, want to know how immigration would affect relative prices and thus affect the welfare of different classes among existing residents. Or we might wish to predict the effects of some proposed tax, or of an import quota. A full analysis of any of these questions would require a model with many industrial sectors (products), many types of labour, and so on. However, a model with only two products and two factors provides a great deal of insight into many of the basic forces determining the welfare consequences of economic change. We, therefore, confine ourselves, for the moment, to the two-sector model, developing it in this chapter and applying it in Chapters 3 and 4 to problems of public finance and international trade. If the framework 'is to be used, it must be checked for consistency. It claims to explain what we observe, i.e., a price system in which relative prices are fairly stable and only change if circumstances change. We need to check this claim. For policy purposes, we also need to know whether the equilibrium of the system is unique, for, if it has more than one equilibrium, only one of them will be the solution to the welfare-maximising problem with which we began. So this chapter has two interrelated purposes: 1. We prove the existence and stability of general competitive ..equilibrium and consider whether it is unique. This question is the main focus of Sec~. -2:r and 2-4.
52

GENERAL

EQUILII3RIUM

53

20 Weshow how thedistribution

ofincome is determined and how it would change with changes in factor supplies (due, for example, to immigration) or in international prices. This is the main focus of Secs, 2-2 and 2-30

lt is convenient to stick to the same analytical sequence that we followed in Seco 1-20 So we first consider the case of consumption without production (Seco 2-1), then that of production without consumption [Secs. 2-2 and 2-3), and finally the general equilibrium in which both are determined together (Seco 2-4)0

2-1 CONSUMPTION WITHOUT PRODUCTION (PURE EXCHANGE)
Bargaining
'e start with a two-person econorny. Manna (x) and quails (y) rain down from eaven, most of the total y falling into the possession of A and most of the total x into the possession of B. So the initial endowment is at E in the Edgeworth x in Figure 2-1. The same endowment arrives every day and has to be consumed , e same day. If A consumed his endowment, his utility would be UA1 and B's would :.. J300 But on this diet A values x more highly than B does, as can be seen z, their respective marginal rates of substitution. (Remember that B's marginal - of substitution, like A's, is rneasured by the slope of his indifference curve , h respect to the x axis, i.e. by - dyidx, even though his consumption bundle -- za asured by its distance south and west of OBo)Thus MRS;< >MRS:x ~
z: -:

,00 By consuming

ihere is an opportunity for fruitful trade in which A gets more x in return at any point in the shaded area, both A and B could be

OB
~----------------------------~

x

B

-----1
Figure 2-1 Pure exchange.

---------------x------------ __

54

WELFARE

ECO

OMICS AND

GENERAL

EQUILIBRIUM

better off than at E-in this sense alI these points "domínate" E. But trade wilI only occur as the individuals learn something about each other's preferences. At first they may not realise how great the difference is between their valuations, and consumption may occur quite near E. But then further gains from trade wilI be available, and eventualIy trade wilI settle down at some point on the efficiency locus of the economy where neither one can become better off without the other becoming worse off. ClearIy this point must make both of them better off than at E, so it must lie on TO TI, which is calIed the eontraet eurue. But where exactly it lies on this curve depends on the bargaining strength of the two parties. If Crusoe (A) is able to push Friday (B) to the limit, they end up at TO, and in the opposite case at TI. But whatever the point, at least one of the parties wilI be better off than he would be with no trade. The indeterminacy of the final solution here arises from the fact that we are dealing with bilateral monopoly, where there is a single selIer and a single buyer.

Existence of Equilibrium
Now assume instead that we move to a competitive market. A now consists of a large number of identical consumers with identical utility functions and identical endowments. The indifference map for A now shows the rate of substitution for each consumer when total consumption of group A is at each particular point in (x, y) space. SimilarIy for group B. Now what happens? For expositional purposes we introduce at this point a fictional character, though he is dropped later. This man is a kind of experimental social scientist, usualIy calIed an "auctioneer." His job is to calI out the day's price (PxlpJ. For convenience we shalI calI this p. There are three questions that interest him: 1. Is there any price that will cIear the market; i.e., does equilibrium 2. Is there more than one such price; i.e., is equilibrium unique? 3. Will the equilibrium be stable? exist?

To tackle these questions he calls out a series of prices on successive days. Suppose he happens to start with the price pO, corresponding to the slope of ETo As Figure 2-2 shows, A indicates that he would like to consume at CA, so that he would be buying GCA of x on the market, financing this by the sale of GE of y. (This construction is sometimes called his tradinq triangle.) By contrast, B wishes to consume at CB, selling only FCB of x and buying FE of y. So there is excess demand for x. This can be seen in two ways. Viewing the matter in terms of quantities traded, we find that market demando GCA exceeds market supply FCB. Alternatively, we could assess the total demand for x by referring the desired consumption levels at CA and CB to their respective origins, and then compare this to the total available supply. The two procedures are equivalent. They reveal net excess demand for x and excess supply of y. These two phenomena are cIosely related. For any good z let us call (he

t Whenever we write an absolute price. For any individual the sum of his excess demands. .-----------------------~----. There is no store of money in this book. Walras' Law. summed over al! markets. So ij' one market has positive excess demand. and if al! but one are in balance. it measures his market demand for z. it measures his market supply. so is that one. weighted by the prices of the goods.. This is Walras' Law. another must have excess supply.~~ lslope] = pO I . which requires that the value of his market demands must equal the value of his market supplies. must equal zero. _-------x-------_ Figure 2-2 A's and B's offers al (pxfp. OB I F I L_________ l. rather than a relative price. must be zero. difference between the ith person's desired endowment i his excess demand for z (EDZi): consumption Zi and his original EDzi = Zi - zi If this is posinve.~-.) = po. we obtain or PxEDX What holds for all individuals + pyEDY = o holds for the nation. CBI o T· : /lB3 CL-----------------.XA) and + plvA ~ yA) = o Adding them. The SUI11 of price-weighted excess demands. as here. the units of money should be thought of only as units of account.GE ERAL EQUIUBRIUM 55 T . Sor pAXA . This follows directly from his budget equation. if it is negative.

Accordingly.e. by Walras' Law. he will choose a higher price (p)Py) in the hope of choking off some of the excess demand for x. negative excess demando (The reader can check why. = O.. turning to Figure 2-3.. At price po the market is not cleared. In other words. Now he finds excess supply of x. Equally. So the argument in the text stands. .. Suppose he chooses p". excess supply of x. there are enough equations to determine the structure of relative prices.fPx = O.fPx = O implies. = e(small). At that point the indifference curves are both tangential to the price line at the same point. Let us concentrate on the market for x and resume our story. and excess demand al eJ». Ir so. if the relationship is expressed by a continuous function linking points A and B. For if we have only two markets. so the auctioneer declares all existing offers to be void and tries another price. If not.. excess demand for y at »J». They are therefore tangential to each other. by continuity. excess demand for J' at p. that is fine: an equilibrium exists. readers may wish to satisfy themselves that if there are more than two goods. and. = O or al p. If so. an equilibrium price exists. This makes the analysis quite simple. equal to the slopeof ET1 in Figure 2-1. t But how is the auctioneer to reach it? He could of course go on calling out prices at random.re» in Figure 2-2.56 WELFARE ECO OMICS ANO GENERAL EQ ILlBRIUM B -------'--------- LD"' Figure 2-3 The existence of an equilibrium relative price. i. there must be some price at which excess demand is zero. excess demand in one implies excess supply in the other. as we rotate the price line from the slope of ETo to that of ET1 there is some point where A's wishes and B's coincide and an equilibrium price exists.. It follows that. we have also established it in the other. If he has any sense. we should have allowed for the possibility that equilibrium occurs al Px/P. as we saw in Figure 2-2. Note that ED"'(pO) equals Ge' .) So. they can turn to Appendix 5. and the point is on the contract curve.Jp. once we have established equilibrium in one market. we have now established two observations on the relationship between price and excess demand: at one price excess demand is positive and at another negative. On a different issue. reverting to Figure 2-1. but he would then only t Strictly. So we have excess supply of x al very high values of P.

reduce it. whatever his initial price. So we can still be sure that the equilibrium will be stable-at pl or v'. In our particular example. provided that the economy had ever experienced any kind of disturbance. there are three equilibrium prices." His "response ~ function. Clearly the number must be odd.also showed how.GENERAL EQUIUBRIUM 57 hit on a solution by chanceo Alternatively. will be unstable. but if the bowl is upside down and the ball is perched upon it. And the evidence of history suggests that competitive markets do indeed respond in this way. This mechanism. for we can perfectly well trace out a path from A to B in Figure 2-4 that gives zero excess demand at more than one price. will equilibrium be stable? By this we mean: if something temporarily disturbs the equilibrium. he would be operating in the manner of an automatic control system. will the underlying forces tend to restore equilibrium or to produce further disequilibrium? (Note that a ball resting inside a U-shaped bowl is in a stable equilibrium.y was positioe. Which of these prices will rule depends on where the system started from. who was among the first to formulate the theory of general equilibrium at the end of the last century. (That this is an appropriate response function can be checked by seeing what would happen if the opposite rule were followed. Excess demand is his "target (or signal) variable" and price his "instrument variable." which governs the response of the instrument to the state of the target variable. We have shown that his job can be done quite well without him. Uniqueness of Equilibrium So an equilibrium exists that is stable. a market clearing price would eventually be found. If excess demand for a commodit. if it is disturbed. (Check why. is that posited by Walras. provided that there is an automatic response ofthe kind specified. like p2.) So the time has come to sack the auctioneer. he would raise its relative price. But how likely are there to be multiple equilibria. In this way. Moreover. He. In proceeding like this. he would be bound ultimately to converge on an equilibrium price. as in this example? To .) The question can be swiftly answered: The control mechanism which brought about equilibrium in the first place will restore it. every alternate equilibrium price. But will there be only one such equilibrium? Nothing we have said so far rules out multiple equilibria. whereby the first response to disequilibrium is a change in price rather than quantities. is (to repeat) the following: if excess demand is positive. and if negative.) But there is no reason why the economy should find itself at p2. by the process of tiitonnement (or groping). as with all dynamic questions. raise price and vice versa. the ball's equilibrium is unstable. Stability of Equilibrium This raises the problem. he could adopt a system.

t The priee consumption line traces the locus of quantities (x.. y) which an individual A endowed with (. y !!. investigate the question. If we treat E as the origin.l». =p! Py L---~--4------~--~~----.. we may cal! the line his offer eL/rue. even though the markets in which they trade is cornpetitive.x"'.. t For sirnplicity the ex position in the next few paragraphs proceeds as though single individuals. the coordinates measuring his excess demand for each good. his excess demand for the other must be negative (he supplies it in the market). Notice again ~hat if his excess demand for one good is positive (he demands it in the market).x Figure 2-5 A's offer curve.O EDx Figure 2-4 Multiple equilibria.1. yA) will consume at each relative price v. A and B were .:. we need to introduce a new geometrical device (see Figure 2-5)..58 WELFARE ECONOMICS AND GENERAL EQUILlBRIUM Px Py ______ .

as we shall see.x were backward-bending and the supply curve of y were not. with OB in the bottom lefthand corner. as PY/Px rises. the "total" effect will be negative. the individual who is buying x becomes worse off. falls then involves the payment of less y in return. Now suppose that neither A nor B have U-shaped offer CUl:V~S. Now take away from him the in come compensation ~y. Thus. at C. Figure 2-7 makes it clear that both are necessary. if we make the following experiment. unless the in come effect is absolutely larger than the substitution effect. But.. i. The supply of y is more complicated: supply is always more complicated and liable to be backward-bending. t By drawing a simple supply and demand diagram for x one might think that multiple equilibria were possible if only the supply curve of . in Figure 2-5. Beyond here the supply of y is now reduced when pJpy falls (or in other words when p)Px rises). rises. When the price rises to p2. When p rises.an equilibrium exists wherever the offer curves of A and B cross. However. .t A's offer curve is the same as the one we derived t Demand for x is inelastic if. after a point. as p rises the demand for x falls continuously. At the equilibrium price. A's supply curve of y is first rising and then backwardbending. So let us give him enough extra y (~y) to enable him to maintain his original level of utility ul. but at the new price p2. which are also at C. So. i. 'Ihis change can be thought of as the sum of two effects.. This shifts him from e to e2: this change in x is called the in come effect. so that total expenditure (px) falls.\So demand and supply of both goods are in balance. This is illustrated in Figure 2-6. he desires x2. The curve of B is similar in form to that of A. so there must be a unique equilibrium. At p' he desires Xl of x. it is one for which demand falls when in come falls and vice versa. and we shall ignore it. Then the curves can only cross once.] However.e.] the increased quantity of x demanded when vJv. the consumption demands of A. the substitution effect of a price rise must be negative (or zero]. So his demand falls. But what has this to do with the uniqueness of general equilibrium? Everything.GENERAL EQUILIBRIUM 59 Now let us see how his market demand (for x) and supply (of y) vary as vJ». are consistent with the consumption demands of B.y) rises . which is less than x 1. The in come effect of a price rise will also be negative if. the demand for x may become inelastic.e. Whenever the offer curve is U-shaped. The total effect of the price rise is thus the sum of the substitution effect and the income effect: e Total effect = substitution effect + in come effect As we show in Chapter 5. even if the income effect is positive. when p falls the proportional rise in x is less than the proportional fall in price.Il'his time we shall ask what happens as vJ». This so-called Giffen case is extremely rare.\At first. the consumption of y falls so that supply (yA . He will now consume at C. But now suppose that both offer curves are U'-shaped (as In Figure 2-7). 'This happens at point e* in Figure 2-5. the good is normal. as p. falls. t:Qr .. this is implied. and only if. as the relative price of y rises./Py falls towards p3. as you can see by looking through the paper at Figure 2-6 from behind. The change in x as we move from el to is called the substitution effect.

And the test of stability is that a rise of pxfpy brings forth an excess supply of x. v'. _But how likely is such a situation? If wherever there is equilibrium it is stable. as in Figure 2-4. if x is normal for him. Demand is reduced. For both A and B there is a substitution effect away from x. . At each of these prices. A becomes worse off. there can only be one equilibrium (see Figure 2-4). . B is better off. so that his income effect also leads to reduced demand. 3. However. and p3. so that his income effect leads to increased demand. the change in demand for x is as follows: 1. So we cannot say unambiguously whether the demand for x falls. And the two offer curves cross each other at three different prices pl. Figure 2-7 Multiple equilibria. the r-------------------------------. 2. in Figure 2-5. What determines whether it does? When pxfpy rises. demand and supply of x are in balance. 08 __-------------x---------------.60 WELFARE ECONOMICS AND GENERAL EQUILlBRIUM Px l Slopel =- Py Figure 2-6 A unique equilibrium. if x is normal for him.

it may not be a sufficient argument against it to say that an even better state could . But they could choose vJ». of course. In our example group A. force group A to consume off A's offer curve. so the excess demand function shifts steadily to the right. Then suddenly it leaps to above p3 There are no well-documented cases of such discontinuities and we can probably sleep safely without worrying about multiple equilibria.yx = Px > MRSB Py yx So it is inefficient. while group A members are dispersed. (Try shifting it over. It is important to know whether multiple equilibria occur in the real world. Instead. would like as high a price as possible. that is. however. would Iike as low a price of x as they can get. But of course it may or may not be ethically superior to point R depending on the needs of the two groups A and B. . So. provided that A and B have similar marginal propensities to spend o~ of additional income. This does not mean. who are buyers of x. For example. Coalitions and Monopoly . In Figure 2-8 this point is at Q. they could choose a point where A's offer curve was tangential to one ofB's indifference curves. for if they do we might be able to improve social welfare by shifting the economy from one equilibrium to another. But it is generally not in the interests of any one group of individuals to act as price takers. So for multiple equilibria to occur~there must 6emarkedly different income effects for A and Band one of the income effects must be substantial.GENERAL EQUILIBRIUM 61 i!tcome loss of A (his Lly referred to above) is approximately equal to the income gain of B. who sell x. given by the slope of ER. The relative price of x is much higher than the free-rnarket price. that the equilibrium which actually comes about will necessarily be a competitive one-it merely says that. a competitive equilibrium will result. suppose that the excess demand function of the economy were initially as shown in Figure 2-4 and the equilibrium price were pl. Then people progressively acquire a stronger taste for x. the two income effef~s (2 and 3 above) mor_~or less cancel out. so that A vo~ntilfily consumed at that point on A's offer curve that maximised the welfare of B. Suppose that all members of group B live close to each other and can organise themselves collectively. as any steelworkers' union leader will tell you. How will they settle the optimum price? They could not. Point Q is not on the contract curve since MRSA . leaving the negative substitution effect. Then there is no reason why group B (the 'OPEC countries) should behave as price takers. they can agree among themselves on the price at which they are willing to sell (oil) and on the quantities of x each one of them shall be allowed to exchange for y. Similarly group B. We should have positive evidence of multiple equilibria if we observed sudden jumps in the economy over time. if individuals act as price takers.) At first price rises slowly and smoothly until it reaches p5.domínant. If it is ethically superior.

whether imposed by unions or by government. We have already seen how some of the costs arise: energy needs to be expended in arranging the carve up. Of course. there are two obvious problems with actions ofthis kind in restraint of trade.__ - -------' pr icc O. At the high. "rnonopoly.----------------------------.\~xA Px u. this must be because the costs of cooperation between members exceed the benefits they obtain. Then if B was endowed ith all the x in the economy. happen. there is the basic problem of carving up the gains accruing to the colluding group (B) among its members. Second.62 WELFA~CONOMICS AND 'GENERAL EQUILIBRIUM T yA . r Clearly. For the moment we simply note that perfect competition will only be found where the transactions costs of collusion exceed the gains fram collusion or where the law is ver y strang." for the tangency between the price line and one of B's indifference curves occurs to the southeast of Q. some members of group B may suffer." union bargaining. be they "optirnal tariffs. we should expect to find such collusive behaviour (sometimes called "cooperative " behaviour) occurring on a wide scale. we need not observe perfect competition. . If the gains cannot be distributed equally. but others may lose from the reduced amount of labour (here x) being bought. OB A's offcr 1 '--------------+. One individual may be a natural monopolist. he would insist on a price exactly analogous to that set by a coalition of individuals who collectively owned all of x. if groups of people can gain by formmg coalitions.. SO quotas will have to be allocated within the group. if these would not. or whatever. We shall revert to imperfect competition in Chapter 8. If it does not. _This is the kind of problem that arises with a minimum wage. in fact. be achieved by lump-sum transfers or by taxes and transfers. But fram now on we shall assume perfect competition. = monopoly Figure 2-8 B's monopoly price." price of x all members would like to sell more x than is being sold at point Q. so that in the end both parties end up worse off. _However. Those who continue in work gain. even if there are no coalitions. So our preceding analysis covers the Case of monopoly. Suppose for example that B consists of only one individual and A of many individuals. First they may encourage retaliation.

Output is produced by very many firms. But at this point average cost would be rising. . where we can again learn important lessons. t This point is discussed more fully in Chapter 7. Clearly the amount of corn that a labourer can buy is his money wage WL divided by the price of corno This is wJpy-the real wage. 'This assumption does not mean that the firms do not experience increasing returns over some initial range of output-it is hard to think of any real firm that does not. each owning 1 unit of L. (ii) stable? Q2-2 Suppose that in the above case consumers of type A could agree among themselves on a price at which they would sell x (but consumers of type B could not collude). Before tackling the two-sector case. if all potential firms are identical. "B = xByB In a competitive market what is an equilibrium relative price of x? Is this equilibrium (i) unique. l/A = xAyA. Clearly the welfare of both groups depends on their incomes and the prices of the things they live on. Similarly the real income of a capitalist will be WK/Py' In the one-sector model these prices are determined in the simplest possible way. this time without bringing in demando From now on we shall dispense with individuals A and B and imagine a world peopled by many identical workers. it might well expand production to a point at which it had decreasing returns to scale. Therefore. price must exceed average cost. we shall start with an even simpler model in order to develop some elementary ideas about the influence of relative factor supplies on the distribution of income. and many capitalists. each having identical production functions and constant returns to scale at its equilibrium level of output. Suppose there is only one good (corn. each owning 1 unit of K. and average cost all be come equal. so that marginal cost would exceed average cost. y). But the equality between the last two implies constant returns to scale. The question is: Would we expect to observe firms actually producing at such low levels of output? If returns to scale are increasing.This tempts new firms to set up in business and. So profits are made. and so we shall in Chapters 2 to 4 treat each firm as having constant returns to scale. ifmarginal cost also equals price (as profit-rnaximising behaviour requires). with price py. In fact.GENERAL EQUILIBRIUM 63 Q2-1 Suppose consumers of type A are endowed with the total supply of x and consumers of type B with the total supply of y.] We can safely disregard all features of the production function except those which describe it in the neighbourhood of equilibrium. marginal cost. Suppose the factors are supplied in fixed quantities (K. What price would they set? 2-2 PRODUCTION WITHOUT CONSUMPTION: ONE-SECTOR MODEL We turn now to the production side of the economy. L). average costs are falling. . and so are alternative assumptions. drives profits down to zero. So price. So any firm that is a price taker will expand its output (or not produce at all).

However. YK K L identical ~---------------.2). For anyfunction y = y(K.. which have the property that increasing all inputs by a multiple A increases output by a multiple AP. we raise output to y = 2.L. Thus the two isoquants must be parallel along any rayo If the function had possessed increasing returns to scale (p = 1.. L) that is homoqeneous ofdeqree 1 tlie absolute level oieach marginal product YL and YK depends only on the input proportions K/L.1.Iv« rises with K/L.LL. say with p = 1.2.2.<:. if we restrict ourselves to constant stronger result: returns to scale we obtain a . v.. The same is true 01 the average products y/K and y/L. If we double either of these input vectors. K y =2 y per K (f) !'!:.. If the function has constant returns to scale..<. so doubling all inputs doubles output. the diagram would have looked exactly the same except that the outer isoquant would have been labelled (y ~2.L (a) l.1).L. (b) The relation . but this is less important.O--------- (b) (f) of YK and K/L...<-<:_. So our first property is as follows: 1.<:. doubling all inputs would change output by a multiple of just over 2. For any homoqeneous [unction y(K.64 WELFARE ECONOMICS ANO GENERAL EQUILIBRI M Homogeneous Functions To understand the implications ofthis for the distribution of income we must pause for a moment to set out the mathematical properties of such functions (they are pro ved in Appendix 3). it is homogeneous of degree 1 (p = 1).... Figure 2-9 (a) The relation of YJYK and K/L. So the y = 2 isoquant cuts all rays from the origin twice as far out as the y = 1 isoquant does. If there were increasing returns to scale. the input mixes at points pO and pl in Figure 2-9a both produce y = 1.<.. As Figure 2-9a shows.<.. Constant returns to scale functions belong to a wider class of homogeneous functions..LL. L) the marginal rate 01 substitution between inputs y¡}YK depends only 0/1 the input proportions K/L. Returning to constant returns to scale. where p is the degree of homogeneity.

Thus. If. ~ \ However. So. and only if. there are constant returns to scale. they will be paid their marginal physical product times the marginal revenue per unit of output (which is less than the product price). nstead. under increasing returns to scale. if you like.llnstead we rely on the reader to keep these in mind. with the entrepreneurial input treated as a variable. All the variables that affect welfare depend only on the factor proportions. I (Üur final property is an adding-up property known as Euler's theorem: 3. for example. L) is homogeneous of deqree 1. and only if. In other words. to repeat: h' < O )'>0 The first of these relationships is illustrated in Figure 2-9b. the [unction y = y(K. as Euler's theorem shows. Euler's theorem. marginal productivity payments will underexhaust the product and leave an entrepreneurial excess profit.. you must remember that the marginal product of labour is higher at P' than at pO and the marginal product of capital is lower. with decreasing returns. L > O So if each factor is paid an amount the factor payments are PyJIK equal to the value of its marginal product. and JIK falls with K/L. YL rises with K/L. points pO and p2 in Figure 2-9a.\This can be thought of. as a return to a factor in fixed supply. under increasing returns. then all K. a more general form of Euler's theorem states where p > 1 for increasing returns and p < 1 for decreasing returns. This makes it clear that the marginal product of capital must be the same at. factors will not be paid the value of their marginal product because the product market will be uncompetitive. Naturally. they would be paid more than the total in come of the firmalsut. the 'true " production function. K + PyYL L By Euler's theorem this exactly equals the value of the product Py y if.GENERAL EQUILlBRIUM 65 So constant returns to scale have the key virtue that scale does not matter. exhibiting constant returns to scale] The case of decreasing returns thus poses no difficulties for our competitive theory and our model los es little by overlooking such cases for the time being.í This logical deduction is sometimes called the law of variable proportions. So if some firms do produce with locally increasing returns to scale.Jln the later sections of this chapter we do not show the relationship between marginal products and factor proportions in any diagrarns. if factors of production were paid the value of their marginal product. If returns to scale are decreasing or increasing but the function is homogeneous of degree p. this .

(K) - The solution for wKjpy is illustrated in Figure 2-9b. L) space.L)=} . when specifying the production function we had chosen to measure L in units such that the total labour supply L consisted of 1 unit of labour. Both are measured in units of product.e. So what can we say about real factor incomes? In perfect competition a firm will hire each factor until its real wage (in units of product) equals its real marginal product. L) = h -= v. o. This whole analysis was developed by J. B.= YK(K. so each will choose the same factor proportions Kijt (due to homogeneity). Given this and constant returns we can clearly describe the output of the y industry by an aggregate production function: y = Y(t t t) = x'. the total shaded area in the figure. Assuming that both K and L are fully employed. it follows that WK .] In order to see the relative shares of labour and capital in the total product.~f we had not fiddled with the units of measurement. So the top triangle must be the absolute real income of labour. Income Distribution Let us see where this leads uso We have reached the point where each firm i is producing with the same constant returns to scale production function i = y(Ki.66 WELFARE ECONOMICS A O GENERAL ECONOMICS is contrary to the spirit of our analysis. in particular. but the ratios of the areas would still be a valid measure of the relative shares. The total national product is the area under the marginal product curve for capital YK integrated between K = O and K = K. the areas would not correspond to the absolute shares. So this is the one-sector theory of distribution. K. t We express some reservations this model in Chapter . i. in other words. Thus the horizontal axis of Figure 2-9b measures Kjl. L) and use this also to compute the marginal products of each factor. This will be our standard practice throughout this part of the book.~ Each firm faces the same factor prices. which can then only be justified on the grounds that some simplification is needed to capture the more important aspects of reality-and.. are well-defined at all points in (K. to do policy-relevant comparative statics. let us imagine that. (K) L L and WL -=YL(K. t). But the bottom rectangle is the absolute real income of capital YK K. t We shall assume it is twice-differentiable about so that marginal products 12. Clark at the end of the last century. y(K.

In Chapter 9 we use two types of labour and point out that labour complementary to immigrant labour will gain from immigration. The condition for this turns out to. We shall discuss the elasticity o. Thus even if YL L/YK K. for example. the total income of labour. The ratio of these proportions is called the elasticity of substitution. What happens to.rces are at work. 2. in fact. what happens to (ydYK)(L/ K)? Clearly as the ratio of labour to capital rises. = 1. if this exceeds unity. rises.t (To check this. i. YK K.e. portray the same inforrnation with L/ K as the dependent variable (see Figure 2-10).. relative shares change. Why. We can equally. s. How will real factor incornes change? We have already seen that YL fa lis and YK rises.f increasing one factor (say L).: This diagram provides a whole range of immediate insights. three main questions that arise in considering the distributional effects o. be that s > YK K. but clearly the less curved the isoquant. the other factor being held constant (at K. t This holds if (i) there are constant returns to scale and (ii) we only distinguish between two factors ofproduction. falls.[\ There are. . is the real wage YL and the standard of living y/L lower in India than Europe? Because K/L is lower. YL L? Since YK rises.). of course. how would imrnigration of labour affect the welfare of capitalists and native workers in a country? Capitalists would gain and resident workers lose. How do. capitalists gaining more (areas a + b) than native workers lost (area b). but yJYK falls (see Figure 2-9a). Or. Whether the relative labour share rises depends on whether the proportional increase in L/K exceeds the proportional fall in yJYK.GENERAL EQUILIBRIUM 67 y per L YL '--- L--_---'- K L Figure 2-10 The relation of YL and L/K. the higher the elasticity of substitution.f the increased factor rises. The marginal product of labour falls as L/ K rises./y (see Chapter 9). YL L may still rise. the share o.f substitution more fully in Chapter 9. two opposite fo. 3. L/K rises. 1. assume K = K. So.

68 WELFARE ECONOMICS AND GE ERAL EQUILIBRIUM Q2-3 If Y = 100 K'/2I!/2 . Px WK) Py If both wd». how does capital accumulation K to K') affect (i) The real wage and real capital rental (ii) The relative shares of national income (iii) The absolute share of capital (Since we have not yet given a formula for s. the effect on capitalists' welfare depends on their utility functions. and L is man-years (millions). and WK/Py rise. USA) (say. whatever the set of prices is. we can writfj uK = f(WK. But how are these prices determined in competitive equilibrium? In a closed economy we cannot answer the question without bringing in demand (i.) The welfare of factor owners now depends on the price of both products relative to their own factor incomes.. and wJpy specifies the maximum amount of y. India) Country 1 2 3 Q2-4 If Y = K'/4¡J/4 and the labour force is constant at L. Similarly for workers. meaning by this that at any given set of factor prices.000 L 100 20 200 (say. where y is output per year (millions of dollars).000. we can do two things. However. solve the question algebraically by computing answers. But if WK/Py rises and WK/Px falls. UK) (say. preferences as between x and y). The interest ofthis arises because one sector may be more capital-intensive than another and the fortunes of factor owners become bound up with the fortunes of different industriesWe shall assume throughout that x is more labourintensive than y. So these two numbers determine the location of his budget line and thus the maximum utility he can get. if both fall. Thus for example WK/Px determines the maximum amount of x an owner of one unit of capital can buy if he spends his whole income on x. x will have a lower K/L ratio than y.e. (x might be clothing and y cars. It also provides us .000 200. what is the real wage and output per worker in the following countries: K 9. This is useful as a building block in the general equilibrium of a closed economy.000 5. K is the capital stock (millions of dollars). If we henceforth use uK to refer to the maximum utility available to a capitalist.) (from actual 2-3 PRODUCTION WITHOUT CONSUMPTION: TWO-SECTOR MODEL We can now move on to a more realistic model with two productive sectors x and y. they are worse off. First we can establish a remarkable one-to-one relationship between the relative prices of products and the welfare of workers and capitalists. and. capitalists are unambiguously better off.

as we have stated it. "International Factor Price Equalisation Once Again. ' The Relation of Relative Product Prices to Real Factor Prices Our first proposition is this. 1941 [also in J Bhagwati (ed.). pp. 1 I The story is ilIustratedin the "Lerner-Pearce " diagram in Figure 2-11. At the same time. provided that some of each . "Protection . and the second two rise. the original price of each is measured by the distance OP. Suppose the relative price of labour rises. and P. Penguin. vol. pp. 181-197. We need to establish a one-to-one correspondence between changes in factor prices and in product prices. and vice versa. London. lnternational t See W. the capital-Iabour ratio in each industry is increased in response to the now higher relative price of labour. But -=XK WK px The first two of these fall.GE ERAL EQUILlBRIUM 69 with a complete theory of income distribution for an open economy in which alI product prices are determined by worId trade. Stolper and P. the price of each good equals the cost of producing 1 unit of it.t The proof is simple.good is being produced. In any particular country a rise in the relatioe (producer) price of the . for labour costs are a higher fraction of the costs of x. Stolper-Samuelson theorem. 1969. A.' Economic Joumal. and Real Wages. 1. Studies. 2. 245-267]. is only necessarily true if prices change exogenously (and not due to changes in taxes). F. So the marginal product of labour rises and the marginal product of capital falIs. For convenience we choose units of x and y such that at the initial relative price of the factors (WdWK).labour-intensine good will make labour better off and capital worse off. A. Since x employs a higher ratio of labour to capital than y. Since there is constant retl!J. So the utility of capitalists unambiguously falIs and that of workers rises. we shalI first think of the causation as running from factor prices to product prices. Px = Py. the price of x must rise relative to that of y. Measured in units of K. This is indicated by the fact that the initial cost line PQ is the same for 1 unit of x and for 1 unit of y. 58-73.?s to scale. Samuelson." Review of Economic Trade. The proposition. 59. pp. Secondly. Samuelson. 1949. vol. Two consequences folIow. we can confirm that in a closed economy the welfare effects of changes in factor supply are the same as in the one-sector model. 9. For convenience.

both prices have fallen but Px has fallen less than Py. prices You can check this is by noting that the cost per unit of x divided by the price of capital is Now we let WdWK rise and trace out our two consequences: in units of K both rise. In the right-hand panel we observe a rise in the relative price of x. 2. 1.70 WELFARE ECONOMICS ANO GENERAL EQUILIBRIUM pY p o '---------"'-------"------=-----+Q L'. . We now raise the relative price of labour. This raises the K/ L ratio in both industries.t Px has risen more than Py(In terms of L. making capitalists worse off and workers better off (assuming full employment of all factors of production). At the same time we have the change in factor proportions resulting from the new relative factor prices. In the left-hand panel we observe the change in factor proportions 'In the t This is beca use of the increase in the price of labour in terms of K. Now 1. J So there is a one-to-one relationship between relative product prices and the welfare of the owners of the two factors. 2. We start with a low WdWK . e Figure 2-11 The effect of factor on product prices.) The basic reason for this is thata revaluation of the original factor inputs for y (at S) and x (at R) raises the cost of the inputs to x more than of those to y. The prices of x and y measured Though both prices have risen in terms of K. This can be usefully illustrated in Figure 2-12.

f3(K)'jB) f4(Kx/E) For given vJ». This is consistent with prices (pxfpy)o. Suppose we start with low (. This situation is illustrated in the left-hand panel of Figure 2-13.dWK)O. for the experiments of Figure 2-11 proved that in such cases wd»« is monotonically related to both vJ». But each marginal product is a function of the relevant K/L ratio. ¡l The functional relationship between relative product price and all other variables follows simply from the requirement that the relative price pxfpy must equal the relative marginal costs of x and y. However. which brings about the welfare changes we have described. one caveat. but that. its capital-labour ratio is always higher. we know the whole structure of real marginal products and relative factor prices. whether these are expressed in terms of capital or labour. We have assumed that there is only one set of (K/Lyand (K/L)Y consistent with a given set of relative product prices. As before. so its relative price begi~s to fal!. KX/E and KY/D'. Having solved for these. however. this gives us two equations in the two unknowns. we obtain from the two preceding equations Px Pr fl(Kr/B) P(Kx/E) Px o. So p.:~ Figure 2-12 The product simple version.GENERAL EQUILlBRIUM 71 x -: (~r(U ----~a'---------'------'-o--------+price/Iact or price relation: . and to the K/L ratios. the relative price of x rises. We now raise the relative price of labour. a time comes when x is no longer the labour-intensive good. YK XK Px YL XL Py v. suppose that at a low relative price of labour y is indeed the capital-intensive good. If y is unambiguously more capital-intensive than x. X becomes capital-intensive. There is. Since the y industry is unambiguously more capital-intensive than the x industry. Writing these functions in a general notation. no matter what the relative factor price. n both of them the capital-labour ratio rises. But if we continue to raise the price of labour. at higher values of WdWK. This .. two industries. this must be so.

from the identity K KXE r=EL+IJL KYIJ . But. How serious a problem is all this? It matters considerably if we wish to compare countries engaged in international trade. The Lerner-Pearce diagram in Figure 2-14 tells the same story. We now raise the relative price discontinuously to (WdWK)l. capital-labour ratios and welfare levels of workers and capitalists. and x now has a higher K/L than y. The reason is of course that the capital-labour ratio responds more to relative factor prices in the x industry than it does in the y industry. but a given relative factor price implies a unique relative product price. But in the meantime the factor intensities have been reversed. The reason is simple. their overall ratio must equal a weighted average of the ratio in the two industries.= 1. it causes no problem. So a given product price is consistent with two sets of relative factor prices. eventually fa lis to its original value and beyond. not that Px/Py has risen. In the case of "[actor-intensit y" reversals. In other words.72 WELFARE ECONOMICS ANO GENERAL EQUILIBRIUM y IV )1 ( IV: o Figure 2-13 A factor-intensity reversaL Py is shown in the right-hand panel. In the economy as a whole the overall capital-labour ratio is given (at K/L). iffactors are fully employed. vJ». a given relative product price is consistent with more than one relative factor price. At this relative price we find. but that it is again unity. As the relative price of labour continues to rise. At these prices (Px/Py)o. The equilibrium capitallabour ratio is lower in x than in y. We start with relative factor prices (WdWK)O. the elasticity of substitution is higher. But if we take any particular country with given factor supplies.

either both of them exceed the number ex or both are less than ex. Instead consider the familiar identity: ~ = (Iow) (~r~+ (~rf1 i (high] (1) . say from pO to r'. if (KjL) < ex. the latter holds and x is the labourintensive good. So. fram Figure 2-13. y is the capital-íntensivegood. However. It also enables us to prove that the transformation curve is convex-outwards. Equally. U O Figure 2-14 A factor-intensity reversal. But how do relative factor prices alter? For the time being we must not rely on any of the geometrical praperties of the drawing. KV (_KL)Xl I Slope] =(W L)O IVK ~=------------------ ¿X. it does not bring out how the pattern of output is changing as prices change. or if x is a closed economy and peoples' tastes change in favour of x. or vice versa. So if (KjL) is greater than ex. it will benefit workers and hurt capitalists. The Edgeworth box provides the best way of showing this. This will be so if x is a trading nation and the world price of x rises. Fram now on we shall assume that. At pl more x is praduced than at pO and less y. The Relation of the Output Mix to Real Factor Prices The preceding analysis pro ves the Stolper-Samuelson theorem.GENERAL EQUILlBRIUM 73 KX. It is impossible for both (KjLy and (KjL)Y to exceed ex and for their weighted average to be less than ex. Figure 2-15 repeats the Edgeworth box in input space. so are (KjLyand (KjLV And in this case x is the capital-intensive good. in our particular country. It follows that if the relative price of x rises. But in any particular economy (KjLy and (KjL)Y are determined by the same factor prices. Let us consider a move between two points on the efficiency locus.

So any discrete increase in x will involve an increase in (K/L)"' and (K/L)". So it follows that. (K/ LY and (K/L)Y will move in the same direction.t But (K/LY < (K/L)Y. a shortage of labour must emerge. Before leaving Figure 2-15. Which of these will happen? Since both industries face the same factor prices.) And since both capital-labour ratios rise. as x rises and rises. as x rises and y falls. (This is how they have been drawn in Figure 2-15. the relative price of the factor in which x is intensive This accords with elementary common sense. . the relative price of labour must rise. unless. But we also know that as WdWK rises.e. If more x is produced at pl than pO. Since (K/LY rises as E rises. Px/Py rises: The transformation curve is conoex-outwards. which implies an increase in C and K' (see Figure 2-15). L' will be higher and I! will be lower. So a higher weight is now being attached to the lower K/L. i. If the labour-intensive output expands. By rewriting (1) with K and L rever sed one can check that the argument which follows in the text applies equally well if KX is raised. either (K/LY or (K/L)Y or both rise. to maintain full employment of K and L. If x were tAn alert reader may question this and suggest that the extra x could be produced entirely by raising KX. we should note some properties of the efficiency locus.. rises. vJ». the curve is convex-downwards.Figure 2-15 The relation of capital intensities and output mix. followed by a general reduction in the labour intensity of production and a rise in the scarcity price of labour. So. they will both rise. This will reduce the right-hand side of the equation. y fal/s.

By the properties of constant returns to scale. The Two-Sector Model of General Equilibrium. Thus if the efficiency locus touches the diagonal at all. London. Figure 2-16 provides the relevan t pair of diagrams in factor space and output space. Suppose there were a point on the diagonal which belonged to the efficiency locus. of course. A geometrical proof is necessarily laborious. So if the rates in the two industries are equal at some point on the diagonal. In the top diagram the output mix at OY is thus (xmax. More and more x is produced.GENERAL EQUILIBRIUM 75 capital-intensive. and reapproach the question about immigration which we originally asked in the context of the one-sector model. hmax). This is consistent with the proposition we have already proved.] Finally let us return to Figure 2-12. But on the efficiency locus. . and at the prevailing factor prices the two industries are both equally capital-intensive. Suppose the world price of x rises continuously. Eventually all the economy's resources go into producing X. pp. that the transformation curve is in fact convex- outwards. O) and at OX it is (O. if x = 1-Xmax. But it could never have points of inñection. If. So the transformation curve lies outside a straight line in output space joining xmax and ymax. the rates of substitution must in any one industry be the same at all points on the diagonal. This breaks the general Stolper-Samuelson relation and is the reason why their proposition is qualified by the remark that "some of each good must be produced. But since (by constant returns to scale) the rate of substitution in any industry depends only on its KjL. is (txmax. Nothing can alter the marginal products of labour and capital in terms of X. 1971. output at point P. 25-26. the transformation curve is straight. the maximum output of y. See H." Effect of Changes in Factor Supply in a Closed Econorny Finally we can revert to the cIosed economy. SO. they must be equal at al! points on the diagonal. The maximum output of x is xmax and of y is ymax. however. while pxfpy rises. In this case. it must lie entirely along it. Then at that K] L the marginal rates of substitution between capital and labour must be the same in both industries. the curve would of course be convex-upwards. G. Johnson. ymax). What is the effect on income distribution? t Our geometry only illustrates that proposition. the transformation curve cannot be straigh t. exceeds hmax. the industries differ in capital intensity. Suppose that K remains at K but L rises from L to [' and that this is not enough to produce a factor-intensity reversal. Still less could it cross the diagonal. The relative price of labour rises. (KjLy = a. Allen and Unwin. half-way along the diagonal. Suppose now that the world price of x rises still further. The reason for this is obvious. Suppose the overall endowrrient of the economy is Kjr = o: When only x is being produced. WJWK remains constant. since any move of a given distance up the efficiency locus increases x by a constant amount and de creases y by a constant amount (due to constant returns to scale).

so output of y must occur along a line through O>" having the slope of OYP. factor proportions in the y industry have to remain constant. We can represent this by a shift of O" to 0>". -. infer the required direction of change of prices. O" held put. -. -. -. "-.e.. '----------. holding prices constant.--'------------"''----------------·x ~x m3x -.'XP . . We can now follow a standard procedure which we shall use time and time again in comparing static equilibria (i.v = ~ymax y y = ymax ymax -. "-. For full . Infer the pattern of supply and demand that would follow. -. Now the supply of labour rises from L to D.:rmax Figure 2-16 Why the transforrnation curve is not straight ir capital intensities differ. -. Since factor prices are constant. -. However. in what is called comparative statics). First assume all prices unchanged. -. If these patterns are inconsistent. we first examine the effect of the immigration on the supply of x and y. .76 WELFARE ECONOMICS AND GENERAL EQUILIBRIUM ----------------L--------------_ . -. -. -. In Figure 2-17 (KjLy is represented by the slope of OXP and (KjL)Y by the slope of OYP. So. -. factor proportions in each industry are unchanged.

and then (ii): (i) You do not know workers' utility functions. U factor and product prices are constant and there are now more owners of some factors. This is the same conclusion as we reached with the one-sector model. Therefore. As regards the relative and absolute shares of the two factors. must fall and this will reduce WJWK' A fall in this reduces the welfare of workers and improves that of capitalists. Q2-S How would you rank the welfare of workers in the following states of the world: Px State l State 2 State 3 l l 1 3 4 2 2 3 1 Assume first (i) below. Now let us examine the demand for y. (ii) u = xy. we should now need to take into account not only the elasticity ofsubstitution in both industries but also the elasticity of substitution in consumption. production must occur at P'. Q2-6 Are workers rational to lobby for tariffs on labour-intensive imports? . another topic treated in later chapters. If product prices are held constant. of y fallen.GENERAL EQUILlBRIUM 77 ox_---- --------L'--------- z----_ Figure 2-17 The Rybczynski theorem. as we saw. oJv. less of y will be supplied. So the output of X has risen and that Rybczynski theorem. more of each good will be demanded. a The striking point of the theorem is that y falls. employment. to clear the market. an increase in the supply offactor L will lead to an increase in outp •. But.• of the L-intensive good and t fall in output of the other good.

We first take two extreme cases. Provided that. as we vary rJ». But some y will be demanded.78 WELFARE ECONOMICS AND GENERAL E(2UILIBRIUM Q2-7 Suppose x = K El'. except that output is no longer given. bear in mind that the relative price r J». so there will be excess demand for y and (by Walras' Law) excess supply of X. we shall focus on the change in demand for x as' pjpy varies.. no y is produced. but with y = 2K. t The reader who wishes to check that. The moment oftruth can be reached quite quickly. no x will be produced. from ver y high to very low. Walras' Law guarantees that if there is zero excess demand for x. At the other extreme if the price is high enough. So. So we can concentrate on looking at the state of the market for x as the price vJ». excess demand for x will fall so long as demand does not in crease.1 relative prices can find satisfaction in Appendix 5. Q2-9 Suppose that with production functions as in Question 2-8 we evaluate x and y such that How does the welfare of workers and capitalists compare with that found in Question 2-8? 2-4 PRODUCTION AND CONSUMPTION We are now ready to bring together consumption and production in a more general way.) (ii) Evaluate the following at points (a) and (b) above: Px P. there must be some price for which excess demand is zero. As before. and is this equilibrium unique? The first question can be answered by following the same general approach as in Seco 2-1. Since a rise in »J». There will be a unique equilibrium provided that at equilibrium a rise in pjpy reduces excess demand for X. always increases the supply of x. only y. as before.'2' El' and everything else as before. At which point is labour better off? Q2-8 Same as Question 2-7. So there will be excess demand for X. Suppose we start with a very low relative price of x. in the n-commodity case. there is zero excess demand for y (assuming factor markets clear). as in Seco 2-1. This is a point of equilibrium. (i) What are the values of x and y on the transformation curve corresponding to first (a) below and then (b): (a) K = KY x2 ' X (b) E = IJ (Do not evaluate cube roots further than is needed to see what is happening. But some x will be demanded. . determines both the pattern of output supplied and the pattern of output demanded.t But will there be multiple equilibria? As we have seen. Once again we are asking: Is there an equilibrium set of prices. there are enough equations to solve for the n . only X. this is the same question as: is equilibrium stable? To examine this. y = xr: IJ' " and the economy is endowed with K and [ measured in units such that K = L = l. varies. excess demand varies in a continuous manner.

must be the line AA. What happens to the demand for x as a result of this rise in its price? 1. y*). as we have seen. so that the total outputs (x*. The income effect of L leads to an increased demando . If we are in equilibrium. where both are consuming nonnegative quantities of each good the lines coincide. WK/Px and WK/Py fall. So far all is simple. . Since the payments to factors exhaust the total product. as we know. the potential welfare of capitalists goes down. But 3. assuming x is a normal good. Suppose we have an equilibrium with outputs (x*. But these in turn.But we are interested in the separate budget lines of K and L. 2.GENERAL EQUILIBRIUM 79 y A OK~'~------~X-*--------~--------~B~K--~X Figure 2-18 General equilibrium How can we picture the demand for x? Clearly it depends on the opportunities open to capitalists (defined by WK/Px and WK/Py) and to workers (defined by wdpx and wdpy). y*) in Figure 2-18. labour and capital taken together must between thern be able to buy the whole product. Suppose we measure K's budget line from OK. yl )-more of labourintensive x and less of y (see Figure 2-19). Over the range. are determined by the pattern of output. the amount of x demanded by K (shown at e!5) plus the amount demanded by L (shown at eL) must add up to the total output ofx. The income effect of K also leads to a reduced demand. with OK taken as origin. Then some line must be able to serve as the common budget line for K (referred to OK) and for L (referred to OL). So some geometry is helpful. In the diagram BK BK is K's budget line and BL BL is Es budget line. Now what happens if vJ». In other words C': and C': must occur at the same point. rises? Output shifts to (x'. Happily this is so in Figure 2-18. Similarly the potential welfare of workers rises: their new line is further from the new OL at both its ends. y*) are demanded. The substitution effects of both K and L lead to a reduced demand for x. So. so their budget line moves in towards OK at both its ends. So the budget line for their combined incornes. but measure L's budget line inwards from (x*.

Production would shift to (Xl. This excess supply of x is what the country is willing to export. Moreover.80 y WELFARE ECONOMICS ANO GENERAL EQUILIBRIUM O KL-------~ __-----I~~----X I xl I I \ \ \ I I I I I \1 'J BL Figure trade. But on normal assumptions the demand for x falls. in return it needs to import y of equivalent value in order to satisfy the excess domestic demand for y indicated by the "trading triangle" shown in the figure. where the sum of K's demands for x (at eK) and Es demands (at eL) do not add up to total output. But once again it is highly likely. no trade would occur. 2-19 Disequilibrium. That is what we shall assume in Chapter 4. i) with higher output of x. we cannot say unambiguously that the demand for x falls. So we end this chapter on the same issue as we began. if opened to world trade. even if the demand for x rises. y*). But we can also use the same apparatus to see how the economy will respond. it is unlikely to rise as much as supply. unless there is foreign So. as in the case of pure exchange. and much less likely in the case of production and consumption than in the case of pure exchange. This is illustrated in Figure 2-19. if we had a world price lower than at (x*. If the world price equalled the equilibrium price at (x*. y*) we should' import x and export y. since a rise in price now induces an increase in supply that is likely to exceed any possible increase in demando This completes our analysis of the equilibrium of a closed economy. We have found that in the two-good case there will exist an equilibrium set of prices that will clear all . assuming the country is a price taker. But suppose the world price equals the higher slope of. Thus multiple equilibria are unlikely. so that the import demand for x increases continually as its relative price falls. So there is excess supply of x. Equally. AA in Figure 2-19. In addition we shall make our normal assumptions about excess demand.

Will such a wage necessarily make workers who cannot get jobs in the x industry worse off? (The x industry may be capital-intensive or labour-intensive. and what is the structure of prices (i) In a rich country with K = 1. L =1 (iii) In a poor country with K = 0.. and is above the equilibrium leve!.GENERAL EQUILIBRIUM 81 markets. For this reason in the next chapter we use the results obtained so far to throw light on the effect of public policy upon the distribution of income. (i) if 1/ = xK"y" (ii) ir uK = xKyK34 u': = xL"yL uL = xLyLll4 Q2-11 To produce 1 unit of x requires 1 unit of L and 2 units of K. The two-good approach has also helped us to see what forces determine real factor incomes. Powerful routines now exist for doing this. Cowles Foundation Press. the higher the capital-labour ratio of the economy. The minimum is expressed in terms of «J». Will there be full employment of labour. the higher the real income of workers. Suppose u = xy.4. In a closed economy. Q2-10 Suppose x =K +E )' = 2K" + IJ X K=[. Equilibrium will be stable and is likely. General equilibrium theory itself is rather dry-it exists for the sake of its applications. to be unique. t: Q2-12 Suppose a minimum wage is imposed in one industry (x). Monograph The Computar ion of Economic Equílibria. the stronger is the public's preference for labour-intensive goods. In an open economy. L = 1.=l What in general equilibrium are P. To produce 1 unit of y requires 1 unit of L and 1 unit of K.5. Thus fairly clear-cut remarks can be made about comparative statics in a two-good. but not certain. there are few general propositions available.j But the two-good case is still important for developing one's intuitions about what kind of results to expect. = 1 (ii) In a less rich country with K = 1. as it often is.8. workers gain the higher the relative world price of labour-intensive goods. Scarf. Workers also gain. two-factor world. Instead one must compute the general equilibrium values corresponding to each set of data. New Haven. Yale University . If it is necessary. for Research in Economics. the wage in y being uncontrolled. 24. to take into account more than two goods or factors.) t H. 1973.

Master your semester with Scribd & The New York Times

Special offer for students: Only $4.99/month.

Master your semester with Scribd & The New York Times

Cancel anytime.