The Economics of Welfare

Pigou, Arthur C.
Display paragraphs in this book containing:
First Pub. Date
London: Macmillan and Co.
Pub. Date
4th edition.
43 of 73

Part III, Chapter VI


§ 1. WHAT has been said in the preceding chapters will serve well enough for the rough purposes of practice. But it is interesting to the economist to probe somewhat deeper, and to set out some of the broader issues that have been raised in a different guise. From an analytical point of view differences between a workmen's union and an employers' association are analogous to differences between two nations. When the differences concern such things as the general conditions of work, the methods of wage payment, the hours of labour and the demarcation of work, the analogy is close. But, when they concern wages, there is an important distinction. Issues between governments usually do not, but issues about wages always do, refer to a rate of interchange between two things, the quantities to be exchanged being left unspecified.*54 In negotiations of this character the one party does not always benefit himself by forcing the other to accept more and more unpalatable terms. On the contrary, after a point, any further raising of the rate against an opponent will bring about such a reduction in the quantity of the thing which he is prepared to buy that, to put it paradoxically, the victor is made actually worse off by getting better terms. Thus, when the bargain is about a rate, there will be certain upper and lower limits outside of which it will not pay either party to go, whereas, when the bargain is about a thing, there are no such limits.


§ 2. When, both on the side of workpeople and on that of employers, competition works with perfect freedom, the result of a bargain about the rate of wages is determinate at a single definite rate, which is settled by the conditions of reciprocal demand. If any worker asks for more than this from any employer, that employer will refuse to engage him and will take somebody else, and, if any employer offers any less than this to any workman, that workman will refuse to work for him and will go to somebody else. Where, however, wage rates are settled, not by the action of free competition, but by bargaining between a workmen's association on one side and an employers' association on the other, the rate of wage is no longer determinate at a single point. There is, on the contrary, a range of indeterminateness. The workmen's association will prefer to the competitive rate something higher than that rate, and the employers' association something lower. In view of the fact that a rise in the rate will lessen the amount of employment obtainable, there will be a certain maximum rate above which the workpeople's association will not wish to go, and, in view of the fact that a reduction in the rate will lessen the amount of labour obtainable, there will be a certain minimum rate below which the employers' association will not wish to go. The range of indeterminateness is constituted by all the rates enclosed within these two limits. Let us suppose that they are 40s. and 30s. respectively. From any rate lying outside this range it is to the interest of both sides to depart in the same direction. It is, therefore, impossible for a settlement to be reached with any wage rate outside that range. If there is any settlement it must fall at some point within it. The extent of the range will be larger the less elastic is the demand for labour by the employers, and the less elastic is the demand for jobs with these employers by the workpeople.


§ 3. In considering their policy, the workpeople's association will reflect that, if they elect to fight a battle about wages, the fight will cost them so much and the terms obtained at the end of the fight are likely to be such and such. Weighing up these things, they will determine on a certain minimum wage which it is worth while, if necessary, to accept rather than fight. This will be, as it were, their sticking point. If they think that a fight would cost them a great deal and that they would get bad terms at the end, this sticking point will be low. It may be a good deal lower than the 30s., which we have supposed to be the lower limit of the range of indeterminateness. On the other hand, if they think that a fight would cost them very little (or, still more, if they think that the actual process of it would benefit them), and that they would get good terms at the end, their sticking point will be high. It may be as high as, but cannot be higher than, the 40s., which is the upper end of the range of indeterminateness. Thus the workpeople's sticking point may, according to circumstances, be any wage less than 40s. By similar reasoning it can be shown that the employers' sticking point may, according to circumstances, be any wage greater than 30s. When the workpeople's sticking point is lower than the employers' sticking point, the range of wages between the two sticking points constitutes, subject to a qualification to be introduced presently, the range of practicable bargains. Thus, if the workpeople would take 32s. rather than fight and the employers would pay 37s. rather than fight, this range is made up of all rates between 32s. and 37s. When, however, the workpeople's sticking point is above the employers' sticking point, when, for example, the workpeople would go as low as 35s., but no lower, rather than fight, and the employers would go as high as 33s., but no higher, rather than fight, there is no range of practicable bargains; and it is impossible for the issue between the two sides to be settled without a fight.


§ 4. If both sides have the same expectations as to the way in which a fight would end and as to the wage rate that would be established in consequence of it, and if each side believes that the process of the fight would involve some positive cost to it, the workpeople's sticking point must be below the employers' sticking point, and there must, therefore, be some range of practicable bargains. If, however, the workpeople expect a fight to yield a higher wage rate than the employers do, this is not necessary, even though each side expects the fight to involve some positive cost: and, if one side expects the fight to be actually of benefit to it,—to involve, as it were, negative cost—it is not necessary, even though workpeople and employers both anticipate the same result to the fight.


§ 5. It should be added in this connection that negative cost is by no means, as might perhaps be thought at first sight, a mere mathematical figment. On the side of the employers it may easily be looked for if their commodity is one the demand for which is highly inelastic, and if, at the time of the conflict, they have large accumulated stocks in hand. Thus it is sometimes alleged that coal-owners are enabled by a conflict "to clear their stocks at famine prices, while postponing the fulfilment of their contracts under strike clauses."*55 They may also anticipate negative costs if they have reason to believe that, by precipitating a conflict in times of bad trade, they can insure themselves against being hampered by one when trade improves.*56 On the side of the workpeople, negative cost is, in the early stages of industrial organisation, fairly common. For in that period what the men are really aiming at is, not the concession of a higher wage, but respect for their Trade Union and consequent increased readiness to deal fairly with them in the future. Or again, the real purpose of the conflict may be to solidify the trade union itself and to attract non-unionists to its ranks. So far as conflicts are undertaken for objects of this kind, the advantage anticipated from the acquisition of those objects needs to be subtracted from the anticipated material losses of the dispute, and, when so subtracted, may leave a net negative result. In these circumstances the men may elect to fight, even though they expect both to be beaten and to be subjected to great suffering in the process. Negative costs are, however, exceptional. As a general rule both sides expect that the actual conduct of a conflict will involve them in loss.


§ 6. The qualification forecasted in § 3 has now to be introduced. Since the workpeople's sticking point may be anything below the highest point in the range of indeterminateness, it may be below the lowest point in that range. In like manner, the employers' sticking point may be above the highest point in that range. Thus, with a range of indeterminateness between 30s. or 40s., we might have the workpeople's sticking point at 25s. and the employers' sticking point at 48s. In these circumstances the range of practicable bargains will not extend to the two sticking points: for no bargain is practicable that lies outside the range of indeterminateness. Though one side may be willing to concede such a wage rather than fight, the other side will not be willing to accept it. Consequently, any wage rate above the workpeople's sticking point and any wage rate below the employers' sticking point, which lies outside the range of indeterminateness, is wholly ineffective. The range of practicable bargains, that is the series of wage rates which both parties would prefer to a fight, is constituted by the wage rates (if any) which both lie above the workpeople's sticking point and below the employers' sticking point and also fall within the range of indeterminateness. This explanation is essential to the practical inferences to which we now pass.


§ 7. The extent of the range of indeterminateness is determined, as was pointed out in § 2, by the elasticities of the reciprocal demands of the two parties. From the present point of view, therefore, it may be taken as fixed. The range of practicable bargains cannot, in any circumstances, be extended beyond the limits so determined. Within these limits it is increased by anything that moves the workpeople's sticking point downwards or the employers' sticking point upwards. The workpeople's sticking point is moved down by anything that increases the cost which they look to suffer in a conflict; and the employers' sticking point is moved up by anything that increases the cost which they look to suffer in it. An increase in the strength of the workpeople's association alone will probably both lower the anticipated cost to them and increase it to employers. Hence it will probably raise both sticking points; but it is impossible to say whether it will widen the gap between them. An increase in the strength of the employers' association alone will, in like manner, lower both sticking points. In practice, since the development of one rival association is almost certain to lead to the development of the other also, the most likely form of increase of strength is an increase on both sides, leading to an increase in the cost anticipated by both. This will mean a lowering of the workpeople's sticking point accompanied by a rise of the employers'. Within the limits set by the range of indeterminateness it will, therefore, extend the range of practicable bargains in both directions, provided that such range already exists, and it may bring such a range into being if none exists. This tendency is exhibited in international, as well as in industrial, negotiations. If the adjustment of some political difference by war means adjustment by a world war embracing powerful alliances on both sides, the enormous probable cost of a fight makes it certain that, unless the stake at issue is, or is thought to be, immensely important, some range of practicable bargains will be available. In like manner, when the organisations of workpeople and employers in some industry are extended to embrace the whole country instead of being merely local, so that a strike or lock-out, if it occurred, would be national in its scope, an issue affecting, say, the wages payable in some small district, for which, before, it may be, no range of practicable bargains was available, is likely no longer to prove intractable. Broadly speaking, as things are to-day, with the great strength which nations and industrial associations alike possess, we may safely conclude that, for nearly all minor matters and interpretation differences, a wide range of practicable bargains will be available. There will always be an enormous number of ways of settling, say, the political status of an obscure African village, or the exact application to a particular pit of a general wage agreement between coal-owners and coal-miners, any one of which would be very much more advantageous to both parties than a conflict could possibly be. When the matter in dispute is an important one, such as the possession of a large and rich territory or the determination of the general standard of wages in an industry, it is not, of course, equally certain that a range of practicable bargains will exist.


§ 8. When two nations, or two associations, representing employers and employed respectively, have entered into an agreement not to fight but to settle differences by arbitration, there is created a real cost, partly moral and partly material (through the probable loss of outside support), to either of them if it breaks the agreement. This is additional to the direct cost which conflict would involve whether there had been an arbitration agreement or not. Thus the setting up of such an agreement increases the cost of war to both parties, and so, within the limits set by the range of indeterminateness, tends to bring into being, and, if it is already in being, to widen, a range of practicable bargains. Workers, who, apart from the agreement, would have fought rather than accept 32s., may, in view of the agreement, keep at work though only 31s. is awarded; and employers, who would have locked out against a 35s. demand, may now go to 36s.


§ 9. When arbitration is voluntary, entrance into it means to each party, in the circumstances contemplated above, the risk of having to accept a rate less favourable to itself than the worst rate it would have accepted without fighting if it had not entered into arbitration. This consideration will not prevent the parties from entering into it if each thinks that the risk of loss, which it assumes by doing so, will be roughly balanced by a corresponding chance of gain. But it is not certain that there will be a balance of this kind. If, for instance, the lower end of the range of practicable bargains is already lying at the lower end of the range of indeterminateness, the introduction of an arbitration agreement cannot extend the range of practicable bargains in a sense favourable to employers, while it may extend it in a sense adverse to them. In these circumstances they will be disinclined for arbitration agreements. The same thing is true of workpeople in the converse case. When, however, conditions are such that entrance into an arbitration agreement leaves both ends of the range of practicable bargains within the range of indeterminateness, the chances of gain and loss resulting from it are likely to balance for both sides, and, consequently, both sides will often be willing to enter into it in spite of the danger of losing more by it than they would look to have lost by war.


§ 10. But this does not exhaust the matter. If it were feasible to construct a voluntary arbitration agreement in such a way as to preclude the arbitrator from making an award outside that range of practicable bargains, which is established by general conditions coupled with the fact of the agreement, there would, indeed, be nothing more to be said. But it is very hard to include a provision of this kind in any workable scheme. Even when arbitration is proposed in respect of a single existing dispute, the parties can hardly give their case away so far as to reveal to the arbitrator beforehand what their respective sticking points are, and bind him not to go outside them. When it is a question of a general arbitration treaty to cover future disputes not yet in being, the difficulty is still greater. If, however, no provision of this kind is included, each disputant will fear that the arbitrator may award terms so unfavourable to him that, in spite of the agreement, he will feel compelled to fight against them, and so make a net loss of his honour. For this danger the knowledge that his opponent may be placed in a similar position by an award of an opposite character is no compensation. Consequently, both sides will tend to confine arbitration agreements to types of dispute, in which, in their judgment, it is practically certain that arbitrators will not give awards falling beyond their respective sticking points. Thus, in international treaties, States have often reserved from arbitration "vital interests" and "questions of national honour"; and in several plans for the League of Nations a distinction was drawn between "justiciable" disputes, to be submitted to a Court whose awards the League would enforce, and non-justiciable disputes, which should go to a Council empowered only to make non-binding recommendations. Similarly in industrial affairs, while employers and workpeople are generally ready to arbitrate minor matters or interpretation differences, they often hesitate to deal in this way with the general question of wage rates.


§ 11. This unwillingness of the parties to enter voluntarily into far-reaching arbitration agreements has been an important factor in determining State authorities to intervene. Compulsory reference of disputes to a body entitled to recommend, but not to enforce, an award, as under the Canadian Industrial Disputes Investigation Act, tends, subject to the conditions which have been enumerated, to widen the range of practicable bargains, because a hostile public opinion adds to the cost of fighting against any settlement which has been recommended by an impartial public authority. It may, therefore, in some instances create a range where none would otherwise exist. The Canadian system is exactly parallel to the arrangement under which international disputes are to be submitted to the Council of the League before war about them is permitted by the League of Nations. Laws which provide, not only that disputes must be submitted to arbitration, but also that the arbitrator's award must be accepted on pain of legal penalties, since they make battle more costly and success in it less probable to a recalcitrant party, will act still more forcibly to create and to extend in both directions the range of practicable bargains.


§ 12. We have now to make clear the bearing of these results upon the prospects of industrial peace. It has already been shown that, when there is no range of practicable bargains, a settlement without conflict is impossible. But it would be wrong to infer that, when there is such a range, a peaceful settlement must occur. This is by no means so. The existence of a range of practicable bargains implies that there are a number of possible adjustments, whether it be of territory or of wages or of anything else, any one of which both disputants would prefer to accept rather than engage in conflict. Each of them, however, naturally wishes to obtain the best terms that he can, and is also, though this is not necessary to the analysis, almost certainly ignorant of how far the other would yield rather than fight. Thus, we may suppose that a Trade Union has decided to fight rather than to accept a wage below 30s., and the corresponding Employers' Association to fight rather than pay more than 35s. There is then a range of practicable bargains including all rates between 30s. and 35s. But, even though the workmen know that 35s. is the sticking point of the employers, and the employers know that 30s. is the sticking point of the workmen, there is still a conflict of will as to the precise point within the range, 30s. to 35s., at which the wage shall be put. Each side tries to push the other to his limit. The employers may think that, if they say firmly enough, "not a penny more than 31s.," the workmen will give way; the workmen may think that, if they insist unfalteringly on 34s., the employers will give way. In the result, both sides, equally the victims of unsuccessful bluff, may find themselves in a fight over an issue which both know is not worth it. In the circumstances the chance that conflict will be avoided is clearly greater, the more cordial are the general relations between workpeople and employers and the better is the machinery available for bringing about friendly negotiations. Thus, when a regular Conciliation Board is ready to hand or, failing that, when a mutually acceptable mediator is prepared to intervene, the prospect of a peaceful settlement is, so far, improved. When an arbitrator is available in the last resort, it is improved still further. For, when once an award is given that falls within the range of practicable bargains, it will not pay either side to fight rather than accept this award. One or other of them may, indeed, think it worth while to bluff in the hope of modifying the award by later negotiation. But, the award having, as it will have,—particularly when the arbitration has been entered upon voluntarily—moral force behind it, is not likely to be resisted up to the point of war, unless it lies beyond the sticking point of one party, that is to say, outside the range of practicable bargains. The wider, however, is the range of practicable bargains, the greater is the chance that an arbitrator will succeed in placing his award somewhere upon it. It follows that, when settlements are arbitrated, the chances of a peaceful adjustment are greater where the range of practicable bargains is wide than where it is narrow. This conclusion, however, does not follow when settlements are negotiated. On the contrary, since, the wider the range of practicable bargains, the greater are the opportunities for bluff by both sides, the opposite conclusion seems more easily defensible.


§ 13. From the general course of this discussion it is easily seen that the need for arbitrators does not, as is sometimes supposed, arise solely out of the ignorance of the parties to a dispute, and that their function is not "simply to find out what the price would naturally have tended to become."*57 Dr. Schultze-Gaevernitz, who enunciates this view, suggests that "the determination of relative strength, which is the function of a contest, could be equally well performed by an exercise of the intelligence, just as we test the pressure of steam by a special gauge instead of finding it out by the bursting of the boiler."*58 The implication of this argument is that, when all the facts are known to both sides, the range of practicable bargains is necessarily a single point, representing a wage equal to the one which battle would have established. This, as we have seen, is not so. It is true that an arbitrator, if he is to be successful, must act as an interpreter, and not as a controller, of economic tendencies, in the sense that he must place his award within the range of practicable bargains—a range, be it remembered, which the appointment of the arbitrator has probably somewhat widened. It is also true that, if all the facts are known to both parties, the wage rate which battle would have established will necessarily be represented by a point within this range. But it is not true that this wage rate is the only award open to the arbitrator which the litigants would be prepared to obey. The range of practicable bargains is not limited to that point unless, not only are all facts known to both sides, but also the costs which a fight is expected to involve is, for both of them, nothing at all.*59

Notes for this chapter

Cf. ante, Part II. Chapter IX. § 14.
Political Science Quarterly, vol. xii. p. 426.
Cf. Chapman, The Lancashire Cotton Industry, p. 211.
Schultze-Gaevernitz, Social Peace, p. 192.
Schultze-Gaevernitz, Social Peace, p. 136.
For a mathematical treatment of the problems discussed in this chapter, cf. Appendix A of my Principles and Methods of Industrial Peace.

Part III, Chapter VII

End of Notes

43 of 73

Return to top