The blue-collar worker of Europe and North America is the relative loser in the present, exceptional world prosperity. Freer trade and advancing technology are the most frequently cited rival explanations. A third alternative, job protection, is offered in this essay as a likely additional cause.

The tail end of the 20th century and the beginning of the 21st have been exceptionally kind to both capital and labour, but kinder to capital than to labour. Overall economic growth, apart from some sluggishness in Continental Europe and Japan and violent but brief upsets in Southeast Asia and Russia in 1998, went on more briskly and for longer than at any time in known history. While the rise from poverty was the most spectacular in China and India, even such hitherto unpromising areas as black Africa and much of Latin America began to share in the benefits of freer trade, relative peace and the rudiments of the rule of law.

Within an expanding world income, profits rose markedly faster than wages, so that the relative share of labour declined fairly continuously. This shift in relative shares concerned labour as a whole, and must not be confused with the quite different shift within total labour income in favour of those with higher skills. In fact, separately from the faster rise of profits than of wages, there was a widening of wage differentials, most pronouncedly favouring managerial, accounting and legal work and computing skills. Unskilled labour and labour using old-established technologies lagged behind. The net effect for labour incomes was a relative loss compared to the income accruing to capital. The era is widely perceived to be one of high profits, low pay.

There was and still is much aggrieved feeling about this by the blue-collar, the casual and the part-time workers in the Western world, which is understandable enough, and much righteous indignation by socialists of all hues which is only to be expected. The more muddle-headed blame the IMF, the World Trade Organisation, “unbridled” liberalism, greedy multinationals and the “dictatorship of the market.” It is fruitless to argue with them; if you do, they win by talking faster and louder. More reasoned inquiry about the cause of low pay in the midst of unparalleled prosperity focuses on two major trends, one in trade, the other in technology. The present paper sets equal store by a third. That trend is less widely understood than the first two, but worth close attention for that very reason. It is the rising ideology and the attendant legal machinery of job protection.

The diagnosis that freer trade favours capital more than it does labour runs roughly thus. In a more or less closed economy, capital formation raises the marginal product of labour and leads to higher demand for it. Since the labour force is limited, the wage rate will quickly catch up with the marginal product. As employment approaches the over-full level, labour’s marginal product may indeed fall, for lower quality workers are employed, labour discipline slackens, and shirking and dawdling involve less risk of sanction. The share of wages in total factor income reaches a maximum. As it now pays to substitute cheap capital for dear labour, the marginal product of capital recovers and capital formation is stimulated. The relative shares of the two factors of production swings back in favour of capital, until it reaches a maximum and the pendulum starts to swing back towards labour. For decades at a time, the relative shares of capital and labour may change only a few percentage points either way, for the pendulum need swing only a little in favour of one factor of production before it is quickly pulled back in favour of the other. A more or less fixed, inelastic supply of labour is the great stabilizer of this distributive machine.

When such an economy is opened up to the wide world, what happens depends on what the world is like, notably in terms of its factor endowments. Its stock of capital and its supply of labour are the decisive determinants of how freer trade affects income distribution.

Frédéric Bastiat envisioned a virtual railroad as a reductio ad absurdum in his essay A Negative Railroad, in Economic Sophisms, Series I, Ch. 17.

In our age, Europe and North America have opened up, first and foremost, to two very large areas in China and India with a huge rural population with no opportunity to deploy its productive potential but eager to do so, and a low stock of capital. As obstacles to trade were partly dismantled and transport costs shrank, the demand for labour of Western capital met, not the limited supply of labour hitherto available to it in the Western world, but the seemingly unlimited supply of Chinese and Indian peasants flocking from rural misery to slightly less miserable urban work. They did not physically move to Europe and North America; the garments, plastic toys, components, electronic sub-assemblies and gadgetry (and no doubt soon complex, highly sophisticated equipment, too) incorporating Asian labour did all the moving from East to West that was necessary to simulate the conditions of an almost infinitely elastic labour supply in the West. De-localisation of production from West to East, painful to its direct victims and politically poisonous, created more protectionist emotion than its tangible effects might have warranted, but it certainly added to the general sentiment that blue-collar workers in advanced countries were getting a raw deal at the hands of ruthless, greedy bosses trying to please ruthless, greedy financiers. Some of the measures proposed in all seriousness to stop de-localisation and curb greedy finance could match Bastiat’s famed virtual railway for silliness.

Under these conditions, as capital accumulation proceeds vigorously in response to the marginal product of capital staying high or rising, the demand for labour increases but the price paid for labour—the price of T-shirts, jeans, plastic articles consumer electronics—does not increase. In the Western world, the pendulum is not swinging back in favour of labour. Wages in “old” industries lag behind overall income growth and even more so behind profits, even as wages in China and India rise fast as they catch up with the sharply increased productivity of urban compared with rural work. The process leads to convergence of factor prices between West and East, though their actual equalisation is no doubt very far off. Meanwhile, in the West “globalization” is blamed, reasonably enough, for low pay.

* * *

Economists worth their salt have a more than merely intellectual commitment to free trade, and regard protectionist arguments with no more sympathy than Vatican prelates regard liberation theology. It is in part their subconscious disgust for findings capable of being turned against free trade, (such as those detecting some ill effects from “globalization”) that induces many economists to reject the thesis of “globalization” being the root cause of low pay. They are only too ready to ascribe it to technological change instead (and are supported in this stand by recent studies done at the OECD that minimise the role of Asian exports produced by cheap Asian labour and stress the effect of information technology).

When we hear the words “technological progress”, we almost automatically couple them with the words “labour-saving”. Indeed, if technological change is progressive, we should expect it to enable a given output to be produced with less labour, or more output produced with no more labour. We have the mental picture of a little man pushing a wheelbarrow filled with earth and next to it a great yellow earthmoving monster driven by another little man doing what it would take a hundred wheelbarrow-pushers to do.

If it is the case that technological progress is intrinsically labour-saving, then one should expect it to be reflected in a lower marginal product of labour (not to be confused with “labour productivity”, which is total output divided by the number of workers engaged in producing it, and includes the contribution to output of capital as well as of labour) or a higher marginal product of capital. It would explain why the share of capital in total income increases more than the share of labour.

But it is quite wrong to suppose that technological progress is typically, or even predominantly, labour-saving. Those who tacitly assume that it is typically labour-saving nowadays have information technology in mind. However, if you reflect that a few decades ago a mainframe computer would fill a good-sized room and cost many times its handy-sized contemporary equivalent, it will dawn on you that even information technology can be capital-saving. In fact, changes in production equipment can go either way and indeed both ways at the same time, though labour-saving may be more characteristic of it

Apart from fixed equipment, though, much of the rest of capital employed in the production process is more likely to be hospitable to capital-saving than to labour-saving technology. Two kinds of capital are involved: work-in-progress and goods in transit.

Work-in progress tied up in producing a given volume of output can be reduced by using statistical probability to estimate the need for various inputs at various times, and by more precise and reliable delivery schedules of materials and parts thanks to advances in logistics. The “just-in-time” methods made famous by Japanese car manufacturers is but a prominent example of a much wider phenomenon that has vastly reduced the amount of capital absorbed in work-in-progress.

Probably more important by a great deal is the effect of advancing transport technology on the volume of both raw materials and finished goods in transit. Depending on the geographical structure of commerce, all goods travel a greater or lesser distance between final seller (the farmer, miner, lumberman or manufacturer) and final buyer (the consumer). If the average good spends three months in transit on road, rail and sea and in warehouses and depots, the transit function absorbs a volume of capital equal to 25 per cent of physical (goods only) GDP. If advances in transport technology cut average transit time to one month, the capital requirement shrinks from 25 to 8 per cent of physical GDP.

These figures, of course, have not the remotest pretension to accuracy, yet may be near enough to reality to illustrate the vast effect that technology is liable to exert in a capital-saving direction. If the numbers are anywhere near reality, the belief that technological progress is intrinsically labour-saving must be at least suspended. As a consequence, it can hardly serve as the most important and most probable explanation of low pay, for capital-saving would, if anything, raise wages.

* * *

Neo-classical economics teaches that in large-number interactions where many agents deal with one another, and all or most act so as to maximise some entity that can be represented by “the measuring rod of money”, capital and labour will each earn their marginal product. It will be only just worth while to employ the last unit of capital at the going rate of interest and the last unit of labour at the going wage rate.

There are two standard objections to this theorem. One is that it works only under diminishing or constant returns to scale, but breaks down under conditions of increasing returns, where paying capital and labour the values of their marginal products would require more than the total product available. The other, close to socialist doctrine, is that it is impossible to identify the marginal product of a particular unit of capital or labour, for all product is social and must be imputed to society as a whole. Therefore society alone is entitled to decide how capital and labour are remunerated. I shall pass by these two objections. A third seems to me more interesting.

Let us admit that in a static economy, which reproduces itself without any change from one day to the next, a firm can both ascertain the marginal product of its labour force and know that tomorrow and the next day it will be the same as today. In that case it will hire labour if its marginal product is higher than the wage rate, and fire it if it is lower.

In a dynamic setting that keeps changing in all kinds of ways, the firm cannot rationally rely on current experience alone. It needs to form expectations about what the marginal product of its staff will be at future dates. These expectations, though obviously unreliable, are still the best guide the firm has as to whether it should hire, fire or do nothing. They form a probability distribution, some values of it lying above the going wage rate, others lying below it. Basic decision theory suggests that if the mathematical expectation (“certainty equivalent”) is lower than the wage rate, the firm should “restructure”, “outsource”, “de-localise” or otherwise contrive to fire some of its workers, lifting the marginal product of the remaining staff.

However, in an economy with freedom of contract, this decision “model” rests on false premises. Suppose that as the future rolls on, times turn out good and the firm’s best expectations prove to have been right. The marginal product is comfortably above the wage rate. It would have been right to hire more labour. Suppose, however, that the firm did not do so, because it was frightened off by the unfavourable half of its expectations, which pulled the “certainty equivalent” down to, or below, the actual wage rate. Now this would have been a rather foolish way to act, for if the firm had hired more labour and found that this did not in fact pay, for times turned out to be bad, it could have without much ado fired those it had hired and suffer little loss; while if it never hired the extra labour and times have in fact turned out to be good, it would suffer an opportunity loss. To rectify its mistake, it could at best belatedly scramble and hire the labour left over by its less timid competitors, while at worst it would miss the chance the good times have offered. Therefore the right decision would have been to hire the extra labour to start with.

The logic of this argument tells us clearly enough that under complete contractual freedom where labour can be hired or fired subject only to the agreed terms of the employment contract, and the length of notice is freely negotiated between employer and employee, there will be a distinct “speculative” incentive for firms to expand. Evidently, if enough firms respond to this biased incentive, it will prove to be a self-fulfilling prophecy. The expansion of many will justify the expansion of each. Subject only to the proverbial slip between cup and lip, expectations held with some, albeit limited probability that times will be good could succeed to bring about full employment and good times.

It needs no great analytical acumen to see that when freedom of contract is suppressed and job protection of some stringency is put in its place, the above argument is turned on its head. If firing workers is made excessively costly, requiring a long-drawn juridical process, or becomes impossible unless justified by manifest problems of the employer’s solvency, the unfavourable half of the probability distribution of future marginal products becomes menacingly relevant, for once it hired them, the firm has to carry its workers almost indefinitely, whether or not it pays to do so. The “speculative” incentive is not to hire, perhaps even not to replace staff lost by natural attrition. A powerful bias toward unemployment is created, and reinforces itself in the manner of self-fulfilling expectations.

As the inexorable force of politics by majority rule continues to strengthen job protection by both labour legislation and the pressure of public opinion, the firm must come to regard its wage bill as becoming dangerously like a fixed cost which it is only prudent to keep lower than would be profitable if it were a truly variable cost. Unemployment, the bias that job protection imparts to the firm’s expectations of probable future outcomes, and the loss of labour’s bargaining power, all combine to keep wages low. Job protection is certainly not the only nor even probably the most important reason for the least well-off getting the worse deal in the present era of burgeoning growth and economic serenity. But it is a cause that was meant to have exactly the opposite effect and that it would be fully within political society’s power to remove if only its perversity were more widely understood.


*Anthony de Jasay is an Anglo-Hungarian economist living in France. He is the author, a.o., of The State (Oxford, 1985), Social Contract, Free Ride (Oxford 1989) and Against Politics (London,1997). His latest book, Justice and Its Surroundings, was published by Liberty Fund in the summer of 2002.

The State is also available online on this website.

For more articles by Anthony de Jasay, see the Archive.