The Rise and Fall of the Sliding Scale
or
Why Wages Aren't Indexed to Product Prices
November, 2003
Christopher Hanes
Department of Economics
SUNY-Binghamton
P.O. Box 6000
Binghamton, NY 13902
(607) 777 5487
[email protected]
Abstract:
Sliding scales - agreements between employers and unions that linked wage rates to product prices
- were widespread in some industries, and tried in others, in Britain and the United States from the
1860s through the 1930s. I describe how sliding scales worked in practice and how they were
viewed by contemporary employers, unionists and outside observers. I argue that sliding scales
were not adopted for the reasons suggested by most theories of wage indexation, but rather in
order to reduce the frequency of strikes in an institutional setting where there was no third-party
enforcement of labor agreements. This can explain both the presence of sliding scales in the
nineteenth and early-twentieth centuries, and the disappearance of sliding scales since the 1930s.
-1Sometime before the year 1841, a British ironmaster named G.B. Thorneycroft decided that the
wages of his puddlers "should fluctuate with the price of 'marked bars' - these words indicating a
quality of iron that then enjoyed a high reputation...The puddlers received, as a rule, 1 shilling for
each pound of the selling price" (Munro, 1889, p.141). By the 1880's, arrangements linking wages to
product prices were generally referred to as "sliding scales." In Britain, they had become standard
practice for skilled iron and steel workers (Munro, 1885) and common in coalmining, where wages
of about 120,000 men were linked to coal prices (Munro, 1885, 1889). In the metals industries
sliding scales were gradually extended to unskilled as well as skilled occupations (Pool, 1938) and
covered 220,000 men as of 1925 (Great Britain Ministry of Labour, 1925, p. 269). In coalmining,
“conciliation boards” set up in the 1890s effectively applied sliding scales (Bowie, 1927; Treble,
1987). In 1921 coal miners’ unions and employers abandoned the sliding scale, but only to replace it
with a similar scheme that linked wages to the mines’ operating profit (Bowie, 1927). In coking and
iron mining sliding scales were used through the 1920s, if not later (Great Britain Ministry of
Labour, 1925, p. 269). At least some “profit-sharing” schemes in gasworks after the 1880s were
actually sliding scales (Matthews, 1988). Textile unions and manufacturers negotiated the terms of a
sliding scale for many months in 1899, but did not adopt it (Price, 1901).
In the United States, as in Britain, sliding scales were most widespread and long-lived in
iron and steel. By the 1870s they determined skilled workers’ wages in most plants manufacturing
iron rails, iron sheets or tin plate (Massachusetts Bureau of Statistics of Labor, 1881, p. 18; Asheley,
1903, p.154) and in many steel plants (American Iron and Steel Association, 1888, pp. 117, 297).
Unions of iron and steel workers were expelled from most plants in the 1900s after the formation of
the U.S. Steel Corporation, but where unions hung on they continued to bargain in terms of sliding
scales through the early 1930s (Robinson, 1920; Chazeau and Stratton, 1937, pp. 143-144). There
were many instances of sliding scales in mining: coking coal (Jeans, 1902, p. 19); Pennsylvania
anthracite, “forced upon the operators in 1869 by trade union pressure” and used off and on through
-2the 1900s (Fisher, 1942, pp. 287-292); Colorado silver mines from 1893 to 1896 and copper mines
from 1907 (Greenfield, 1960, p. 116-122). A few instances of sliding scales can be found elsewhere:
in the textile mills of Fall River, Massachusetts from 1905 to 1910 (Howard, 1920), in glass
factories from the 1880s through the 1900s (Davis, 1949, pp. 132, 187), and in copper smelting and
refining through the 1930s (U.S. Bureau of Labor Statistics, 1943, pp. 28-29).
Sliding scales fascinated contemporary economists and public officials involved in labor
relations. In 1881 Carroll Wright recommended the adoption of sliding scales in Massachusetts
(Massachusetts Bureau of Statistics of Labor, 1881, p. 74). Members of special commissions on
labor relations conducted by the British parliament in the early 1890s, and the U.S. Congress around
1900, frequently inquired about the use of sliding scales, spoke favorably of industries that used
them, and quizzed representatives of other industries as to why they did not (Great Britain Royal
Commission on Labour, 1892, 1894; U.S. Industrial Commission, 1901). An Edinburgh University
professor named J.E.C. Munro proclaimed the sliding scale to be "the greatest discovery in the
distribution of wealth since Ricardo's enunciation of the law of rent. That it has a great future before
it, not only in the iron and coal trades, but in other industries, I have no doubt" (1885, p. 26). One
noted benefit of sliding scales was the adjustment of wages to monetary shocks:
For there can be little doubt that the appreciation of gold is at present a disturbing factor of no small
magnitude in industrial relations, and that in some instances workmen exhibit a somewhat
pertinacious insistence upon nominal wages...But the automatic adjustment of wages to prices
effected by a sliding scale entirely avoids the difficulty occasioned by a general appreciation of gold.
(Price, 1887, p. 96).
Alfred Marshall recommended that labor arbitrators “adopt a self-adjusting sliding scale” that took
into account the cost of raw materials as well as the product price, with a caveat that “it is often
difficult for working men even to ascertain the real price at which the product of their labour is
being sold” (Marshall and Marshall, 1881, p. 216-17; see also Marshall 1887, p. xix). Arthur Pigou
similarly judged that a sliding scale “may easily, for all its inaccuracies, do a considerable amount of
good” (1905, p. 104), though he later observed that “In an industry engaged in making a number of
-3different articles, particularly if the quality and nature of these varies from time to time - ships, for
example, or ladies hats - the technical difficulty of finding any price, or combination of prices, that
will correctly indicate variations in the demand for labour is very great...Hence the range over which
the remedy of sliding-scales can be applied is restricted somewhat narrowly by purely technical
incidents” (1927, p. 287).
Modern economic theory agrees with those early economists on the potential usefulness of
linking wages to product prices. Weitzman (1984) argues that the widespread adoption of wageindexing schemes like sliding scales would boost aggregate employment and dampen the real effects
of monetary shocks. Apart from the macroeconomic benefits described by Weitzman, models of
long-term employment contracts imply that employers and employees would have strong private
incentives to index wages to prices of their product, or to product prices less costs of non-labor
inputs. This holds even if employees are risk-averse, and whether or not wages are also linked to
other variables like prices of consumption goods. Conditioning the wages of a workforce on future
realizations of output value can benefit both sides by preventing separations when the value of a
worker’s product in the enterprise falls, but remains above the prospective value of his next-best
job. An employer would gain a more certain stream of profit. An employee would gain a smaller
risk of layoff, which could more than compensate for increased variability in his wage if he is not
laid off. To rule out the use of sliding scales, it is argued (or simply assumed) that such indexation is
impossible because of the “technical difficulty” referred to by Pigou and Marshall: employees are
unable to observe variables sufficiently correlated with the value of output and/or raw materials cost
(for example Blanchard, 1979; Hart, 1983; Azariadis and Stiglitz, 1983; Hall and Lazear, 1984).
Historical sliding scales have been discussed in this context, when they are discussed at all, in
economics literature of recent decades (Weitzman, 1984 pp. 78-79; Hall and Lazear, 1984, p. 255).
Treble (1987) argues that sliding scales were adopted in nineteenth-century British coalmining to
transfer risk from mineowners to workers, and the industry moved away from sliding scales “as a
-4part of a process by which the burden of the risk induced by shifts in the demand function for the
output was shifted from the employees to the employers” (p. 85).
From this point of view, the history of sliding scales in general is puzzling. In Britain,
sliding scales never spread beyond metals and mining. There they remained in force until the advent
of general wage and price controls in the Second World War, but they were not revived after the war
(Haynes, 1953, p. 26; Burn, 1961, p. 27). In the United States, it is hard to find instances of sliding
scales after the 1930s. In 1940 the U.S. Bureau of Labor Statistics reported the existence of sliding
scales in just one industry, the mining and smelting of nonferrous metals (such as copper, zinc, and
nickel) (U.S. Bureau of Labor Statistics, 1940, p. 13). In 1951, the BLS observed that even in this
industry “the general practice of gearing wages to prices declined appreciably in the early 1940's”
(U.S. BLS, 1951, p. 48). It is almost (though not entirely) impossible to find sliding scales in recent
years.1 Perhaps there was an decrease after the 1930s in the share of workers making goods with
easily definable prices. But there was certainly a vast increase across the 1930s and 1940s in the
number of employers making long-term contracts with unions (Freeman, 1998). There was also an
increase in information available to workers about prices of products and materials, in the form of
wholesale or producer price series collected by government agencies. In the U.S., at least, the quality
and scope of such series have improved steadily over the course of the twentieth century (Hanes,
forthcoming), and there have been many instances of indexing long-term contracts other than wage
agreements to PPIs (U.S. BLS, 1991).
Why were sliding scales common and long-lived in metals and mining, but rare in other
sectors before the 1930s? Why did sliding scales disappear almost entirely after the 1930s? In this
paper, I attempt to answer these questions. First, I describe the characteristics of sliding scale wage
1
Union contracts with the Magma Copper Company linked hourly wages to
copper prices, along with the CPI, through 1992, when the arrangement was replaced by
one linking hourly wages to operating profit (Charlier 1991, 1992) along the lines of the
1920s British coal scheme. In the 1990s union contracts with Inco, a Canadian nickel
producer, linked wages to the spot price of nickel (Bagnell, 1997).
-5agreements in Britain and the U.S., how sliding scales were viewed by contemporaries, and why
employers and unions abandoned sliding scales after trying them, or rejected proposals to adopt
sliding scales. I find that it was hard for employees to observe output values and materials’ costs
even in enterprises where outputs and materials were relatively homogenous goods with observable
open-market prices. When employers and unions did adopt sliding scales, it was not for the reasons
suggested by models of long-term contract indexation, but rather to reduce the frequency of strikes
in an institutional environment where long-term employment contracts did not exist. I present a
model, based on standard models of strikes as a consequence of asymmetric information (such as
Hayes, 1984; Hart, 1989), to demonstrate how sliding scales could prevent strikes by revealing
employers’ private information about product demand. I argue that this best explains what
contemporaries said about sliding scales in the nineteenth and early-twentieth centuries, and helps
explain why sliding scales disappeared after the 1930s.
1. The structure of sliding scales
Most sliding scales were negotiated between a labor union or alliance of unions, on the one
hand, and a firm or trade association of firms on the other. The terms of a sliding scale defined a
relation between a product price, or set of prices, and money wage rates to be paid for specified jobs
(rather than the wage for a specified employee) by piece- or time-rate. Nearly all sliding scales set
minimums below which wages could not fall, no matter what happened to the product price(s);
many set maximums. In some scales the relation between wages and price(s) was nearly continuous,
for example, a schedule giving percent increases over the minimum wage rate as a function of per
cent differentials between the price(s) and a base level. Other scales fixed wide, discrete steps.
Sometimes the price governing a wage was that of a good actually handled by the worker,
but that was not always or even usually the case. Often wages for workers producing a variety of
goods were set by the price of just one or two products. In American iron mills, for example, all
wages were set by the price of iron bars (Robinson, 1920). Sometimes wages were set by the price
-6of a good further down the chain of production from the workers’ own output. Wages of British iron
ore miners were set by the price of pig iron (Munro, 1885, p. 8). In Wheeling, West Virginia, where
the principal industry was the manufacture of nails, nail prices determined coal miners’ piece rates
(Massachusetts Bureau of Statistics of Labor, 1881, p. 18).
It was generally recognized that the terms of a sliding scale had to be adjusted in response to
changes in outside labor-market conditions or prices of nonlabor inputs. Munro (1885) observed:
A sliding scale does not bind the miner to remain in the employ of his master. It only fixes the
wages whilst the relation of employer and employed continues, and leaves either party free to
determine that relation by the usual notice to quit. That there are many obstacles to the free
migration of labour is well known, but in so far as it does exist it tends to establish an economic
minimum wage for every sliding scale (p. 25)
Sometimes an employer would refrain from making a wage cut allowed by an existing agreement
because he feared it would cause workers to quit and take other jobs, as the agreed-upon wage scale
would be lower than the wages workers could earn elsewhere (Massachusetts Bureau of Labor
Statistics, 1908, p. 267). In British coalmining,
the wage and the price ruling at the time have been taken as’standards.’ Then..the colliers have
become dissatisfied with this basis, and pressed for its revision; the reason invariably being that they
found, or though they found, that their average wage under the sliding scale did not give them the
same net advantages as were being earned by other trades with which their labor could be put into
comparison. In other cases the masters have revised the standards for exactly the opposite reason
(Smart, 1895, p. 74).
The British Commission on Industrial Relations observed that:
A common difficulty in settling or revising the basis of sliding scales arises from one party or the
other contending that other circumstances besides the average wages and selling prices for a
preceding term of years should be taken into consideration, such as changes in the cost of material,
or the state of the labour market, or the relative wages of men in other districts and like industries,
or competition with other districts and countries. An experienced witness giving evidence on behalf
of employers in the iron and steel trade thought that, in view of this,..any sliding scale would require
revision every few years (Great Britain, 1894, p. 42)
In practice, sliding scales were revised much more often than “every few years.” In British metals,
agreements were not expected to last more than a year or two (Pool, 1938, p. 158). In American
-7iron and steel the sliding scale was renegotiated once a year (U.S. Bureau of Labor, 1904, p. 237).
Many sliding scale agreements did not have a term at all, allowing either party to withdraw after
giving notice for a specified number of weeks (for example in textiles [Massachusetts Bureau of
Labor, 1908, p. 264]). Adjustments to materials costs were frequently anticipated by including
materials prices in the sliding scale formula. In the British tinplate and galvanizing industry (Pool,
1938, p. 161) and Fall River textiles (Thayer, 1909), for example, wages were set as a function of
the difference between a product price and prices of raw materials in the quantity required to
produce a product unit. This differential was referred to as the employer’s “margin.”
2. Practical problems with sliding scales
Why did unionists and employers abandon sliding scales, or refuse to adopt them in the first
place? In many cases, it was because it was impossible to define a product price, or impossible for
workers to observe product prices. In the early 1890s the secretary of a union in the nut and bolt
industry, one R. Juggins, told members of the British Royal Commission on Labour that the sliding
scale:
does not apply in my trade at all, because it is very difficult to get at the prices. It is not like the iron
or coal trade where the prices are published. Ours being a trade which embraces a great number of
technicalities and specialities it is impossible for us to find out what price the employers get for
them (Great Britain, 1892, p. 471).
A few years later the president of the glass workers’ union told the United States Industrial
Commission that, despite his union’s experience with sliding scales in some lines of goods,
our union does not like the sliding scale...The sliding-scale system must be based now on certain
rebates and discounts that you can hardly find out. You must depend a great deal on the honesty and
integrity of the manufacturers. It is hard to reach. They have an understanding among themselves
that they will sell a certain amount of glass to a dealer, and if he buys so much he will get a certain
rebate (U.S. Industrial Commission, 1901, p. 934).
A Massachusetts state labor arbitrator was asked:
Q. Do you think it is possible in the boot and shoe industry in the State to establish between the
workmen and the employers...a sliding scale agreement, changable, say, in 90 days by the working
committee on each side. Would that be more advantageous to the business in giving stability of
employment to the men, and a perfect knowledge on the part of the employers as to what would be
-8the cost of the goods and what they could market them for?
A. Theoretically, I should say yes; practically, I should have little expectation of anybody carrying it
out in good faith...I can conceive of a shoe manufacturer meeting his workmen to make up a new
list, as you say, on a sliding scale that would take into consideration the present cost of production
and stock, but the first difficulty you would run against would be this, that the workmen would not
believe a word of what the manufacturer said as to what his goods cost him or what he sold his
goods for (U.S. Industrial Commission, 1901, p. 918).
In the textile mills of Fall River, Massachusetts, unions and employers adopted a sliding
scale in 1905. The first scale adjusted wages every two weeks as a function of the “margin” between
the price of one type of plain print cloth and the price of raw cotton required to produce the priced
quantity of cloth, with both prices taken from a New York City commercial publication
(Massachusetts, 1906, p. 195). It became apparent that this product’s price was not an adequate
indicator of changes in prices for the mills’ full range of products. Though Fall River was “the
center of the plain rather than of the fine goods manufacture...products are less diversified than
elsewhere” (Howard, 1920, p. 14), “probably less than one-third of the product” was plain goods,
“the main product being ‘fancies’ and ‘odds.’ Under the present system a fall in the price of print
goods would lower wages, even though the mill was not running on prints at all and indeed it might
be making a large profit on fancies” (Massachusetts, 1906, p. 195-96).
Another problem was accounting for the cost of raw cotton. The cotton price in the original
scale was a spot price, but the workers (or the union leaders) soon came to understand that this was
not the same as the cost of the cotton input, because of “the buying of raw materials in the future on
present quotations, and the existence of stocks” (Chapman, 1903, p. 191). The unions “contended
that an advance in the price of cotton did not mean an increase in the cost of manufacturing unless
the agents actually bought cotton at the higher price,...The mills had been stocked with cotton, so
that it was unnecessary to buy at the high quotation” (Lincoln, 1909, p. 458).
To mitigate these problems, the scale was amended to base wages on an average of prices of
-9two types of plain goods rather than one, and set wages for six months at a time, based on six-month
averages of cloth and cotton prices. According to a contemporary observer,
No doubt the six-month basis was an improvement. But there is a difficulty present no matter upon
what intervals agreement is reached. There is not much object in averaging raw cotton prices for a
week or even for six months with a view to determining wages unless it is known that the cotton
used in production was actually bought at the prices averaged. Examination of commercial statistics
shows that by far the heavist buying of cotton stocks by mills..is in the months of October,
November, December and January of each year...It is evident that any system of basing wages on
margins averaged for six-month periods ending in May and November does not fairly allow for the
uneven distribution of cotton purchases throughout the year. And yet an adjustment of the dates of
margin and wage calculations to meet this seasonal fluctuation is not easy for the fluctuation is not
regular (Howard, 1920, p. 36).
In 1910 the unions and firms gave up on sliding scales and reverted to negotiation over fixed wage
rates (Howard, 1920).
In mining, materials costs were unimportant. Products were fairly undifferentiated
commodities. It could nonetheless be difficult to observe the product price, because most mines sold
output on long-term contracts rather than on spot markets. Thus, prices for sliding scales for most
mining sliding scales were not taken from business publications, but rather were calculated from
employers’ books. This could require an investigation by at least two accountants, one hired by the
union (for example Smart, 1895, p. 66). The cost and trouble of examining the books caused a
“difficulty of adjusting wages at short intervals of time...in a large coal firm the frequent
examination of books would probably cause much inconvenience” (Munro, 1889, p. 144). But if
wages were adjusted infrequently,
when prices rise, wages do not follow them until the end of the revision period...the colliers become
impatient...their attention is attracted by additional men being drafted into the pits in order that their
masters may secure the rise in price over a larger output. They thus become aware that, if not bound
by the sliding scale, this would have been a favourable opportunity to press for an advance of wages,
and they perhaps throw off the scale (Smart, 1895, p. 94)
Apart from the costs of ascertaining contract prices, there were difficulties caused by the difference
between them and current spot prices, since “the rise of wages which might be expected from the
- 10 current newspaper quotations for coal does not come at the end of the revision-period, and may
never come at all if prices fall again before the contracts expire” (Palgrave, 1896, p. 411).
“When..the miner, expecting an advance of wage, is told that half or more of the output of the
colliery has been sold for months ahead at the previous low price, it is very natural that he should
either disbelieve the figures, or ask by what right his labour has been thus disposed of” (Smart,
1895, p. 95). In one coalmining sliding scale this problem “was met by the clause that ‘any contract
for the sale of coal for a period of more than twelve months shall not be taken into account for more
than six successive audits of two months each” (p. 86).2
The iron and steel industries’ sliding scales were generally judged to be the most successful.
Unlike mines, metals firms did not usually enter into long-term sales contracts: “the ‘long contract’
or selling-ahead’ system, which..has always endangered the sliding scale in the coal industry, does
not exist at all in the iron and steel trades” (Asheley, 1903, p. 150). Prices of most goods were
closely correlated, so prices of just a few goods could stand for output prices in general: “when
anything goes up on one piece of iron, no matter what its shape, it comes up all along the line in the
same degree” (U.S. Industrial Commission, 1901, p. 98). Still, the product market was such that it
could be hard to determine what any goods’ prices really were. In the 1870s and 1880s, American
iron plants’ sliding scales linked wages to the official published prices of the employers’ trade
association, “an organization whose sole function, it appears, was to fix prices,” but: “This scheme
caused considerable friction, because iron was frequently reported to be selling above the
official..rate” (Robinson, 1920, pp. 146). The agreement was revised to determine an average selling
2
Pennsylvania anthracite was subject to a problem that was unique, but which
illustrates the range of problems that could arise in practice. At first, wages were based on
coal prices at a shipping point within the mining region. But the railroads shipping the
coal out of the region were owned by the same companies operating the mines, and the
miners “thought themselves victimized by the high freight rates which the large
companies..as common carriers charged themselves as operators” (Virtue, 1900). When
arbitrators re-established a sliding scale after the great strike of 1902, they were careful to
base wages on coal prices “at New York Harbor” (Fisher, 1942, p. 292).
- 11 price from a sample of transactions provided by the trade association, where “in case of doubt, a
committee of the union might examine the employers’ books and bills of sales” (p. 147). This
process took time and effort on both sides, partly because the manufacturers wanted to keep their
sale prices secret from other firms. In 1899 a former union president described it to the U.S.
Industrial Commission:
they meet us in these conferences, and agree with us that we shall select a number of mills, and that
those manufacturers will send every two months to the secretary of their organization - there is an
organization of the manufacturers, but they do not control the selling price as they used to do - and
that they shall take a statement of all sales made in those particular mills for those two months, and
the price at which the articles were sold; make a sworn statement to that effect, and send it sealed to
the secretary at the end of every two months. Our committee goes over and these statements are
opened in their presence and by the committee...The other manufacturers are very jealous of the
prices at which they sell, so that we are in secrecy ourselves. We appoint a committee and they
determine it for our men. They are satisfied with the determination of the thing; and the
manufacturers themselves do not know what the others have sworn to except as they gather it on the
outside - on the market, so we have a pretty tight compact (U.S. Industrial Commission, p. 96)
Often, as in this case, it was necessary for union members to trust their leaders’
interpretation of information that could not be revealed to the members, or easily understood by
them. An employer told the British Labour Commission that in his trade problems were caused by
the difference between producers’ selling prices and “the prices which speculators get.” Employees
“saw, say, [a price of] 66 s[hillings per ton] quoted in the newspapers, they only found (I am
instancing a specific case) 45 s realized over the quarter” on the employers’ books. While the
union’s leaders understood the difference, they “have behind them a large body of persons not very
familiar with the operations of the trade whom they have to persuade” (Great Britain, 1892, volume
1 p. 77). One reason the Fall River sliding scale was amended to adjust wages at six-month rather
than two-week intervals was to make “the relations between the fluctuations of the cotton market
and the rate of wages clearer to the uninformed operatives” (Lincoln, 1909, p. 460; see also pp. 457458). The British Commissioners concluded that that to make sliding-scales work, in some
situations “the only remedy seems to be..that the workmen at large should trust their representatives,
- 12 and delegate the fullest powers to them.” (Great Britain, 1894, p. 42).
3. Why unions and employers adopted sliding scales
3.1 What contemporaries said
Some contemporary observers of the 1860s to the 1930s stated that the benefits of sliding
scales include stabilization of profits and/or employment. These are the benefits of indexing wages
to prices within a long-term contract. In 1882, a British labor arbitrator noted that “in times of
depression its tendency must be to enable the capitalist to keep his works going, and his hands
employed until brighter times dawn again upon them” (quoted in Price, 1887, p. 79). In 1889, the
chairman of the U.S. Industrial Commission asserted that a sliding scale “makes labor more stable
in its hours and days” (U.S. Industrial Commission, 1901, p. 386). The authors of a 1940s BLS
bulletin stated that in copper mining “The automatic adjustment of production cost in response to
changing metal prices tends to stabilize company earnings and may contribute somewhat toward
regularity of employment (U.S. BLS, 1943, p. 8).
However, according to every contemporary statement on the subject that I have found, the
main reason that union(s) and employer(s) adopted sliding scales - usually, the only reason
mentioned - was to reduce the frequency of strikes and lockouts, or “industrial warfare” in the
language of the time. Such statements include Great Britain (1894, p. 41), U.S. Industrial
Commission (1901, p. 344), Pool (1938, p. 172), Marshall (quoted by Smart, 1895, p. 63). Jevons
(1915) describes the development of the sliding scale in British coalmining as follows:
During the nineteenth century, whenever a slump of trade set in, coal owners found the margin of
profit being turned into a loss, and were forced to try and economise by reducing wages. More often
than not this led to a strike, in which the men had to give way,..The men, finding that they had to
accept a reduction of wages from time to time, took care to agitate for an increase when trade
improved, with the result that there were sometimes also strikes when trade was improving as well
as when it was collapsing..a colliery owner would give way after resistance because with increasing
prices there was ample room for higher wages.
The frequency with which labour disturbances arose solely from the rise and fall of the
market, led to a general desire to make some arrangement betweenn masters and men which would
allow for, and regulate, changes in wages in accordance with the state of trade.
- 13 After a good deal of local discussion, the idea of the Sliding Scale, in which wages varied
automatically with the price of coal, took definite shape (p. 490).
Read (1894) argued that sliding scales were especially useful in the British coal fields of
South Wales, because their product market made strikes especially costly. Most South Wales coal
was used to fuel steam engines,
the Welsh mineral being far away the best fuel yet discovered for steam-raising purposes. But it is
an indispensable requirement that deliveries from the Welsh ports be regular...the foreign railways
and navies, and the great steamship lines, must be able to depend without doubt upon the due
delivery of the primary requisite for power. Hence the need for continuity of output; and this is
obtained through the sliding scale agreement (p. 332).
(See also Morris and Williams 1960, p. 175).
Most observers, even those otherwise critical of sliding scales, judged that generally “sliding
scales have succeeded..in doing that for which they were chiefly proposed, in making the process of
distribution more peaceful” (Chapman, 1903, p. 195). Smart (1895) observed that in British coal
“during the period of depression since 1891, which resulted in the great strike of 1892 in Durham
and the disastrous lockout of 1893 in the Midlands, South Wales, regulated entirely by the sliding
scale, passed through the crisis without more than two or three weeks’ interruption of work” (p. 64).
In the 1890s Sidney and Beatrice Webb (1897) observed that, in the branches of the British iron
trades where arbitration set wages by sliding scales, “on every occasion the arbitrator’s award has
been accepted by both employees and employer” (p. 232); “the Board itself was able, on eight
separate occasions, to agree to advances or reductions without troubling the arbitrator at all” (p.
233). As the use of sliding scales spread throughout British iron and steel in the early twentieth
century, the absence of strikes in the industry became notable (Knowles 1952, p. 172). According to
Pool (1938), iron and steel was:
the only important trade both to use these scales and the remain free from major industrial disputes
(apart from the general strike of 1926) for half a century..During the post-war period, for instance,
iron and steel workers accepted wage reductions greater than those imposed in most other
industries, without any conflict arising...there can be no doubt that, in the absence of sliding scales,
- 14 they could not have been negotiated without serious risk of a stoppage (pp. 172-173).
Even in Fall River textiles, where the sliding scale was abandoned after a few years, it was judged to
have been successful in forestalling strikes. It was adopted soon after the “Great Strike of 1904",
when workers struck for nearly six months because employers had cut wages 12.5 percent following
a prolonged decrease in demand for their cloth (Howard, 1920, p. 20). The sliding scale was in
force, with wages adjusted every six months, when the demand for output fell again after the Panic
of 1907.
The events of the next six months, ending in May 1908, illustrate best of all the chief benefit to be
derived from a sliding scale of wages...the sullen preparations for resistance to an expected
reduction of wages, which usually marked a sharply declining market, were entirely absent..When
on May 25 the time again came for the semiannual adjustment of wages,...it was found that [the
scale] called for a reduction in wages..amounting to a general reduction in wages of all the workers
of 17.94 per cent (Lincoln, 1909, p. 463).
There was no strike, even though this was “the heaviest reduction that has ever been made at one
time in the history of Fall River” (Massachusetts Bureau of Labor Statistics, 1908, p. 260).
3.2 Sliding scales and strikes in the absence of long-term contracts
Why did contemporaries describe sliding scales as a way to avoid strikes? Models of strikes
as a consequence of asymmetric information show that strikes can occur if the employer knows
more than the union about the value to the employer of the union’s cooperation (Hayes, 1984; Hart,
1989; Cramton and Tracy, 1992). “Profitable firms lose more from a strike than unprofitable firms
and hence will settle early for high wages, while unprofitable firms will be prepared to delay
agreement until wages fall. The reason that the parties cannot do better by avoiding the strike and
sharing the gains from increased production is that there is no way for an unprofitable firm to
‘prove’ that it is unprofitable except by going through a costly strike” (Hart, 1989, p. 25). In other
words, it is best for the union to follow a negotiating strategy that occasionally results in strikes,
despite the costs of lost production opportunities, because it allows the union to gain a higher wage
- 15 in response to information initially possessed by the employers alone. At the same time, a union
may be willing to enter a contract with an employer that would prevent the union from going on
strike within the term of the contract, in order to attract more capital to the enterprise, which
improves the combination of wages and employment available to the union. Absent a long-term
contract, an employer would reduce capital investment in anticipation of the union’s future efforts to
extract capital’s quasi-rents (Grout, 1984). Within a long-term labor contract, wages are linked to
variables like product prices in order to condition contract terms on information that neither party
can observe when the contract is signed but which will be available both parties in the future. A
contract is possible if each party faces a substantial cost of violating the terms of a contract, beyond
the cost of an immediate breakdown in the employment relation. The cost may be imposed by the
state or, under limited conditions, correspond to the loss of a valuable reputation for holding to
agreements.
Contracts between employers and unions were not possible in Britain or the United States
before the 1930s. As far as the state was concerned employees were always equally free (or equally
forbidden) to strike, whether or not the employer had a “contract” with a union or contracts with
individual members of the union (except during and just after the First World War, when both
countries’ governments regulated employment bargaining to prevent strikes and maintain war
production) (Smith, 1950, p. 839; Steinfeld, 1991). A union’s actions might be constrained by
concern for its reputation, but it is easy to find examples of unions abandoning agreements and
going on strike before the end of an agreed-upon term (such as Fisher 1942, p. 283; Read, 1894, p.
334).
I argue that the sliding scales of the nineteenth and early twentieth centuries were not an
instance of long-term contract indexation, but rather an alternative to strikes as a device to extract an
employer’s private information about product demand. A sliding scale could give union members
high wages when the value of labor’s product was high, and low wages to preserve employment
- 16 when product value was low, while avoiding strikes and maintaining higher capital investment in
the enterprise. This can explain why contemporaries described the prevention of strikes as the chief
benefit of sliding scales. It can also help account for the disappearance of sliding scales after the
1930s. In the next section of the paper, I present a model to demonstrate my argument.
4. Model
4.1 Overview of the model
This model describes interactions between a union and an industry - a set of firms producing
the same good. I assume that any agreed-upon terms of employment set the wage per worker but
leave the employers free to choose the employment level. This assumption is hard to justify
theoretically but anything else would be wildly unrealistic (for a discussion of this point see Farber
[1986]). The union’s preferences across wage rates reflect those of a union member who acts to
maximize his expected wage income subject to given union rules on voting and allocation of
employment. The firms behave competitively in their product market, which is arguably realistic for
some cases of sliding scales. An assumption that the union bargains with a product-market
monopolist or cartel gives similar results, but is a bit more complicated. As in the simplest models
of strikes as a result of asymmetric information, the union proposes terms of employment which the
employers can accept or reject (making this a “screening” model in the terms of Kennan and Wilson
[1990]), and the union can commit itself to strike for an interval of time or equivalently there is a
fixed interval between the union’s offers (for a discussion of this see Hart [1989]). The marginal
revenue product of labor in the industry depends on a fixed capital stock, the state of product
demand and the relative cost of raw materials. The capital stock is chosen by firms in expectation of
future interactions with the union. The realized state of product demand and relative raw materials’
cost is observed by the firms, but not by the workers.
Under these assumptions, I show what happens in three situations: first, if the union and the
firms “bargain” over fixed wage rates in the absence of long-term contracts; second, if they bargain
- 17 over fixed wage rates with an ability to sign a long-term contract; third, if the union can offer the
firms a sliding scale. In the absence of contracts, the union’s best fixed-wage strategy sometimes
results in strikes. The strategy gains the union member a higher wage in the event that labor’s value
is high, but the anticipation of future strikes lowers the firms’ investment in fixed capital, which
tends to reduce the union member’s wage in any event. With long-term contracts, the union can
commit not to strike in the future. That prevents the union from gaining a higher wage in the event
that labor’s value is high, but it causes the firms to choose a larger capital stock, so it may increase
the union member’s expected wage income: if it does, the union will sign a long-term contract.
Finally, if the union member can observe the product price and materials cost, the union can
offer the firm a sliding scale. The form of the sliding scale that maximizes the union member’s
expected income matches the characteristics of historical sliding scales described above. Like the
strike strategy, the sliding scale gains the union member a higher wage when labor’s value is high.
But the firms always accept the offer of a sliding scale immediately: there are no strikes, whether or
not the union can sign a long-term contract. The sliding scale is better for the union member than
fixed-wage bargaining whether or not the union can sign a long-term contract. But the advantage of
the sliding scale is greater in the absence of long-term contracts.
4.2 General assumptions and notation
The industry is made up of many identical profit-maximizing, risk-neutral firms that
produce a homogenous, nonstorable good. Time is divided into three periods. In period zero a firm
chooses its capital stock K subject to a unit cost of capital (or required expected return) R . In each
of the following two periods - period one and period two - a firm’s capital stock is fixed, while a
firm can produce output Y using materials and labor. The quantity demanded in each period, per
firm, is:
YD '
D
P
Γ
where
Γ > 1
(1)
- 18 where P is the price of the good and D is a demand variable.
Each unit of output requires one unit of a raw material. The cost of the raw material relative
to the product price is M < 1. Production also requires capital and labor, which are subject to
diminishing marginal productivity. A firm can hire workers on an open competitive market at a
going wage W per period, or from a set of workers belonging to a union. Effective labor input is
higher if the firm employs the members of the union (perhaps because they have special skills, or
because they are incumbents and it is costly to bring new workers into a firms' plants). Thus, if a
firm employs union members its production function is:
Y ' min ( X , K Θ L Λ )
where Λ % Θ < 1
(2)
where X is the number of units of raw material and L is the number of workers employed. If the firm
does not employ the union members, the production function is:
Y ' min ( X , K Θ L / Φ Λ )
where Φ > 1
(3)
In addition to costs of labor and materials, a firm is subject to a fixed operating cost Ω every period
in which production takes place.
The members of the union can coordinate their behavior, in a limited way. They cannot
enforce transfers of income between members, or prevent a member from leaving the union to take
employment on the outside competitive labor market. But they can make a common take-it-orleave-it offer of employment conditions to the firms in the industry, and commit to withdraw their
labor for a period if the firms do not accept the offer. If the firms accept the union's offer but choose
to employ fewer workers than the number of union members, union members are laid off in a
predetermined order (for example, by seniority). Any union offer to the firms must satisfy a certain
number of union members. (This number might be all of the members if union actions require a
consensus, or a majority of members under majority rule.) This number of members, divided by the
- 19 number of employers, is N. To simplify notation, we can set N equal to one by scaling all of the
other variables to N. A union member supports actions that maximize his expected wage income,
where I e denotes the expected value of wage income per period. Under these assumptions, the
union will generally act to maximize expected wage income for the marginal member of the group
required to approve an action. I refer to him as the “union member.”
The firms in the industry accept or reject the union's offers as a group, so as to maximize
expected profit for a firm. If they reject the union's offer, they can continue to negotiate with the
union and wait for another offer, or replace the union workforce with new workers hired from the
outside labor market. If the firms reject a union offer but continue to negotiate, there is a strike: for
one period the firms' plants are closed, producing no output and incurring no costs (other than the
required return to capital). The period’s potential sales are lost - they cannot be made up in the
second period. Meanwhile the union members receive no labor income. If the firms choose to
replace the union with new workers, the strike is aborted; the firms pay the new workers the going
wage W and produce according to expression (3), while union members take other jobs at the going
wage. Outside negotiation with the union, the firms are unable to collude. Thus, in periods one and
two firms will produce the quantity of output that equates price and marginal cost. In period zero
firms install capital to the point that the required return equals the expected value of the marginal
revenue product of capital.
What would happen if the union offered, and the firms accepted, a fixed wage W for all
workers? The cost of production in a period, excluding the fixed cost of capital, would be:
1
Λ
C ' Ω % M P Y % WL ' Ω % M P Y % W Y K
&
Θ
Λ
(4)
This is less than the cost of producing with a replacement workforce as long as:
W < W Φ
(5)
- 20 Taking the capital stock as given, setting the product price equal to marginal cost MC/MY determines
the quantity of output for the period as a function of the wage, the capital stock, product demand and
the relative price of raw materials. The corresponding product price, marginal revenue product of
capital (taking the product price as given), and number of workers employed are:
P(W,Z , K) '
R (W, Z , K) ' P
MY
MK
Λ&Λ K &Θ Z Γ(1&Λ) W Λ
1
Λ%Γ(1&Λ)
' Θ ( Λ / W)Λ(Γ&1) K &(Θ%Λ) & Γ(1&Θ&Λ) Z Γ
L ( W, Z , K ) '
where
ΛΓ K (Γ&1) Θ
(6)
1
Λ%Γ(1&Λ)
1
Λ%Γ(1&Λ)
Z Γ
(7)
(8)
W
Z ' D (1&M)
Γ&1
Γ
The variable Z summarizes the effects of product demand and raw materials cost. Z can take
one of two values: a high value ZH and a low value ZL . (This would be true if, for example, product
demand D takes two values, high or low, while low product demand is sometimes offset by low raw
materials prices.) Z’s value is realized and observed by the firms at the beginning of period one. It
remains unchanged through period two. As of period zero, the probability that Z will turn out to be
low is Π . This is known to both firms and workers.
What wage would the union offer the firms in period one or two, taking the capital stock as
given, if the union member knew the realized value of Z? The union will offer the highest wage that
ensures employment to the marginal union member given Z. Thus, if Z is high, the union would
offer the wage for which employment defined by expression (8) equals N (=1) for Z ' ZH , unless
that wage would make it cheaper for the firms to replace the union members with workers hired at
the outside going wage. Thus the union’s offer would be:
- 21 WH,K ' Λ K
(Γ&1) Θ
Γ
ZH
for
ΛK
(Γ&1) Θ
Γ
(9)
ZH < W Φ
WH,K' W Φ otherwise
If Z is low, the union’s offer is defined by expression (8) for Z ' ZL , unless that wage is less than
the outside going wage, in which case the union would offer the outside going wage:
WL,K ' Λ K
(Γ&1) Θ
Γ
ZL
for
ΛK
(Γ&1) Θ
Γ
ZL > W
(10)
WL,K' W otherwise
Assume, however, that the union member cannot observe the realized value of Z.
4.2 Fixed wage rates
What happens if the union’s offers to the firms consist of fixed wage rates? That depends on
whether the union can enter into a binding long-term contract.
Strikes in the absence of binding union contracts
Suppose that the union is unable to constrain its future behavior in a binding contract. Then
in period one, the union will take the capital stock as given and make the firms an offer that
maximizes expected labor income for the union member. In period two the union can make another
offer, unless the firms chose to replace the union with new workers. In the preceding period zero,
firms choose the capital stock based on rational expectations of the relation between any given
capital stock and the union’s behavior in the subsequent periods.
For a range of values of the model’s parameters, the sequence of fixed wage offers that
maximizes the union member’s expected wage income at each point in time will occasionally result
in strikes. The capital stock chosen by the firms in period zero, anticipating strikes, is KSTR . The
union’s first offer to the firms is the wage defined by expression (9) for K ' KSTR , denoted WH,K .
STR
- 22 If Z ' ZH the firms accept this offer; in the second period the union will offer the same wage and the
firms will again accept it. If Z ' ZL there is a strike: the firms reject the union’s first offer but
continue to negotiate with the union. After a strike, in the second period the union offers the firms
the lower wage WL,K defined by (10). The firms accept the second offer. In period zero, looking
STR
forward to this sequence of offers, the union member’s expected wage income is:
e
ISTR ' (1&Π) WH,K
STR
Π
%
WL,K
2
(11)
STR
while the firms choose K STR such that:
R ' (1&Π) R(WH,K , ZH , K STR) %
STR
Π
(12)
R(WL,K , ZL , K STR)
2
STR
which means:
K STR ' Λ
Λ(Γ&1)
R
Θ
&Λ&Γ(1&Λ)
(1&Π)
Γ &Λ(Γ&1)
Z H WH,K STR % Π
2
Γ &Λ(Γ&1)
Z L WL,K STR
1
Θ% Λ% Γ (1 & Θ & Λ)
(13)
For this sequence to prevail, it must be true that the firms reject the union’s first offer when
demand is low, but accept it when demand is high. This holds as long as the fixed operating cost Ω
falls within a certain range relative to the other parameters in the model. It must also be true that the
union member is willing to run the danger of a strike, rather than immediately offer a wage low
enough to ensure his employment in the event that demand is low. This holds as long as the possible
loss of one period’s low wage is outweighed by the chance to gain a higher wage in the event that
labor demand is high, which is to say if:
(1&Π) (WH,K & WL,K ) >
STR
STR
Π
2
WL,K
(14)
STR
Assuming the wage offers are within the bounds determined by W and WΦ (a further
condition on the model’s parameters), the capital stock is:
- 23 KSTR '
Θ
R
Π
(1&Π) ZH %
ZL
2
Γ
Γ& ( Γ & 1 ) Θ
(15)
The union member’s expected income is:
(Γ&1)Θ
e
ISTR ' ΛK STRΓ ( 1 & Π ) ZH%
Π
2
(Γ&1)Θ
Γ&(Γ&1)Θ
Θ
ZL ' Λ
R
Π
(1&Π) Z H %
2
ZL
Γ
Γ&(Γ&1)Θ
(16)
The union member is willing to run the danger of a strike ( expression 14 holds) as long as:
(1 & Π)
ZH
%
ZL
Π
2
(17)
> 1
Note that the size of the capital stock is positively related to the union member’s expected income,
but it does not affect the union’s decision to strike.
Long-term fixed-wage contracts
Now suppose that the union can sign a long-term contract with the firms. A contract prevents
the union from striking within the term of the contract if the employers pay the specified wage. The
union will sign a contract if the union member’s expected wage income is higher under the contract
than if the union were left free to strike. This condition may be satisfied by a contract that the union
signs in period zero, setting a wage for periods one and two equal to WL , K
CON
- that is, the wage
determined by expression (10) for the capital stock firms choose subject to the contract. Given this
wage and capital stock, the union member would be employed in periods one and two whether Z
turns out to be high or low, so his expected wage income under the contract is simply WL,K
CON
. This
exceeds expected wage income without a contract (expression 11) if:
(1&Π) (WH,K & WL,K
STR
STR
) <
Π
2
WL,K
STR
% (WL,K
CON
&WL,K )
(18)
STR
Expression (18) can hold along with (14) because the firms will choose a larger capital stock if the
- 24 union signs the long-term contract:
K CON ' Λ
Λ(Γ&1)
R
&Λ&Γ(1&Λ)
Θ
Γ &Λ(Γ&1)
(1&Π) Z H WL,K CON
Γ &Λ(Γ&1)
Π Z L WL,K CON
%
1
Θ% Λ% Γ (1 & Θ & Λ)
(19)
Assuming the wage offers are within the bounds determined by W and WΦ :
Θ
KCON'
R
ZH
(1&Π) ZH % Π Z L
(Γ&1)Λ
Λ%Γ(1&Λ)
Γ
Γ& (Γ&1)Θ
(20)
ZL
and (18) holds if:
(1 & Π)
ZH
%
ZL
KCON / KSTR '
where
Π
2
K CON
<
(Γ&1)Θ
Γ&Θ(Γ&1)
(21)
K STR
(1&Π) Z H % Π Z L
(1&Π)Z H %
(Γ&1)Λ
Z H Λ%Γ(1&Λ)
ZL
Π
2
Γ
Γ&Θ(Γ&1)
ZL
4.2 Sliding scales
If union members can observe the product price and the raw materials cost, the union can
offer the firm a sliding scale. The sliding scale that the union would offer as a long-term contract in
period zero is the same as the scale it would offer in the absence of a contract in period one and
again in period two. It sets the sliding-scale wage WSS as a function of the product price and
materials cost:
Θ
WSS ' Λ KSS (P & P M )
for
Θ
W # Λ KSS (P & P M) # WΦ
Θ
WSS ' W
for
Λ K SS (P & PM) < W
WSS ' W Φ
for
Λ K SS (P & PM) > W Φ
Θ
(22)
- 25 where KSS is the capital stock the firms chose in period zero. Note that this function matches the
form of actual sliding scales: the wage is scaled to the manufacturer’s margin - the product price less
materials cost - with a minimum (equal to the outside going wage) and a maximum (determined by
the outside wage and the union workers’ productivity advantage). The wage and product price that
would result from this scale are determined by (22) together with expression (6) above. This gives:
WSS ' WH,K
SS
if Z ' ZH
WSS ' WL,K
SS
if Z ' ZL
(23)
In the event that labor demand is high, the sliding scale gives the union the same wage it would have
received in the first period in the absence of a long-term contract, given a capital stock equal to KSS
. If labor demand is low, the sliding scale gives the same wage the union would have received
following a strike. But there will never be a strike under the sliding scale, even in the absence of a
contract. The firms will accept the sliding scale offer in the first period whether labor demand is
high or low. Thus, as of period zero the union member’s expected labor income is:
e
ISS ' (1&Π) WH,K % Π WL,K
SS
(24)
SS
while the capital stock satisfies:
K SS ' ΛΛ(Γ&1)
R
&Λ&Γ(1&Λ)
Θ
Γ &Λ(Γ&1)
(1&Π) Z H WH,K SS
%
Γ &Λ(Γ&1)
Π Z L WL,K SS
1
Θ% Λ% Γ (1 & Θ & Λ)
(25)
The union member’s expected wage income is higher under the sliding scale than under either the
fixed wage contract or fixed wages and strikes in the absence of contracts. But the advantage of the
sliding scale must be greater in the absence of contracts, for a union that would otherwise be willing
to sign one. In the absence of contracts, the sliding scale promises the union member a gain in
expected wage income equal to:
- 26 e
e
ISS & ISTR' (1&Π) (WH,K & WH,K ) %
SS
STR
Π
2
WL,K % ( WL,K & WL,K )
SS
SS
(26)
STR
Assuming that the wage offers are within the bounds given by W and WΦ :
Θ
KSS '
R
(1&Π) ZH % Π Z L
Γ
Γ& ( Γ & 1 ) Θ
(27)
Γ
Γ&(Γ&1)Θ
(28)
while:
e
ISS ' Λ
Θ
(Γ&1)Θ
Γ&(Γ&1)Θ
R
(1&Π) Z H % Π ZL
and:
e
ISS
&
e
ISTR'
Π WL,K % ( (K SS / KSTR)
(Γ&1)Θ
Γ
SS
where K SS / K STR '
(1&Π) Z H % Π Z L
(1&Π) Z H %
Π
2
e
& 1 ) ISTR
(29)
Γ
Γ&Θ(Γ&1)
ZL
Expression (29) shows that, in the absence of long-term contracts, there are two reasons the union
member’s expected wage income is higher under the sliding scale. First, the sliding scale allows the
union member to receive a wage in the first period, rather than nothing, in the event that labor
demand is low. Second, it boosts the capital stock chosen by the firms, which increases the wage the
union member receives in any event. Both of these reasons can be viewed as consequences of the
elimination of strikes. Thus, it would make sense to describe the elimination of strikes as the
advantage of the sliding scale.
5. Why did sliding scales disappear after the 1930s?
After the 1930s, institutional developments made sliding scales less useful in the set of
industries where they were feasible. In Britain, the mining and metals industries came under
- 27 government controls that broke the connection between product demand, labor demand and the
wages that unions could extract from employers. The beginning of the Second World War brought
controls of product prices, output and employment. Coal mines were nationalized in 1947. In the
1950's, it was observed that the board regulating the industry “clearly does not aim to maximize
profit” (Alexander, 1956, p. 165), as “Nobody disputes that the existing output of coal could be sold
at a higher price” (p. 174), while the miners’ union maintained a “voluntary renunciation of the
strike weapon” (p. 164) because of “the importance attached by influential leaders of the union
to..the stability of political institutions in general and of the public corporation type of
nationalization of coal-mining in particular” (p. 166). The iron and steel industry was under private
ownership over most of the 1950's, following an aborted nationalization at the beginning of the
decade, but “As maximum prices were set by the Iron and Steel board for most categories of steel
products the producers were unable to raise their prices” (Blair, 1997, p. 573). After the industry
was nationalized again in the 1960's it was “a participant in various counter-inflationary
programmes as steel prices were subject to restraint by the government..not able to operate freely in
the market..under political pressure not to declare redundancies [layoffs]..refused permission to
increase its selling prices ” (p. 575, 576). By the 1970s firms in both the metals and mining
industries were losing money, with workforces far greater than any profit-maximizing enterprise
would choose to employ (Pryke, 1981, pp. 48, 59, 183-188). Over the 1980s the subsidies were
stopped and the unionized industries ceased to exist (Fraser, 1999, pp.235-36, 241-242).
In the United States, meanwhile, unions gained the ability to enter long-term binding
contracts with employers as a result of the Wagner Act of 1935, the Taft-Hartley Act of 1947 and
related court decisions. (In Britain, union wage agreements never became binding contracts in this
way [Shackleton, 1998]). Workers that strike in violation of a contract lose protections afforded by
the National Labor Relations Board system (against lockouts, for example), while a union
organization that orders such a strike can be sued for damages (Mills and Brown, 1950, pp. 470-
- 28 513). By the 1940s “the underlying goal of bargaining strategy..was to minimize industrial strife by
negotiating long-term contracts...a multi-year contract without periodic reopenings” (Garbarino,
1962, p. 10). Long-term contracts came to prevail not only in the unionized mining and metals
industries, but in almost all industries subject to strike threats. According to the model presented
above, the advantage of a sliding scale over fixed wage rates is not as great within a long-term
contract as in the absence of long-term contracts. Thus, the spread of long-term contracts reduced
unions’ incentives to bargain over sliding scales rather than fixed wage rates.
6. Conclusion
The sliding scales of the ninteenth and early-twentieth century were not adopted in order to
stabilize firms’ profits or maintain efficient employment levels within long-term employment
contracts. They were instead an alternative to strikes as a mechanism for dealing with employers’
private information about product demand in an environment where unions were unable to enter
into long-term contracts. Sliding scales were usually unsuccessful outside metals and mining
because it was hard for workers to observe employers’ product prices and materials costs, even in
industries where outputs and inputs were homogenous goods with observable open-market prices.
Materials cost to the employer could differ from spot prices because employers held stocks.
Employers could be producing for long-term sales contracts rather than spot markets. Union
members might distrust their representatives’ interpretation of employers’ price information that
could not be revealed to, or understood by, the union rank and file.
The problems that prevented the use of sliding scales in most industries before the 1930s
were hardly alleviated by the increase in some kinds of information about product and materials’
prices. Long-term sales contracts that set prices, but not quantities are still common in goodsproducing industries (Blinder et. al., 1998, p. 92-95). It remains difficult to determine the cost to a
producer of materials held in inventory, subject to transport and storage costs and uncertainty about
future spot market prices. Union members still distrust their leadership, and economic theory
- 29 suggests they are right to do so (Ashenfelter and Johnson, 1969; Cai, 2000). Meanwhile, in Britain,
nationalization and other forms of government control made product demand irrelevant for union
bargaining in mining and metals. In the U.S., unions gained the ability to enter long-term
employment contracts, which reduced the potential advantages of sliding scales over fixed wage
rates.
Bibliography
American Iron and Steel Association Bulletin, Volume 22, 1888.
Alexander, K.J.W. “Wages in Coal-mining since Nationalization.” Oxford Economic Papers, June
1956, 8, 164-180.
Asheley, W.J. The Adjustment of Wages: A Study of the Coal and Iron Industries of Great Britain
and North America. London: Longmans, Green 1903.
Ashenfelter, Orley and George E. Johnson, “Bargaining Theory, Trade Unions, and Industrial Strike
Activity.” American Economic Review, March 1969, 59, 35-49.
Azariadis, Costas and Joseph E. Stiglitz, "Implicit Contracts and Fixed-Price Equilibria," Quarterly
Journal of Economics, 1983 Supplement, 98, 1-22.
Bagnell, Paul. “Inco’s Sudbury Workers Approve Pact.” Toronto Financial Post, July 1 1997, p. 4
Blair, Alasdair M. “The British Iron and Steel Industry Since 1945.” Journal of European Economic
History, Winter 1997, 26, 571-581.
Blair, Douglas H. and David L. Crawford. “Labor Union Objectives and Collective Bargaining.”
Quarterly Journal of Economics, August 1984, 99, 547-566.
Blanchard, Olivier Jean. “Wage Indexing Rules and the Behavior of the Economy.” Journal of
Political Economy, 1979, 87(4), 798-815.
Blinder, Alan S., Elie R.D. Canetti, David E. Lebow and Jeremy B. Rudd. Asking About Prices: A
New Approach to Understanding Price Stickiness. New York: Russell Sage Foundation, 1998.
Bowie, J.A., "A New Method of Wage Adjustment in the Light of the Recent History of Wage
Methods in the British Coal Industry," Economic Journal, September 1927, 37, 384-394.
Burn, Duncan. The Steel Industry 1939-1959, A Study in Competition and Planning. Cambridge,
- 30 UK: Cambridge University Press, 1961.
Cai, Hongbin. “Bargaining on Behalf of a Constituency.” Journal of Economic Theory, June 2000,
92(2), 234-273.
Carnegie, Andrew. The Gospel of Wealth and Other Timely Essays. Cambridge: Harvard University
Press, 1962, reprint of 1900.
Chapman, S.J., "Some Theoretical Objections to Sliding Scales," Economic Journal, June 1903, 13,
186-196.
Charlier, Marj. “Magma Copper’s Workers Approve Pact the Protects Against Strikes for 7 Years.”
Wall Street Journal, October 23 1991, p. A2:3.
-- “Magma Copper Heals Its Workplace and Bottom Line.” Wall Street Journal, April 6 1992, B4:3.
Daugherty, Carroll R, Melvin D. de Chazeau and Sameul S. Stratton. The Economics of the Iron and
Steel Industry, Volume I. New York: McGraw-Hill, 1937.
Davis, Pearce. The Development of the American Glass Industry (Harvard Economic Studies,
volume 86). Cambridge: Harvard University Press, 1949.
Farber, Henry. “The Analysis of Union Behavior.” In Orley Ashenfelter and Richard Layer, eds.
Handbook of Labor Economics, Volume II. Amsterdam: North-Holland, 1986.
Fisher, Waldo E., "Anthracite," in Harry A. Millis ed., How Collective Bargaining Works, New
York: Twentieth Century Fund, 1942.
Fraser, W. Hamish. A History of British Trade Unionism, 1700-1998. London: MacMillan, 1999.
Freeman, Richard. "Spurts in Union Growth: Defining Moments and Social Processes." In
Michael D. Bordo, Claudia Goldin, and Eugene N. White, eds., The Defining Moment: The Great
Depression and the American Economy in the Twentieth Century. Chicago: University of Chicago
Press, 1998. Pp. 265-95.
Garbarino, Joseph W. Wage Policy and Long-term Contracts. Washington, D.C.: The Brookings
Institution, 1962.
Great Britain, Ministry of Labour, Ministry of Labour Gazette, "Adjustment of Wages by
Sliding-Scale Arrangement," August 1925: 269-270.
-- Royal Commission on Labour. Minutes of Evidence, Group A, Volumes I and II. London: HMSO
1892.
--------- Fifth and Final Report, Part I, The Report. London: HMSO, 1894.
- 31 -- Report of the Royal Commission on Trade Unions and Employer’s Associations. Chairman Lord
Donovan, London: HMSO 1968.
Greenfield, Henry I., Sliding Wage Scales, Columbia University Doctoral Dissertation, 1960.
Greenwald, Bruce C. and Stiglitz, Joseph E., "Asymmetric Information and the New Theory of the
Firm: Financial Constraints and Risk Behavior," American Economic Review, 80, May 1990,
160-165.
Grossman, Gene M. “Union Wages, Temporary Layoffs, and Seniority.” American Economic
Review, June 1983, 73, 277-290.
--, Oliver Hart and Eric Maskin. “Unemployment with Observable Aggregate Shocks.” Journal of
Political Economy, December 1983, 91, 907-928.
Grout, Paul A. “Investment and Wages in the Absence of Binding Contracts: A Nash Bargaining
Approach.” Econometrica, March 1984, 52(2), 449-460.
Hall, Robert E. and Lazear, Edward P., "The Excess Sensitivity of Layoffs and Quits to Demand,"
Journal of Labor Economics, 1984, 2, 233-257.
Hart, Oliver. “Optimal Labour Contracts under Asymmetric Information: An Introduction.” Review
of Economic Studies, 1983, 50, 3-35.
-- “Bargaining and Strikes.” Quarterly Journal of Economics, February 1989, 25-43.
Hanes, Christopher. "Prices and Price Indexes: Introduction," in Historical Statistics of the United
States, Millenium edition. Cambridge University Press, forthcoming.
Haynes, William Warren. Nationalization in Practice: The British Coal Industry. Boston: Harvard
University Graduate School of Business, 1953.
Hart, Oliver. "Optimal Labour Contracts under Asymmetric Information: An Introduction," Review
of Economic Studies, 1983, 50, 3-35.
Hayes, Beth. “Unions and Strikes with Asymmetric Information.” Journal of Labor Economics,
January 1984, 2, 57-83.
Howard, Stanley, The Movement of Wages in the Cotton Manufacturing Industry of New England
since 1860, Boston: National Council of American Cotton Manufacturers, 1920.
Jeans, J. Stephen. American Industrial Conditions and Competition: Reports of the Commissioners
appointed by the British Iron Trade Association to enquire into the Iron, Steel and Allied Industries
of the United States. London: British Iron Trade Association, 1902.
- 32 Jevons, H. Stanley. The British Coal Trade. London: Kegan Paul, Trench, Trubner and Company,
1915.
Knowles, K.G.J.C. Strikes: A Study in Industrial Conflict. New York: Philosophical Library, 1952.
Marshall, Alfred and Mary Paley Marshall, The Economics of Industry, London: MacMillan and
Co., 1881.
Massachusetts Bureau of Statistics of Labor, Twelfth Annual Report, 1881, Boston: Reed, Avery
and Company, 1881.
Massachusetts Bureau of Statistics, Thirty-Ninth Annual Report on the Statistics of Labor for the
Year 1908, Boston: Wright and Poter, 1909.
Matthews, Derek. “Profit-Sharing in the Gas Industry, 1889-1949.” Business History, July 1988,
30(3), 306-328.
Millis, Harry A. and Emily Clark Brown. From the Wagner Act to Taft-Hartley. Chicago: University
of Chicago Press, 1950.
Morris, J.H. and L. J. Williams. “The South Wales Sliding Scale 1876-1879: An Experiment in
Industrial Relation.” Manchester School of Economic and Social Studies, May 1960, 28(2), 161176.
Munro, J.E.C., "Sliding Scales in the Iron Industry," Transactions of the Manchester Statistical
Society, Sessions 1885-1886, 1-43.
-- "Sliding Scales in the Coal Industry," Transactions of the Manchester Statistical Society, Sessions
1889-1890, 119-171.
Palgrave, R. H. Inglis. “Sliding scale (wages).” Dictionary of Political Economy, volume II, 410411. London: MacMillan and Co., 1896.
Pigou, Arthur C., Principles and Methods of Industrial Peace, London: MacMillan and Company,
1905.
Pool, Arthur George. Wage Policy in Relation to Industrial Fluctuations, London: MacMillan and
Company, 1938.
Porter, "Wage Determination by Selling Price Sliding Scales 1870-1914," The Manchester School
of Economic and Social Studies, March 1971, 39, no. 1, 13-21.
Price, L. L. Industrial Peace: Its Advantages, Methods and Difficulties, London: MacMillan, 1887.
Pryke, Richard. The Nationalised Industries: Policies and Performance since 1968. Oxford: Martin
- 33 Robertson, 1981.
Robinson, Jesse S., "The Amalgamated Association of Iron, Steel and Tin Workers," Johns Hopkins
University Studies in Historical and Political Science, Series 38 No 2, 138-162. Baltimore: Johns
Hopkins Press, 1920.
Shackleton, J.R. “Industrial Relations Reform in Britain since 1979.” Journal of Labor Research,
Summer 1998, 19(3), 581-605.
Smart, William, Studies in Economics, London: MacMillan, 1895.
Smith, Russell A. Labor Law: Cases and Materials. Indianapolis: Bobbs-Merrill, 1950.
Steinfeld, Robert J. The Invention of Free Labor: The Employment Relation in English and
American Law and Culture, 1350-1870. Chapel Hill: University of North Carolina Press, 1991.
Treble, John G., "Sliding Scales and Conciliation Boards: Risk-Sharing in the Late 19th Century
British Coal Industry," Oxford Economic Papers 39, 1987, 679-698.
United States Bureau of Labor, Eleventh Special Report of the Commissioner: Regulation and
Restriction of Output, Washington: GPO 1904.
United States Bureau of Labor Statistics, "Wage-Adjustment Provisions in Union Agreements,"
Monthly Labor Review, January 1940, 50, 6-15.
-- Wage Structure of the Nonferrous Metals Industry, 1941-1942, Bulletin 729, 1943.
-- “Escalation and Producer Price Indexes: A Guide for Contracting Parties.” Report 807, September
1991.
United States Industrial Commission, Report of the Industrial Commission on the Relations and
Conditions of Capital and Labor, Volume 7, Washington: GPO 1901.
Virtue, George O. “The Anthracite Miners’ Strike of 1900.” Journal of Political Economy,
December 1900, 9 (1), 1-23.
Martin L. Weitzman, The Share Economy. Cambridge: Harvard University Press, 1984.