Jump to ContentJump to Main Navigation
Managing Intellectual CapitalOrganizational, Strategic, and Policy Dimensions$

David J. Teece

Print publication date: 2002

Print ISBN-13: 9780198295426

Published to Oxford Scholarship Online: November 2003

DOI: 10.1093/0198295421.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2017. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 24 February 2017

(p.225) Appendix B: The Glass Industry and the Pilkington Float Process*

(p.225) Appendix B: The Glass Industry and the Pilkington Float Process*

Source:
Managing Intellectual Capital
Publisher:
Oxford University Press

B.1 Introduction

Pilkington's float glass process is one of the great process innovations of this century. It transformed the industry and caused prices to fall dramatically. Since its development and introduction in 1959, the float process has almost totally displaced other methods of manufacturing flat glass. It completely replaced the plate process and almost completely replaced the sheet process. World‐wide glass production has increased many fold since 1959, due in large measure to the massive reduction in the cost of glass brought about by the almost‐universal adoption of the float process. This adoption was enabled by a worldwide licensing programme engineered by Pilkington.

Pilkington's innovation provided large returns to the many licensees of the float process. The float process also benefited consumers and producers. Consumers have gained from the significantly lower prices and consistently high quality of flat glass produced by the float process. The consumer benefits flowing from lower prices, high quality and increased output, and the private benefits to the licensees, dwarf the returns obtained by Pilkington itself. Calculations summarized in Table B.1 and explained in section B.6.1, show that Pilkington's private benefits from its invention of the float process were only 5 per cent of the total social benefits (the benefits to society as a whole) attributable to its innovation. We estimate that Pilkington realized $5.3 billion (in 1992 dollars) from its float innovation, while social benefits have totalled between $105.4 billion and $117.2 billion. A review of the economics literature reveals that, in comparison with other innovators, Pilkington has captured a significantly smaller fraction of the total social benefits than is typically the case.

Float technology consists of both patentable and non‐patentable elements. Patents (now largely expired) covered specific aspects of the process, but there was and still is a large volume of know‐how in the design and operation of the (p.226) process which can only be protected by trade secret. As discussed in Chapter 1, there are a number of substantial differences between patents and trade secrets. In particular, patents are granted for a definite period, and protect the innovator against use of the patented product or process, even by someone who has independently developed the technology. Trade secrets on the other hand, have an indefinite and potentially infinite life, but do not protect the innovator against subsequent replication of the innovation by a third party.

B.2 Pilkington's Licensing Practices: Motivation and Effect

When it invented the float process, Pilkington was a relatively small UK‐based glass maker, without the balance sheet or experience needed to introduce a major innovation world‐wide on its own. Pilkington believed there would be more profit in selling glass than in just selling the float technology; to the extent feasible, it therefore wished to expand its own production in the UK and world‐wide employing its innovation. It did not have the financial and managerial capabilities to expand in this way immediately, having had its cash drained by the long, expensive development of the float process. Moreover, converting its UK facilities to the new process also would involve the investment of large sums of money. As a practical matter, Pilkington itself did not have the resources, in terms of financial and managerial capabilities, to commercialize its float innovation on a world‐wide basis using direct investment.1

Reflecting these constraints, Pilkington's strategy for commercializing its float technology evolved over time. In the initial stage, Pilkington chose to license its invention. To commercialize the process rapidly and broadly, Pilkington needed the support of the existing plate producers, as they had the necessary manufacturing, marketing, and distribution capabilities to rapidly commercialize float. This was in part because the float process replaced only a part of the overall glass production process—namely, the grinding and polishing stages associated with plate glass production. Existing manufacturers had the manufacturing capacity and skills to implement float production successfully.

Pilkington's decision to grant licences to existing plate producers created competition that would not otherwise have existed. It benefited consumers by making the technology widely available rather quickly, and this in turn led to prices falling rapidly and significantly. In short, there were advantages which flowed to Pilkington and to consumers because Pilkington chose to license (p.227) industry incumbents. Without access to Pilkington's innovation, incumbents would have had to compete using inefficient and obsolete plate technology.

Pilkington granted its first licence in 1962, to an American firm, PPG, after lengthy negotiations over the terms under which PPG would gain access to this revolutionary technology. By 1970, just eight years after this licence, every plate glass producer in the world had obtained a licence to practise Pilkington's technology.2 Indeed, no new plate plants were built anywhere in the world after 1962. Later, after float glass manufacturing costs had fallen sufficiently, glass produced by the float process became economically viable as a replacement for glass produced by the sheet process, thereby making glass of the highest quality available for the same cost as the lower‐quality glass produced by the sheet process.

Pilkington naturally granted geographic rights to existing plate glass producers to use its float technology in their home markets, and they generally received the right to sell the resulting glass almost anywhere. Pilkington kept its options open as to how float glass technology would be commercialized in other markets which did not yet have established plate producers.

The markets that were easiest for Pilkington to enter de novo were those that did not yet have a plate glass manufacturing facility. In many of these regions Pilkington had a longstanding market presence as a producer of sheet glass. In early years, the glass market in many of these regions had not yet reached a size where a float plant could be justified. Pilkington chose not to grant licences to producers in these regions immediately, anticipating that, when demand grew to a level which would support float glass facilities, Pilkington itself would establish those facilities (often via local affiliates, subsidiaries or joint ventures). In many of these regions, Pilkington has subsequently granted licences to local sheet manufacturers.

In summary, Pilkington first licensed to existing manufacturers, and later expanded its own production capabilities via affiliates where appropriate, while postponing direct foreign investment until it had sufficient funds to do so. It even granted certain licensees the right to sublicense its technology in certain territories as opportunities and business conditions required and allowed. As a result of this phased approach, float technology has been introduced in different world markets in a variety of ways according to circumstances, including direct entry by Pilkington (either on its own or via joint ventures), licences to other manufacturers, and sublicences granted by Pilkington licensees.

B.3 Licensing Terms: Provisions and Rationales

Pilkington licensed its technology by granting affirmative rights for the non‐exclusive use of its float technology in specific territories. Licensing by territory (p.228) is extremely common in international business. Indeed, it is quite normal for licensors to grant rights to use the licensor's intellectual property within specified territories or specified technical ‘fields of use’ (or both).3 Moreover, territorial licensing was in line with traditional practice in the glass industry. Both the sheet glass process and the twin grinder plate glass process which had preceded float were licensed on a territorial basis.4

Without the ability to grant licences only for specific territories and thereby to preserve to itself the use of its technology in certain other territories, Pilkington's best strategy would probably have been not to license at all, but to use the UK as a base to gradually introduce float through Pilkington‐owned and operated facilities throughout the world. The ability of an innovator such as Pilkington to license to others less than the full panoply of intellectual property rights associated with its innovation is widely recognized as pro‐competitive, since it facilitates wider and more rapid dissemination of the innovation than would otherwise occur. Territorial and field‐of‐use limitations thus promote the licensor's incentive to license by providing limited protection against cannibalization from licensees using the licensor's own technology.

Pilkington's licences also provided for the exchange of improvements to the process among all users of the new technology, including Pilkington. This exchange of improvements assured licensees that they would not be stranded with an obsolete version of the technology in the event of rapid improvements. And, just as importantly, it assured Pilkington that it was not empowering a competitor to use Pilkington's technology to leave it stranded, an assurance that was crucial to Pilkington's willingness to share its technology with others.

Another significant aspect of Pilkington's licensing terms was the form of the intellectual property rights that protect different aspects of the technology. Float technology is a combination of patentable and non‐patentable intellectual property. (See Fig. B.1.) Patents covered specific aspects of the process, but there was and is a large volume of know‐how in the design and operation of the process which could only be protected by trade secret. Thus the licences covered both types of intellectual property required to operate a float plant. They included provisions for maintaining the confidentiality of Pilkington's trade secrets, and procedures for resolving situations which might arise if the patents were no longer used by licensees. Licensing terms that remain operative rest on trade secrets that have remained confidential after the expiration of the patents.5 (p.229)

Appendix B: The Glass Industry and the Pilkington Float Process*

Fig. B.1. Float glass technology: changes in structure of proprietary knowledge over time

In short, Pilkington's licensing terms reflect the common concerns of innovators endeavouring to capture a portion of the benefit from innovation. The terms of the licence grants are ‘restrictive’, but only in the limited sense that licensees were not granted unconditional world‐wide manufacturing and sub‐licensing rights. Such grants are an exception and not the norm in technology licensing. It typically makes little sense for an innovator to license its technology carte blanche.

B.5 Diffusion of Float Glass Technology

Float glass technology has diffused very rapidly compared with other process innovations, whether in the glass industry or in other industries. By all benchmarks—the lag between invention and first outside adoption, the rate of adoption throughout the plate glass industry, and the rate of displacement of existing capacity by the new technology—the adoption of the float process was rapid. Float has also demonstrated its potential for continuing technological improvement to open up new areas for diffusion. The growth in the use of float technology continued after its adoption by the plate glass industry, as it (p.230) replaced the low cost sheet process, and as new markets for glass were created. The rapid diffusion of float was one of the major reasons why consumers and licensees reaped such large benefits from Pilkington's innovation.

The first stage of the diffusion of the float process was the replacement of plate glass. Diffusion speed may be measured in three ways: by the adoption lag between invention and the first outside adoption, the time for a proportion of potential adopters to adopt, and the rate of displacement of existing capacity. The diffusion of float technology throughout the plate glass industry was at the time one of the most rapid and complete cases of diffusion of a major innovation in industrial history by each of these measures. First, plate firms made their decisions to adopt float within a very short time after its announcement. The first licence was made within three years of the announcement of the innovation, and all ten existing plate manufacturers world‐wide had taken out licences by 1964, within only two years of the first licence. They did not, as is typical of many innovations, wait to observe the experience of the first adopters before deciding to adopt themselves. The adoption lag was very short, since most of the three years between the announcement of float in 1959 and the first licence in 1962 was taken up negotiating the licence terms. This compares with innovation lags for other major innovations which have been anything from a few years to centuries.6

Second, the diffusion rate, measured as the time for an innovation to be adopted by a given percentage of the potential users, was equally fast. Float glass adoption reached 80 per cent of the plate glass manufacturers within one year of the first licence, and 100 per cent within two years. This compares remarkably well with other studies of diffusion of process innovations in the literature, in which gradual diffusion over periods of decades is the norm (Mansfield 1968; Davies 1979).7 Also, it is rare to reach 100 per cent penetration—often an innovation becomes obsolete before it is fully diffused. Float also diffused much quickly than previous process innovations in the glass industry (Anderson and Tushman 1990).8

(p.231) Third, the float process displaced existing plate glass capacity within a very short period. The cost savings were so great that the combined capital and variable costs of installing and running a float line were lower than the variable costs of continuing to run an existing plate glass line, indicating that immediate adoption was worthwhile. Manufacturers replaced their plate plants almost as quickly as they could arrange to do so while keeping up continuous production, and well before the end of the planned lives of the plate plants (Skeddle, 1977: 310).9 Float capacity increased from about 1 per cent of world plate capacity in 1962 to about 80 per cent by 1976 (i.e. 14 years). This is faster than most comparable innovations (Ray 1974; 1984; Lynn 1982; Sahal 1981; Derclaye 1982).10

The second diffusion stage was the more gradual, though equally impressive, replacement of sheet glass, as the cost of float glass was gradually lowered by further technological improvements. Continued improvements in float technology resulted in continued commercial demand for the technology by making float glass cost‐competitive with sheet glass. This opened up new markets for float by making it possible for smaller or less prosperous countries, which previously could not justify a plate glass plant and had to import plate glass, to afford a float plant. From 1965 the expansion in the use of float glass came first from markets which previously had no plate plants. Then, after 1970, once float had become cost‐competitive with sheet, float licences were offered to major sheet glass producers, and float began to displace sheet production, a trend which accelerated rapidly after about 1974. The continued diffusion of the technology and the continuing demand for new float plants reflect the longstanding, gradual and continuing process of improvement, combined with the expansion of world markets. Fig. B.2 shows the actual number of licences and float plants world‐wide.

B.6 Quantifying Social and Private Benefits from Licensing Float Glass

The social benefits from float technology have been substantial, whether measured by the reduction in manufacturing costs for flat glass, or by the reduction in glass prices to consumers. This not only indicates that the innovation was of great importance, but also that competition has been sufficiently (p.232) great to cause the bulk of these benefits to be passed on to the consumer in terms of lower prices.

Appendix B: The Glass Industry and the Pilkington Float Process*

Fig. B.2. Number of float plants and licensees world‐wide

To illustrate these points we estimate the total social benefits, both from the point of view of cost savings and consumer surplus. We estimate the private benefits retained by Pilkington as a share of total social benefits, and calculate a social rate of return to the innovation from the total development costs of float. These are close order of magnitude estimates, rather than precise figures, intended to illustrate the overall scale of the benefits, and make some general points as to the impact of the float process and the benefits it has had for society as a whole.

B.6.1 Cost Surplus

Production Cost Saving

The net cost surplus due to the float process is the reduction in total flat production costs attributable to the float process, relative to the savings that would have occurred if the float process had not been developed. The surplus (p.233) is calculated in two steps. First, the savings achieved by float are calculated for the period 1962–92, compared with a ‘baseline’ case of no change (i.e. static technologies). Second, the projected savings are calculated for the ‘but for’ case, as the cost reductions that would reasonably have been achieved by improvements in existing plate and sheet technologies (or other new technologies), again compared with the ‘baseline’ case. The net cost surplus is the difference between these two figures (see Table B.2).

To calculate the savings attributable to the float process, one needs the annual production cost savings per ton of glass, for each year during 1962–92, compared with static ‘baseline’ technologies (i.e. 1962 plate technology and 1974 sheet technology).11 The cost trajectory is estimated from the relative behaviour of unit production costs using the float process compared with costs using the plate process, together with actual factory costs of production of float glass. Total float production costs in 1962, when the float process was introduced, were about 59 per cent of plate glass costs.12 Float glass costs fell as the process was improved, and it was shown that costs could be brought below those of sheet by about 1970. The float process began replacing the sheet process on a large scale in about 1974 (see US data below), so it is assumed that average float costs had fallen below average sheet costs by then.13 Since sheet glass prices themselves were between 50 per cent and 25 per cent of plate glass prices (see US data below), we assume that float manufacturing costs fell from 59 per cent to 25 per cent of plate 1962 costs, in real terms, over the period 1962–74. Since then they have continued to fall. It is assumed that float production costs have fallen during 1974–92 at an average rate of productivity change in the manufacturing sector of the economy as a whole. This productivity change is estimated using US data for productivity improvements in output per hour for the business sector.14 This gives a relative cost trajectory for float glass costs for 1962–92, compared with 1962 plate glass costs. To summarize, float glass costs are thus assumed to fall at a constant proportional rate per (p.234) year from 59 per cent in 1962, to 25 per cent in 1974 of 1962 plate glass costs, and then to fall from 1974 to 1992 at the same proportional rate as overall US productivity.

Actual manufacturing costs of world sales for 1962–92 are calculated using the relative cost trajectory by projecting costs back for 1961–91 from actual 1992 costs. Actual 1992 costs are taken from Pilkington's average factory cost of production per ton for 1992, consisting of direct manufacturing costs plus depreciation. (See Table B.3).

Before calculating the cost surplus, we also need to account for the fact that some float glass is used as a replacement for plate glass and some for sheet. The cost savings calculation is disaggregated into two components: ‘plate‐equivalent’ glass which would otherwise have been produced by the plate process, and ‘sheet‐equivalent’ glass which would otherwise have been produced by the sheet process.15 Thus for years after 1974, float production is divided between plate‐ and sheet‐equivalent glass, and the cost savings relative to plate and sheet are calculated separately. All float glass is taken as ‘plate‐equivalent’ until 1974, after which a proportion was ‘sheet‐equivalent’. A transition period is assumed between 1974–82, as the proportion of plate‐equivalent glass in total flat glass output is assumed to decrease linearly from 100 per cent in 1974, to an equilibrium proportion by 1982. The ‘steady state’ proportion from 1982 on is taken as the same proportion of plate output before the introduction of float glass, using US shipment data, which was 35 per cent plate.16

The surplus is the net cost difference between the actual (estimated) costs and the static ‘baseline’ technology costs, multiplied by the world float capacity for each year, separated into plate‐ and sheet‐equivalent components. The cost savings for plate‐equivalent float glass, compared with 1962 vintage plate glass costs, are multiplied by the capacity of plate‐equivalent glass each year. 1962 vintage plate costs are calculated as (1/0.593) of estimated float costs in 1962 (from the costs trajectory). Similarly, cost savings for sheet‐equivalent float glass, relative to 1974 vintage sheet glass production costs, are multiplied by the capacity of sheet‐equivalent production each year. 1974 vintage sheet glass (p.235) costs are assumed equal to estimated float costs in 1974. The two surplus figures are summed to give the total surplus for attributable to the float process, compared with the baseline.

The second step is to estimate the cost savings in a ‘but for’ case, due to likely improvements in the existing technologies. It is assumed that no other radical innovation in glass manufacture comparable to float would have been developed in the period 1962–92, and that improvements in both plate and sheet technologies would have occurred at the rate of improvements in the productivity of the manufacturing sector as a whole. The cost surplus calculations above are repeated for the cost savings which would have occurred if the total float glass production 1962–92 had been produced using the plate or sheet processes, assuming that both would have improved in productivity at the same rate as the US business sector productivity (as in the 1974–92 sheet‐equivalent cost calculation above).17 Estimated cost trajectories for plate and sheet glass are calculated from the plate and sheet glass costs in 1962 and 1974 (respectively) implied by the float glass trajectory above.18 This gives the projected surplus for existing technologies over the 1962–92 period, using the same method as for float glass, and compared with the same baseline (i.e. compared to 1962–74 vintage plate and sheet technologies).

The net cost surplus attributable to the float process is the (gross) cost surplus due to the float process minus that estimated for likely productivity improvements in plate and sheet technologies.19 In presenting the surplus figures in a consistent form we need to account for (a) the changes in purchasing power of money over the thirty‐year period, due to inflation, and (b) the time value of money (which makes cost savings for earlier periods of greater value than for later periods). The surplus figures are presented in two ways. In the first, we present the savings in real 1992 dollars (i.e. savings in each year and adjusted by the consumer price index to express them in constant 1992 dollars), summed for the whole period 1962–92. In the second, we use the same real (p.236) figures, but also take account of the time value of money, by ‘projecting’ the savings in each year forward to 1992.20 We use a 5% social discount rate, which is the normally accepted figure for social benefit accounting.21

The production cost savings results are shown in Table B.1. The undiscounted savings are $105.3 billion for float, and $45.6 billion for projected plate and sheet technologies, giving a net savings of $59.7 billion for the float process. When these figures are compounded to 1992, the net savings are $101.6 billion.22

Capital Costs Savings

To obtain the total cost surplus we add the capital cost savings for the 1965 float plants constructed around the world during 1962–92 (see Table B.4). The cost of a float plant is about 50% of the cost of a plate plant.23 Assuming that a third of the plants that would have been built in the absence of Pilkington's invention would have been plate plants (i.e. using the 35 per cent ratio of plate to total US flat glass capacity for 1960–68), and a current float plant costs about $150 million, this implies savings of $8.7 billion, in 1992 dollars, shown in Table B.1. The figure discounted to 1992 is $15.6 billion.

Summing production cost and capital cost effects, the total cost savings are $68.4 billion, or $117.2 billion compounded to 1992.

B.6.2 Consumer Surplus

The consumer surplus due to Pilkington's invention of the float process is the price reduction to levels below what users would have otherwise paid for plate (p.237) or sheet glass, multiplied by the appropriate unit volumes. This is estimated using price and shipment data for the US market (the only large market for which detailed data are readily available). The US surplus is then used to estimate the total world consumer surplus from float glass (see Table B.5).

US glass shipments over the period 1960–94 are shown in Fig. B.3. US flat glass real prices are shown in Fig. B.4. The float process began to replace the plate process after 1965, with no new plate glass capacity added after 1962. By the mid‐1970s most plate production in the US had ceased.24 However, it appears that float glass only began to affect plate glass prices significantly in about 1969, when float production had become a large enough proportion of US plate/float glass production to start to bring down prices. This is indicated by the behaviour of price and shipment data. Shipments of plate/float glass were fairly constant over the period 1965–69, then started to increase steadily. This increase coincided with significant (real) reductions in plate/float glass prices from 1969 onwards.25

We assume that this acceleration in the price decline of plate/float glass was due to the introduction of the float process, and measure the consumer surplus for float glass from 1969.

After about 1974, it is also apparent that float glass began to significantly replace sheet glass, as float glass prices fell closer to sheet glass prices. This is indicated by the rapid decrease in sheet glass shipments from 1974 onwards. By 1977, sheet glass production had fallen to the point that it was no longer counted as a separate item in US Commerce Department figures.26 Thus we assume that float glass began to substitute for plate glass from 1969 on, and for sheet glass from 1974 on.

As in the cost surplus calculation, we calculate the consumer surplus in two stages. We first calculate the consumer benefits due to float glass price reductions after 1969 (effectively compared to a baseline case of no change in plate glass prices after 1969, and in sheet glass prices after 1974). We then calculate the projected surplus in the ‘but for’ case, if the float process had not been introduced, and plate glass prices had continued to fall at their pre‐1969 rates and sheet prices at their pre‐1974 rates. The net consumer surplus is the difference between these two figures.

To calculate the actual surplus due to Pilkington's invention of the float prices, we take the difference between the actual price paid for plate/float glass (p.238)

Appendix B: The Glass Industry and the Pilkington Float Process*

Fig. B.3. Projected US domestic shipments of flat, sheet, and plate glass

and the ‘reservation prices’ of groups of users, multiplied by the volume of glass shipments to each group of users.27 A ‘reservation price’ is the maximum price that a user is willing to pay for glass. At any price below the reservation price the user is receiving some consumer surplus (the difference between the maximum price he or she was willing to pay and the price he or she in fact did pay). Different users have different valuations of glass, and hence different reservation prices.

The consumer surplus calculation must account for users who were already buying plate glass in 1968, but also for new users who entered the market after 1968. Their reservation prices are taken as the maximum prices they have shown that they are prepared to pay for glass, that is the price level in the year in which they entered the market. Since prices have been falling, we assume that each year's increase in demand occurs by attracting new users who were (p.239)

Appendix B: The Glass Industry and the Pilkington Float Process*

Fig. B.4. Projected US real domestic prices for flat, sheet, and plate glass

indifferent between buying and not buying the product at the previous year's price level. These new users would not have bought plate glass at the earlier prices (e.g. they were sheet users), but are willing to buy float glass at the new prices, and receive consumer surplus if prices fall below their entry price.28 In other words, for ‘new’ sales after 1969, the surplus is calculated for ‘vintage’ users from the price reduction below the level ruling when this ‘vintage’ group entered the market. Any reduction below that level is their surplus.29 To calculate (p.240) the surplus we use plate/float data for 1969–76, then total float data from 1977–92.30

In practice the consumer surplus is calculated as follows. For each year, t, the price reduction (P t−1P t) creates an incremental consumer surplus of (P t−1P t) per unit for all the purchasers Q t−1 who previously, in year t−1, were prepared to pay at least P t−1 (this understates the surplus slightly as some new buyers would have been prepared to pay above P t in the range P t to P t−1). The incremental surplus applies over the remaining years until the end of the period, T, i.e. for a further Tt years. Thus the net incremental surplus associated with each year is Q t−1 (P t−1P t) (Tt). The total surplus over the period is the sum of the individual years. This automatically gives a correct net value for the surplus, accounting for years in which prices rise (since this gives a negative increment for that year) and for years in which the total demand decreases over the previous year (also giving a negative increment for that year). In calculating the discounting surplus (net present value) the increment in each year is discounted to the current year, t. The individual year increments are then discounted to the desired year, 1992.

The second step is to calculate the estimated surplus in the ‘but for’ case, assuming that float technology had not been introduced. This requires that plate and sheet glass are calculated differently, since the float process began to replace the plate process in 1969, and to replace the sheet process in 1974. The projected shipment volumes for plate and sheet glass are calculated by assuming that revenue demand increases at the rate of US GDP as a whole (i.e. plate and sheet revenues stay the same percentage of GDP).31 This is calculated for plate for the period 1969–92 and for sheet for the period 1974–92.32 The volume projections (i.e. adjusted for the projected prices) are shown in Fig. B.3.

We also need the projected prices for plate and sheet glass for the relevant periods. Projected prices for plate glass for 1969–92 are calculated from the average rate of decline during 1960–68. The projected prices for sheet glass for (p.241) 1974–92 are calculated from the average rate of decline 1964–73. These are shown in Fig. B.4.33

The projected consumer surplus for plate glass is calculated as above, using the projected prices and volumes for the period 1969–92. This is repeated for projected sheet glass prices and shipments, for the period 1974–92. The total ‘but for’ surplus is the sum of the plate glass and sheet glass surpluses.34

The net consumer surplus is the float surplus minus the estimated plate and sheet surpluses. This gives the US consumer surplus for each year for flat glass shipments. World consumer surplus is estimated from US flat glass figures in two steps. First, US shipment data used here are for flat glass only. Float glass production is also used to manufacture laminated glass. The surplus for flat glass is adjusted upwards, using the average ratio of the value of shipments of laminated to flat glass over the period 1979–91, when laminated sales revenues were an average 0.84 as large as flat glass sales.35 The raw surplus figures for the USA are multiplied by a factor of 1.84. Second, the world‐wide surplus is estimated by multiplying the US surplus by the share of US float capacity in the world total float capacity each year (as used above).

The consumer surplus results are shown in Table B.1. The undiscounted surplus is $118.0 billion for float, and $48.5 billion for projected plate and sheet technologies, giving a net surplus of $69.5 billion attributable to Pilkington's invention of the float process. When these figures are compounded to 1992, the net surplus is $105.4 billion.

B.6.3 Pilkington's Private Earnings from Float and the Ratio of Private to Social Benefits

The main component of Pilkington's earnings from having invented the float process is the royalty and licensing fees from the float technology licences (see Table B.6). In undiscounted 1992 dollars, this amounts to $2.1 billion, or $4.1 billion when discounted forward to 1992.36 A further component may be Pilkington's earnings from glass production using the float process, associated with Pilkington's increased sales in an industry where the float process was available. Some of Pilkington's increase in sales has been related to expansion in the world market for glass, and some to its increased share of (p.242)

Table B.1 Summary of Private and Social Benefits from Float Technology

Surplus (1992 $bn)

Undiscounted

Discounted (1992)††

1. Cost Surplus

Float surplus*

105.3

175.9

Projected plate/sheet surplus**

45.6

74.4

Net cost surplus

59.7

101.6

Plant cost savings

8.7

15.6

Total cost surplus

68.4

117.2

2. Consumer Surplus

 Float surplus

118.0

170.6

 Projected plate/sheet surplus††

48.5

65.2

Net consumer surplus

69.5

105.4

3. Pilkington Private Earnings

Licensing royalties

2.1

4.1

Other earnings before tax

.8

1.2

Total

2.9

5.3

4. Consumer + Producer Surplus

Total (consumer + producer) surplus

72.4

110.7

5. Pilkington Earnings as Percentage of Social Benefits

Cost surplus

4.2%

4.5%

Consumer surplus

4.2%

5.0%

(Consumer + producer) surplus

4.0%

4.8%

() Discounted present value of benefit stream—to 1992, in 1992 dollars, using 5% discount rate

(*) Compared to baseline technology: 1962 plate and/or 1974 sheet

(**) Assuming plate and sheet increase productivity in line with overall manufacturing productivity, compared to baseline

() Compared to baseline prices: 1962 plate and/or 1974 sheet

(††) Assuming plate and sheet prices continue to fall at 1960–68/73 trend rate, compared to baseline

the market.37 A conservative view (i.e. one that puts the highest value on Pilkington's incremental earnings attributable to the float process) is that this increase in market share has been made feasible by the availability of the float process.38 Pilkington's share of the world glass market had risen from about (p.243) 15–20% in 1962 to about 20–25% in 1992.39 If we use the lower figures in both cases, an increase from 15% to 20% means that 25% of Pilkington's 1992 revenues are explained by its increase in market share, compared to what they would have been had its share stayed at 15%. We make the further assumption that 25% of its total earnings before tax, after excluding royalties, may also be attributed to this increase in share. If we assume that Pilkington's market share has increased linearly over the period 1962–92, we are then able to estimate the portion of earnings in each year which is attributable to this increased share.

The results of this calculation are that Pilkington's incremental earnings due to increased world market share are $0.8 billion, in 1992 dollars, or $1.2 billion when discounted forward to 1992. The total of royalty and other earnings are $2.9 billion in 1992 dollars, or $5.3 billion compounded to 1992. Comparing the private and social benefits, it is shown in Table B.1 that, using undiscounted figures, Pilkington has retained 4.2% of the total social benefits measured by the cost surplus or 4.2% of the benefits measured by the consumer surplus. Using compounded figures, Pilkington has retained 4.5% or 5.0% of the total benefits, depending on the measure of social surplus. If the total social surplus is measured by the sum of the consumer surplus and the producer surplus (represented by Pilkington's total private earnings) this gives a total surplus of $72.4 billion undiscounted or $110.7 billion compounded to 1992, of which Pilkington has retained 4.0% or 4.8% respectively.

B.6.4 Private and Social Rates of Return

The benefits from innovation may also be measured by the private and social rates of return for float glass and other glass technology over the period. The main development costs for float glass were Pilkington's R&D costs leading up to the invention in 1959 plus subsequent R&D expenditures for improvements to float glass over the next 15 years, from 1960 to 1975, after which time there were few new patents. Total Pilkington development costs for the original innovation plus improvements were about $174 million in undiscounted 1992 dollars, or $740 million when compounded to 1992. The private earnings to Pilkington were $2.9 billion undiscounted, or $5.3 billion compounded to 1992, as above. The private rate of return to Pilkington is estimated to be about 21%. If we include the estimated R&D costs for improvements made by licensees, the private rate of return to innovation in float glass to all producers as a group falls to about 18%.40

(p.244) We may also calculate the total social rates of return for the float glass innovation. The social rate of return using the cost surplus, allowing for Pilkington and licensee development costs, is estimated to be about 62%. The rate of return using the consumer surplus is about 38%, and using the total consumer plus producer surplus it is about 40%. We may further estimate the social rates of return for all innovation effort in the glass industry over the period by including the additional costs of other ‘dry hole’ R&D in glass technology, which did not lead to successful innovations and was generally rendered obsolete by float glass.41 The social rate of return to innovation of all technologies in the glass industry using the cost surplus, for the period from the mid 1950s until 1992, is estimated to be 41%. The rate of return using the consumer surplus is 29%, and using the total consumer plus producer surplus it is 30%.

These figures may be compared with other estimates of private and social rates of return for innovation in the literature. A few studies have estimated cost and consumer surplus figures for innovation, using various methods. These have reported private rates of return for specific industrial innovations mostly in ranges from about 5% to 40% and corresponding social rates of return in ranges from about 15% to 100%, with a few rates of return outside even these ranges.42 Other studies looking at industry‐wide rates of return to innovation have reported social rates of return in ranges from about 10% to 50%.43 It appears that the rates of return for float glass are towards the upper part of these ranges. This is all the more remarkable given the profound impact that float glass has had on the industry, the almost complete displacement of other (p.245) technologies, and the length of time over which improvements took place. The ratio of social to private rates of return for float glass is also consistent with these other studies, which show social rates of return that are two to three times the magnitude of private rates. Of course, the absolute share of returns going to social benefits is a larger ratio—for float glass the social surplus was over 20 times the private benefits.

B.6.5 Conclusion

In summary, there have been huge social benefits from the introduction of float technology. Estimates of cost savings and consumer surplus attributable to float both give estimates amounting to tens if not hundreds of billions of dollars. The fact that consumer surplus figures are high, and of the same order as the cost savings, implies that most of the benefits have been passed to consumers in the form of lower prices. The price reductions themselves since the introduction of float have been dramatic. There have also been a great number of benefits which are more difficult to quantify, but which have had a great impact on the market for glass, including new products, new uses of flat glass, and a higher quality of glass. In addition, there have been environmental benefits such as lower pollution and energy consumption.

The extraordinary divergence between private and social returns in this case underscores the liberality of Pilkington's market entry strategy and its licensing policy. It may also speak to the inherent limits of licensing as a vehicle to capture value from innovation. Pilkington's licensing policy led to the widespread diffusion of float, which in turn led to the large social returns. The private‐vs.‐social‐benefit calculation provides compelling evidence of the positive, pro‐competitive effect of Pilkington's licensing strategy. Whether Pilkington left too much money on the table (as compared to an alternative investment strategy) will never be answered definitively.

Our calculations indicate that social returns to innovation are high, and that the social returns are markedly greater than private returns. Studies in the literature show that social returns to innovation are generally high and that they greatly exceed the private returns, although rarely by as much as we observe for float. Mansfield showed that private rates of return on a sample of innovations varied from over 200% to less than zero, with a median return of 25%. Social returns for the same innovations were also in a range with lower bounds of about 200% to less than zero, but with a median return of 56% (Mansfield 1976, 1991; Mansfield et al. 1971, 1977). These were innovations of moderate impact, however, not on the same scale of impact on their industries as the float process was to the glass industry. Moreover, the Mansfield studies did not include a deduction for the ‘but for’ world. That is, we assume that the float process (or its economic equivalent) would have eventually been discovered. Our estimates would be even higher had we assumed that the gap with the older technologies was maintained throughout the period.

(p.246)

Table B.2 Float Glass Production—Cost Savings Calculation

Parameters

Social discount rate (real)

5%

(13)

Plate Capacity Share 1965–69

35%

(12)

Factory cost/tonne (1992)

264.29

(14)

Cost redn. rate (62–74) (calc.)

0.9306

(a) Capacity (also used in consumer surplus calculation)

1962

1963

1964

1965

1966

1967

1968

1969

1970

1971

1972

1973

World float capacity (tons)

59534

75451

91955

247111

601026

868454

1034560

1283935

1285745

2627695

3195357

3863029

US capacity

0

0

19398

63380

187437

409878

518023

627490

869750

1180614

1453003

1801008

(11)

US capacity share

0.0%

0.0%

21.1%

25.6%

31.2%

47.2%

50.1%

48.9%

47.6%

44.9%

45.5%

46.6%

(b) Cost Surplus Calculation ‐ Float (versus baseline)

US productivity (bus. op/hr.)

70.5

73.3

76.5

78.6

80.7

82.8

85.3

85.8

87

89.8

92.7

95

(1)

Cost mult.‐ rel. 1962 plate

0.593

0.552

0.513

0.478

0.445

0.414

0.385

0.358

0.333

0.310

0.289

0.269

(2)

Cost mult.‐ rel. 1974 sheet

Cost/tonne (92$)

776.9

723.0

672.7

626.0

582.5

542.1

504.4

469.4

436.8

406.5

378.2

352.0

(5)

plate‐equiv. prop.

1.000

1.000

1.000

1.000

1.000

1.000

1.000

1.000

1.000

1.000

1.000

1.000

(6)

capacity ‐ plate‐equiv.

59534

75451

91955

247111

601026

868454

1034560

1283935

1825745

2627695

3195357

3863029

capacity ‐ sheet‐equiv.

(3)

man cost ‐ plate‐equiv. (92$bn)

0.05

0.05

0.06

0.15

0.35

0.47

0.52

0.60

0.80

1.07

1.21

1.36

(4)

man cost ‐ sheet‐equiv. (92$bn)

(7)

CS‐plate (92$bn)

0.03

0.04

0.06

0.17

0.44

0.67

0.83

1.08

1.59

2.37

2.98

3.70

(8)

NPV‐plate (nvp 62, 92$bn)

$38.11

(9)

CS‐sheet (92$bn)

NPV‐sheet (92$bn)

$2.60

Total (npv 62, 92$bn)

$40.70

Total (npv 92, 92$bn)

$175.92

FLOAT

(c) Cost Surplus Calculation ‐ No Float (‘but‐for’ versus baseline)

(10)

Cost mult‐plate

1

0.96

0.92

0.90

0.87

0.85

0.83

0.82

0.81

0.79

0.76

0.74

(15)

Cost mult. ‐sheet

(16)

Cost/tonne ‐ plate (92$)

1310.13

1260.08

1207.37

1175.11

1144.53

1115.51

1082.81

1076.50

1061.65

1028.55

996.37

972.25

(17)

Cost/tonne ‐ sheet (92$)

man cost ‐ plate

0.0780

0.0951

0.1110

0.2904

0.6879

0.9688

1.1202

1.3822

1.9383

2.7027

3.1838

3.7558

man cost ‐ sheet

CS ‐ plate (92$bn)

0.00

0.00

0.01

0.03

0.10

0.17

0.24

0.30

0.45

0.74

1.00

1.31

NPV‐plate

$14.61

CS ‐ sheet (92$bn)

NPV‐sheet

$2.60

Total (npv 62, 92 $bn)

$17.21

Total (npv 92, 92 $bn)

$74.37

NOFLOAT

(d) Net Cost Surplus Calculation

Total (npv 62, 92 $bn)

$23.50

Total (npv 92, 92 $bn)

$101.56

NET***

Parameters

Social discount rate (real)

(13)

Plate Capacity Share 1965–69

(12)

Factory cost/tonne (1992)

(14)

Cost redn. rate (62–74) (calc.)

(a) Capacity (also used in consumer surplus calculation)

1974

1975

1976

1977

1978

1979

1980

1981

1982

1983

1984

1985

World float capacity (tons)

4732886

5415299

6239974

6833105

7536339

8207491

8498127

9302704

10136242

10731717

11146631

11857886

US capacity

2135617

2428842

2704423

2749106

2963030

3044411

3126328

3346475

3499773

3521736

3596283

3683624

(11)

US capacity share

45.1%

44.9%

43.3%

40.2%

39.3%

37.1%

36.8%

36.0%

34.5%

32.8%

32.3%

31.1%

(b) Cost Surplus Calculation ‐ Float (versus baseline)

US productivity (bus. op/hr.)

93.2

95.5

98.3

99.9

100.5

99.4

98.6

99.9

100

102.3

104.8

106.3

(1)

Cost mult.‐ rel. 1962 plate

0.250

0.244

0.237

0.233

0.232

0.234

0.236

0.233

0.233

0.228

0.222

0.219

(2)

Cost mult.‐ rel. 1974 sheet

1.000

0.976

0.948

0.933

0.927

0.938

0.945

0.933

0.932

0.911

0.889

0.877

Cost/tonne (92$)

327.5

319.6

310.5

305.6

303.7

307.1

309.6

305.6

305.3

298.4

291.3

287.2

(5)

plate‐equiv. prop.

0.907

0.814

0.721

0.629

0.536

0.443

0.350

0.350

0.350

0.350

0.350

0.350

(6)

capacity ‐ plate‐equiv.

4293404

4409601

4501696

4295095

4037324

3634746

2974344

3255946

3547685

3756101

3901321

4150260

capacity ‐ sheet‐equiv.

439482

1005698

1738278

2538010

3499015

4572745

5523783

6046758

6588557

6975616

7245310

7707626

(3)

man cost ‐ plate‐equiv. (92$bn)

1.41

1.40

1.41

1.31

1.23

1.12

0.92

0.99

1.08

1.12

1.14

1.19

(4)

man cost ‐ sheet‐equiv. (92$bn)

0.14

0.32

0.54

0.78

1.06

1.40

1.71

1.85

2.01

2.08

2.11

2.21

(7)

CS‐plate (92$bn)

4.22

4.37

4.50

4.31

4.06

3.65

2.98

3.27

3.56

3.80

3.97

4.25

(8)

NPV‐plate (nvp 62, 92$bn)

(9)

CS‐sheet (92$bn)

0.00

0.01

0.03

0.06

0.08

0.09

0.10

0.13

0.15

0.20

0.26

0.31

NPV‐sheet (92$bn)

Total (npv 62, 92$bn)

Total (npv 92; 92$bn)

(c) Cost Surplus Calculation ‐ No Float (‘but‐for’ versus baseline)

(10)

Cost mult‐plate

0.76

0.74

0.72

0.71

0.70

0.71

0.72

0.71

0.71

0.69

0.67

0.66

(15)

Cost mult. ‐sheet

1.000

0.976

0.948

0.933

0.927

0.938

0.945

0.933

0.932

0.911

0.889

0.877

(16)

Cost/tonne ‐ plate (92$)

991.03

967.16

939.61

924.56

919.04

929.21

936.75

924.56

923.64

902.87

881.34

868.90

(17)

Cost/tonne ‐ sheet (92$)

327.532

319.644

310.539

305.565

303.741

307.21

309.594

305.565

305.260

298.396

291.278

287.168

man cost ‐ plate

4.2549

4.2648

4.2299

3.9711

3.7105

3.3775

2.7862

3.0103

3.2768

3.3913

3.4384

3.6062

man cost ‐ sheet

0.1439

0.3215

0.5398

0.7755

1.0628

1.4043

1.7101

1.8477

2.0112

2.0815

2.1104

2.2134

CS ‐ plate (92$bn)

1.37

1.51

1.67

1.66

1.58

1.38

1.11

1.26

1.37

1.53

1.67

1.83

NPV‐plate

CS ‐ sheet (92$bn)

0.00

0.01

0.03

0.06

0.08

0.09

0.10

0.13

0.15

0.20

0.26

0.31

NPV‐sheet

Total (npv 62, 92 $bn)

Total (npv 92, 92 $bn)

(d) Net Cost Surplus Calculation

Total (npv 62, 92 $bn)

Total (npv 92, 92 $bn)

Parameters

Social discount rate (real)

(13)

Plate Capacity Share 1965–69

(12)

Factory cost/tonne (1992)

(14)

Cost redn. rate (62–74) (calc.)

(a) Capacity (also used in consumer surplus calculation)

Total

1986

1987

1988

1989

1990

1991

1992

(Undis)

World float capacity (tons)

12345941

13165669

14831271

15711974

16694213

17261566

18235397

262.43

US capacity

3635784

3682516

3865628

4119181

3969718

3731017

3839205

78.51

(11)

US capacity share

29.4%

28.0%

26.1%

26.2%

23.8%

21.6%

21.1%

29.9%

(b) Cost Surplus Calculation ‐ Float (versus baseline)

US productivity (bus. op/hr.)

108.5

109.6

110.7

109.9

110.7

111.8

115.5

(1)

Cost mult.‐rel. 1962 plate

0.215

0.213

0.210

0.212

0.210

0.208

0.202

(2)

Cost mult.‐rel. 1974 sheet

0.859

0.850

0.842

0.848

0.842

0.834

0.807

Cost/tonne (92$)

281.3

278.5

275.8

277.8

275.8

273.0

264.3

(5)

plate‐equiv. prop.

0.350

0.350

0.350

0.350

0.350

0.350

0.350

(6)

capacity ‐ plate‐equiv.

4321079

4607984

5190945

5499191

5842975

6041548

6382389

capacity ‐ sheet‐equiv.

8024862

8557685

9640326

10212783

10851238

11220018

11853008

(3)

man cost ‐ plate‐equiv(92$bn)

1.22

1.28

1.43

1.53

1.61

1.65

1.69

(4)

man cost ‐ sheet‐equiv(92$bn)

2.26

2.38

2.66

2.84

2.99

3.06

3.13

(7)

CS‐plate (92$bn)

4.45

4.75

5.37

5.68

6.04

6.27

6.67

$100.14

(8)

NPV‐plate (nvp 62, 92$bn)

(9)

CS‐sheet (92$bn)

0.37

0.42

0.50

0.51

0.56

0.61

0.75

$5.15

NPV‐sheet (92$bn)

$105.29

Total (npv 62, 92$bn)

Total (npv 92, 92$bn)

(c) Cost Surplus Calculation ‐ No Float (‘but‐for’ versus baseline)

(10)

Cost mult‐plate

0.65

0.64

0.64

0.64

0.64

0.63

0.61

(15)

Cost mult.‐sheet

0.859

0.850

0.842

0.848

0.842

0.834

0.807

(16)

Cost/tonne ‐ plate (92$)

851.28

842.74

834.36

840.44

834.36

826.15

799.69

(17)

Cost/tonne ‐ sheet (92$)

281.345

278.522

275.754

277.761

275.754

273.041

264.294

man cost ‐ plate

3.6785

3.8833

4.3311

4.6217

4.8752

4.9912

5.1039

man cost ‐ sheet

2.2578

2.3835

2.6584

2.8367

2.9923

3.0635

3.1327

CS ‐ plate (92$bn)

1.98

2.15

2.47

2.58

2.78

2.92

3.26

$40.44

NPV‐plate

CS ‐ sheet (92$bn)

0.37

0.42

0.50

0.51

0.56

0.61

0.75

$5.15

NPV‐sheet

$45.59

Total (npv 62, 92 $bn)

Total (npv 92, 92 $bn)

(d) Net Cost Surplus Calculation

Total (npv 62, 92 $bn)

Total (npv 92, 92 $bn)

$59.70

Notes (Table B.2)

((1)) Direct manufacturing costs relative to 1962 plate costs, taken as 59.3% in 1962, fall from 59.3% to 25% by 1962–1974, and matching overall US prod. improv. 1975–92

((2)) Direct manufacturing costs relative to 1974 sheet costs, taken as matching overall US prod. improv. 1975–92

((3)) Estimated manufacturing cost using 1992 technology, in 1992 $, using 1992 Pilkington factory cost data pro rated by annual capacity

((4)) Estimated direct manufacturing costs at actual technology (adjusted by cost multiplier)

((5)) 1962–73: float 100% plate‐equivalent; transition 1974–80: % plate‐equivalent linear fall to steady state; 1981–92: % plate‐equivalent float glass in US 1965–69 plate/sheet avg. (35%, note (13))

((6)) Plate act. manuf. cost 5 total manuf. costs * plate volume share of market (assume all grades same cost per unit); cost saving = cost * (1‐r)/r, where r = cost ratio float/plate.

((7)) Cost surplus below projected manufacturing costs if using 1962 plate process, for plate equivalent float production [CS = Cost (1‐r)/r, where r = cost multiplier relative to 1962 plate cost]

((8)) Net present value of surplus stream 1962–92, discounted to 1962 at 5% social discount rate

((9)) Cost surplus below projected manufacturing costs if using 1974 sheet process, for sheet‐equivalent float production

((10)) Assumes manufacturing cost of plate process 1962–92 would have fallen at same rate as overall improvement in US productivity (output per man hour) each year

((11)) Calculated for use in consumer surplus calculation—in Table B.5

((12)) Pilkington average factory cost/ton for 1992 (92$)—from Table B.3

((13)) US 1965–69 plate/sheet avg. proportions (35%)—from Table B.5 (see note (5)) (35% is proportion of output by sq. ft. plate glass in flat; assume capacity share by tonnage in same proportion [rated capacity & costs determined by sq. ft./hr output, tonnage tank capacity])

((14)) Assume float costs fall 1962–74 at a constant percentage rate from 59% of 1962 plate costs in 1962 to 25% in 1974; then fall in line with overall US productivity improvements 1974–92.

(i.e. rate in cell D35 used in row 36, 1962–74, rate in cell P35 not used)

((15)) Assumes manufacturing costs of sheet process 1974–92 would have fallen at same rate as overall improvement in US productivity (output per man hour) each year

((16)) Plate costs estimated as (1/59.3%) of float costs 1962, declining according to plate multiplier 1962–92, as at note (10)

((17)) Sheet costs estimated as equal to estimated float costs 1974, declining according to plate multiplier 1974–92, as at note (15)—(hence assumed equal to estimated float costs)

((nu)) Not used—information only

(p.247) (p.248) (p.249) (p.250) (p.251)

(p.252)

Table B.2s Summary Results (Cost Surplus)

Undisc.

NPV 92

Float

$105.3

$175.9

No Float

$45.6

$74.4

Net

$59.7

$101.6

(p.253)

Table B.3 Breakdown of Average Costs of Float Glass Production at Pilkington Sites (1988–94)

1988

1989

1990

1991

1992

1993

1994

A + B

Raw Materials

37.5269

40.6741

47.0374

43.2985

43.1595

45.57

44.7237

C + F

Labor

28.3906

38.1947

39.2663

39.882

40.688

45.4515

51.3079

D

Energy

19.5213

24.3612

25.5589

23.986

23.5945

25.3035

26.0079

G

Factory Overhead

6.04

9.26

10.79

9.71

11.37

11.87

12.24

J

Non Factory Overhead

23.71

23.77

23.02

21.51

19.65

18.96

17.26

H

Depreciation

11.8288

13.7053

16.3253

15.984

18.2275

20.4435

20.5137

K

Transport

19.2781

19.4818

20.7395

19.939

22.8175

25.1515

25.7537

E

Miscellaneous

6.12

10.12

10.58

9.92

11.36

13.36

13.84

I

Provisions

2.57

3.5

3.45

1.03

1.3

1.19

1.98

(3)

Total Costs/tonne

154.986

183.067

196.767

185.26

192.167

207.3

213.627

(1)

A to I

Factory cost (£)

111.998

139.815

153.008

143.811

149.7

163.189

170.613

(2)

Factory cost ($)

199.512

229.255

273.073

254.458

264.294

283.948

272.981

Exchange rate ($/£)

1.7814

1.6397

1.7847

1.7694

1.7655

1.7400

1.6000

Notes

((1)) Direct factory costs (£/metric ton)

((2)) 1992 average factory costs used in cost surplus calculation—to Table B.2

((3)) All costs in £/metric ton

(p.254)

Table B.4 Float Cost Savings on New Plate Glass Plants

Parameters

Social discount rate (real)

5%

A

Cost/plant (92$k)

150

B

Plate‐equiv. plant %

35%

Cost Saving Calculation

1962

1963

1964

1965

1966

1967

1968

1969

1970

1971

1972

C

Total New Plants

1

0

1

4

3

2

4

3

7

5

5

Cost plants (92$bn)

0.15

0.00

0.15

0.60

0.45

0.30

0.60

0.45

1.05

0.75

0.75

1973

1974

1975

1976

1977

1978

1979

1980

1981

1982

1983

Total New Plants

6

8

3

6

4

5

3

3

5

6

5

Cost plants (92$bn)

0.90

1.20

0.45

0.90

0.60

0.75

0.45

0.45

0.75

0.90

0.75

1984

1985

1986

1987

1988

1989

1990

1991

1992

Total

Total New Plants

3

9

6

11

14

9

12

4

8

165

Cost plants (92$bn)

0.45

1.35

0.90

1.65

2.10

1.35

1.80

0.60

1.20

24.75

Und 92$

D

NPV—Cost plants (npv 62, 92$bn)

$10.32

(1)

Saving (62, 92$bn)

$3.61

Saving

$8.66

Saving (npv 92, 92$bn)

$15.61

Und 92$

Notes

((1)) Cost savings from float plants built instead of plate plants of the same capacity = NPV (Annual A*C)*B [where B equals historical plate capacity share—see World Capacity note (13)]

Table B.4s Summary Results (Plant Cost)

Undisc.

NPV 92

Float

No Float

Net

$8.7

$15.6

(p.255)

Table B.5 Consumer Surplus Calculation—US Domestic Glass Shipments

I US Domestic Glass Shipments Data

1957

1958

1959

1960

1961

Price Deflator—U.S. Consumer Price Index (CPI):

CPI

28.1

28.9

29.1

29.6

29.9

base 1957

100.00

102.85

103.56

105.34

106.41

C

base 1962

93.05

95.70

96.36

98.01

99.01

base 1992

20.03

20.60

20.74

21.10

21.31

Total U.S. Shipments (millions of sq.ft.):

A

Flat Glass

1297.18

1114.32

1828.30

1492.53

1429.80

Sheet Glass

1054.00

895.55

1307.25

1036.90

1041.80

Plate/Float Glass

243.18

218.77

521.05

455.63

388.00

(17)

5yr avg share plate

0.2550

Total Revenues—(Nominal $k):

B

Flat Glass

216583

189019

333733

281928

260451

Sheet Glass

106947

92339

134502

106476

110971

Plate/Float Glass

109636

96680

199231

175452

149480

(Laminated) ($m)

(lam + flat)/flat

(lam + flat)/flat [avg.]

Total Revenues—Real—(1962 $) (dir):

Flat Glass

232768.9

197521.6

346348.3

287642.8

263064.2

Sheet Glass

114939.5

96492.7

139586.3

108634.3

112084.4

Plate/Float Glass

117829.4

101028.9

206762.1

179008.5

150979.8

Real Price U.S. Domestic Shipments—Indexed to 1962—Deflated/Set to 1962 $ Absolute Value:

(18)

Flat Glass

179.44

177.26

189.44

192.72

183.99

Sheet Glass

109.55

107.75

106.78

104.77

107.59

Plate/Float Glass

484.54

461.80

396.82

392.88

389.12

II Consumer Surplus Calculations—Projections from US Shipments

Parameters

Social discount rate (real)

5%

CPI multiplier 1962–92

4.6457

(15)

Avg. share plate 1965–69

0.3531

(13)

Float/lamin.: US mult.

1.8428

Base Data Series—Shipments and Prices

(16)

US/World float cap share

(1)

US Ship. float/flat

243.2

218.8

521.1

455.6

388.0

Price float/flat (62$/tonne)

484.5

461.8

396.8

392.8

389.1

Increm. price redn

Remaining yrs. (undisc.)

dummy for NPV

(2)

Remaining years (NPV)

(a) Consumer Surplus (CS)—Original plate users only

(a.1) Consumer Surplus (CS)—Original plate users only—Float

(7)

CS/yr.

C. Surplus (NPV)

CS/yr. (flat + lam)

NPV CCS (f + l).

World npv CS/yr.(flat + lam)

NPV CS world

(a.2) Consumer Surplus (CS)—Original plate users only—No Float

(7)

CS/yr.

NPV‐Cons. Surplus

CS/yr. (flat + lam)

NPV‐CS (f + l).

NPV‐CS‐World/yr.(flat + lam)

NPV‐CS World

Net world

I US Domestic Glass Shipments Data

1962

1963

1964

1965

1966

1967

1968

1969

1970

1971

1972

1973

Price Deflator—U.S. Consumer Price Index (CPI):

CPI

30.2

30.6

31

31.5

32.4

33.4

34.8

36.7

38.8

40.5

41.8

44.4

base 1957

107.47

108.90

110.32

112.10

115.30

118.86

123.84

130.60

138.08

144.13

148.75

158.01

C

base 1962

100.00

101.32

102.65

104.30

107.28

110.60

115.23

121.52

128.48

134.11

138.41

147.02

base 1992

21.53

21.81

22.10

22.45

23.09

23.81

24.80

26.16

27.66

28.87

29.79

31.65

Total U.S. Shipments (millions of sq.ft.):

A

Flat Glass

1587.90

1724.64

1716.75

1809.45

1729.04

1613.44

1751.80

1866.19

1761.19

2131.81

2389.12

2530.47

Sheet Glass

1163.15

1248.40

1216.90

1203.90

1146.50

1066.25

1095.80

1160.95

1069.70

1188.75

1197.50

1125.90

Plate/Float Glass

424.75

476.24

499.85

605.55

582.54

547.19

656.00

705.24

691.49

943.06

1191.62

1404.57

(17)

5yr avg share plate

0.2695

0.2810

0.2823

0.2896

0.3022

0.3155

0.3354

0.3531

0.3649

0.3883

0.4230

0.4622

Total Revenues—(Nominal $k):

B

Flat Glass

285987

317299

324955

354308

343138

331976

387469

416870

382942

464674

550292

585290

Sheet Glass

126448

141479

144753

140559

136785

131476

139391

150123

131551

150344

157187

152242

Plate/Float Glass

159539

175820

180202

213749

206353

200500

248078

266747

251391

314330

393105

433048

(Laminated) ($m)

(lam + flat)/flat

(lam + flat)/flat [avg.]

Total Revenues—Real—(1962 $) (dir):

Flat Glass

285987.0

313151.3

316569.1

339685.8

319838.5

300169.9

336251.8

3430337.4

298063.1

346497.6

397579.4

398102.7

Sheet Glass

126448.0

139629.6

141017.4

134758.2

127497.1

118879.5

120965.8

123534.5

102392.8

112108.4

113565.7

103552.0

Plate/Float Glass

159539.0

173521.7

175551.6

204927.6

192341.4

181290.4

215286.1

219503.0

195670.3

234389.3

284013.7

294550.7

Real Price U.S. Domestic Shipments—Indexed to 1962—Deflated/Set to 1962 $ Absolute Value:

(18)

Flat Glass

180.10

181.57

184.40

187.73

184.98

186.04

191.95

183.82

169.24

162.54

166.41

157.32

Sheet Glass

108.71

111.85

115.88

111.93

111.21

111.49

110.39

106.41

95.72

94.31

94.84

91.97

Plate/Float Glass

375.61

364.36

351.21

338.41

330.18

331.31

328.18

311.24

282.97

248.54

238.34

209.71

II Consumer Surplus Calculations—Projections from US Shipments

Parameters

Social discount rate (real)

CPI multiplier 1962–92

(15)

Avg. share plate 1965–69

(13)

Float/lamin.: US mult.

Base Data Series—Shipments and Prices

(16)

US/World float cap share

0.0%

0.0%

21.1%

25.6%

31.2%

47.2%

50.1%

48.9%

47.6%

44.9%

45.5%

46.6%

(1)

US Ship. float/flat

424.7

476.2

499.9

605.6

582.5

547.2

656.0

705.2

691.5

943.1

1191.6

1404.6

Price float/flat (62$/tonne)

375.6

364.4

351.2

338.4

330.2

331.3

328.2

311.2

283.0

248.5

238.3

209.7

Increm. price redn

16.93

28.28

34.43

10.20

28.63

Remaining yrs. (undisc.)

24

23

22

21

20

dummy for NPV

1

1

1

1

1

(2)

Remaining years (NPV)

14.49

14.16

13.82

13.46

13.09

(a) Consumer Surplus (CS)—Original plate users only

(a.1) Consumer Surplus (CS)—Original plate users only—Float

(7)

CS/yr.

0

0

0

0

0

0

0

11.11

29.66

52.24

58.93

77.72

C. Surplus (NPV)

$1,076

CS/yr. (flat + lam)

0

0

0

0

0

0

0

20.47

54.65

96.27

108.60

143.22

NPV CCS (f + l).

$1,983

World npv CS/yr.(flat + lam)

0

0

0

0

0

0

0

41.89

114.73

214.28

238.83

307.19

NPV CS world

$5,877

(a.2) Consumer Surplus (CS)—Original plate users only—No Float

(7)

CS/yr.

0

0

0

0

0

0

0

3.40

7.82

12.15

16.40

20.55

NPV‐Cons. Surplus

$384

CS/yr. (flat + lam)

0

0

0

0

0

0

0

6.26

14.41

22.40

30.21

37.87

NPV‐CS (f + l).

$707

NPV‐CS‐World/yr.(flat + lam)

0

0

0

0

0

0

0

12.81

30.26

49.85

66.44

81.22

NPV‐CS World

$2,213

Net world

$3,664

I US Domestic Glass Shipments Data

1974

1975

1976

1977

1978

1979

1980

1981

1982

1983

1984

1985

Price Deflator—U.S. Consumer Price Index (CPI):

CPI

49.3

53.8

56.9

60.6

65.2

72.6

82.4

90.9

96.5

99.6

103.9

107.6

base 1957

175.44

191.46

202.49

215.66

232.03

258.36

293.24

323.49

343.42

354.45

369.75

382.92

C

base 1962

163.25

178.15

188.41

200.66

215.89

240.40

272.85

300.99

319.54

329.80

344.04

356.29

base 1992

35.14

38.35

40.56

43.19

46.47

51.75

58.73

64.79

68.78

70.99

74.06

76.69

Total U.S. Shipments (millions of sq.ft.):

A

Flat Glass

2252.75

1905.78

2603.64

2897.03

3099.84

2974.46

2749.95

2652.61

2255.75

2916.03

3300.68

3463.84

Sheet Glass

857.15

453.25

550.70

Plate/Float Glass

1395.60

1452.53

2052.94

(17)

5yr avg share plate

0.5085

Total Revenues—(Nominal $k):

B

Flat Glass

543382

467994

644751

739919

829449

858130

868459

952283

892059

954927

955088

1174380

Sheet Glass

132541

76229

101739

Plate/Float Glass

410841

391765

543012

(Laminated) ($m)

(lam + flat)/flat

759

647

666

697

808

906

1043

(lam + flat)/flat [avg.]

1.884481

1.744998

1.699372

1.781338

1.846138

1.948604

1.888128

Total Revenues—Real—(1962 $) (dir):

Flat Glass

332862.8

262703.0

342205.3

368738.5

384192.6

356963.2

318294.4

316380.1

279172.9

289546.1

277609.8

329612.2

Sheet Glass

81191.4

42790.3

53998.6

Plate/Float Glass

251.671.4

219912.7

288206.7

Real Price U.S. Domestic Shipments—Indexed to 1962—Deflated/Set to 1962 $ Absolute Value:

(18)

Flat Glass

147.76

137.85

131.43

127.28

123.94

120.01

115.75

119.27

123.76

99.29

84.11

95.16

Sheet Glass

94.72

94.41

98.05

Plate/Float Glass

180.33

151.40

140.39

II Consumer Surplus Calculations—Projections from US Shipments

Parameters

Social discount rate (real)

CPI multiplier 1962–92

(15)

Avg. share plate 1965–69

(13)

Float/lamin.: US mult.

Base Data Series—Shipments and Prices

(16)

US/World float cap share

45.1%

44.9%

43.3%

40.2%

39.3%

37.1%

36.8%

36.0%

34.5%

32.8%

32.3%

31.1%

(1)

US Ship. float/flat

1395.6

1452.5

2052.9

2897.0

3099.8

2974.5

2749.9

2652.6

2255.8

2916.0

3300.7

3463.8

Price float/flat (62$/tonne)

180.3

151.4

140.4

127.3

123.9

120.0

115.7

119.3

123.8

99.3

84.1

95.2

Increm. price redn

29.38

28.93

11.01

13.11

3.34

3.93

4.26

−3.53

−4.49

24.47

15.19

−11.05

Remaining yrs. (undisc.)

19

18

17

16

15

14

13

12

11

10

9

8

dummy for NPV

1

1

1

1

1

1

1

1

1

1

1

1

(2)

Remaining years (NPV)

12.69

12.27

11.84

11.38

10.90

10.39

9.86

9.31

8.72

8.11

7.46

6.79

(a) Consumer Surplus (CS)—Original plate users only

(a.1) Consumer Surplus (CS)—Original plate users only—Float

(7)

CS/yr.

96.99

115.97

123.19

131.79

133.98

136.56

139.36

137.04

134.10

150.15

160.11

152.86

C. Surplus (NPV)

CS/yr. (flat + lam)

178.73

213.71

227.02

242.86

246.90

251.65

256.81

252.54

247.12

276.69

295.05

281.70

NPV CCS (f+l).

World npv CS/yr.(flat+lam)

396.09

476.47

523.80

603.65

627.98

678.44

698.06

702.04

715.72

843.17

914.52

906.80

NPV CS world

(a.2) Consumer Surplus (CS)—Original plate users only—No Float

(7)

CS/yr.

24.61

28.60

32.49

36.31

40.05

43.71

47.29

50.80

54.23

57.60

60.89

64.11

NPV‐Cons. Surplus

CS/yr. (flat+lam)

45.36

52.70

59.88

66.91

73.80

80.54

87.15

93.61

99.94

106.14

112.20

118.15

NPV‐CS (f+l).

NPV‐CS‐World/yr.(flat+lam)

100.52

117.49

138.16

166.32

187.71

217.14

236.88

260.22

289.45

323.43

347.78

380.32

NPV‐CS World

Net world

I US Domestic Glass Shipments Data

1986

1987

1988

1989

1990

1991

1992

Price Deflator—U.S. Consumer Price Index (CPI):

CPI

109.6

113.6

118.3

124

130.7

136.2

140.3

base 1957

390.04

404.27

421.00

441.28

465.12

484.70

499.29

C

base 1962

362.91

376.16

391.72

410.60

432.78

450.99

464.57

base 1992

78.12

80.97

84.32

88.38

93.16

97.08

100.00

Total U.S. Shipments (millions of sq.ft.):

A

Flat Glass

3740.06

4073.89

4225.84

4733.77

4658.67

4327.43

4597.48

Sheet Glass

Plate/Float Glass

(17)

5yr avg share plate

Total Revenues—(Nominal $k):

B

Flat Glass

1259746

1457587

1484949

1543242

1471447

1350658

1528698

Sheet Glass

Plate/Float Glass

(Laminated) ($m)

1145

1390

1293

1249

1218

1009

(lam+flat)/flat

1.908913

1.953631

1.870737

1.809335

1.827757

1.747043

(lam+flat)/flat [avg.]

1.842806

Total Revenues—Real—(1962 $) (dir):

Flat Glass

347119.8

387492.3

379082.5

375854.1

339997.7

299485.1

329056.9

Sheet Glass

Plate/Float Glass

Real Price U.S. Domestic Shipments—Indexed to 1962—Deflated/Set to 1962 $ Absolute Value:

(18)

Flat Glass

92.81

95.12

89.71

79.40

72.98

69.21

71.57

Sheet Glass

Plate/Float Glass

II Consumer Surplus Calculations—Projections from US Shipments

Parameters

Social discount rate (real)

CPI multiplier 1962–92

(15)

Avg. share plate 1965–69

(13)

Float/lamin.: US mult.

Base Data Series—Shipments and Prices

(16)

US/World float cap share

29.4%

28.0%

26.1%

26.2%

23.8%

21.6%

21.1%

(1)

US Ship. float/flat

3740.1

4073.9

4225.8

4733.8

4658.7

4327.4

4597.5

Price float/flat (62$/tonne)

92.8

95.1

89.7

79.4

73.0

69.2

71.6

Increm. price redn

2.35

−2.30

5.41

10.31

6.42

3.78

−2.37

Remaining yrs. (undisc.)

7

6

5

4

3

2

1

dummy for NPV

1

1

1

1

1

1

1

(2)

Remaining years (NPV)

6.08

5.33

4.55

3.72

2.86

1.95

1.00

(a) Consumer Surplus (CS)—Original plate users only

(a.1) Consumer Surplus (CS)—Original plate users only—Float

Total (Und)

(7)

CS/yr.

154.40

152.89

156.44

163.20

167.41

169.89

168.33

$2,974

US

‐62 $

C. Surplus (NPV)

$5,000

US

(npv 92, 62$)

$13,818

US

‐92$

CS/yr. (flat+lam)

284.53

281.75

288.29

300.75

308.50

313.0679

310.206322

$5,481

US (f+1)

‐62$

NPV CCS (f+l).

$39,822

$9,214

US (f+1)

(npv 92, 62$)

$25,463

US (f+1)

‐92$

World npv CS/yr.(flat+lam)

966.18

1007.29

1106.07

1147.15

1297.38

1448.41

1473.41323

$17,450

World

‐62$

NPV CS world

$27,303

World

(npv 92, 62$)

$81,065

FLOAT

‐92$

(a.2) Consumer Surplus (CS)—Original plate users only—No Float

$118,001

FLOAT

(npv 92, 92$)

(7)

CS/yr.

67.27

70.36

73.39

76.35

79.25

82.09

84.87

$1,135

US

‐62$

NPV‐Cons. Surplus

$1,782

US

(npv 92, 62$)

$5,271

US

‐92$

CS/yr. (flat+lam)

123.96

129.66

135.23

140.70

146.04

151.28

156.40

$2,091

US (f+1)

‐62$

NPV‐CS (f+l).

$14,196

$3,285

US (f+1)

(npv 92, 62$)

$9,713

US (f+1)

‐92$

NPV‐CS‐World/yr.(flat+lam)

420.94

463.55

518.86

536.66

614.16

699.88

742.87

$7,003

World

‐62$

NPV‐CS World

$10,281

World

(npv 92, 62$)

$32,533

NOFLOAT

‐92$

Net world

1957

1958

1959

1960

1961

1962

1963

1964

1965

1966

1967

1968

(b) Full Consumer Surplus Calculation—Float

(b.1) CS—undisc.—float vs. baseline

Cons. Surplus (undis.)

CS US (f+l)

CS World (f+l)

(b.2) CS—NPV—float vs. baseline

(3)

NPV‐CS/yr.

0.00

0.00

0.00

0.00

0.00

0.00

0.00

(4)

NPV‐Cons. Surp.

$1,934

(5)

NPV‐CS/yr. (flat+lam)

0.00

0.00

0.00

0.00

0.00

0.00

0.00

NPV CS (f+l).

$3,564

(6)

NPV‐CS‐World/yr.(flat+lam)

0.00

0.00

0.00

0.00

0.00

0.00

0.00

CS world (npv62, 92$bn)

$8,494

CS world (npv92, 92$bn)

(c) Full Consumer Surplus Calculation—No float

(c.1.1) Shipments projections—No float—(GDP growth rate: plate 1969–92; sheet 1974–92)

(14)

US GDP (1987 $)

1973.2

2025.6

2129.8

2218.0

2343.3

2473.5

2622.3

2690.3

2801.0

(10)

Plate (proj.)

215286.1

Sheet (proj.)

120965.8

Total (proj.)

336251.8

Plate (proj.)

656.0

Sheet (proj.)

1095.8

Total (proj.)

1751.8

(c.1.2) Price projections—No float—(plate 1969–92, sheet 1974–92)

avg (3yr)

447.72

417.17

392.94

385.87

376.36

363.73

351.33

339.93

333.30

329.89

(11)

rate plate %

−2.09%

avg (3yr)

107.86

106.43

106.38

107.02

109.38

112.15

113.22

113.01

111.54

111.03

rate sheet %

(8)

Plate (proj.)

484.54

461.80

396.82

389.12

375.61

375.61

364.36

351.21

338.41

330.18

331.31

328.18

(9)

Sheet (proj.)

109.05

107.75

106.78

104.77

107.59

108.71

111.85

115.88

111.93

111.21

111.49

110.39

(12)

Inc price redn‐plate

Inc price redn‐sheet

(c.2) CS projections—undisc.—no float vs. baseline

CS‐Plate

CS‐Sheet

CS‐Total

CS‐Total (f+l)

CS‐World (f+l)

(c.3) CS projections—NPV—no float vs. baseline

CS/yr‐plate

0

0

0

0

0

0

0

CS‐plate

$526

CS/yr‐sheet

0

0

0

0

0

0

0

CS‐sheet

$157

CS/yr‐total

0

0

0

0

0

0

0

CS‐total

$683

NPV CS/yr.(f+l)

0

0

0

0

0

0

0

NPV CS (f+l).

$1,259

NPV‐CS‐World/yr. (flat+lam)

0

0

0

0

0

0

0

NPV CS‐World (f+l).

$3,245

1969

1970

1971

1972

1973

1974

1975

1976

1977

1978

1979

1980

(b) Full Consumer Surplus Calculation—Float

(b.1) CS—undisc.—float vs. baseline

Cons. Surplus (undis.)

266.61

458.65

523.75

201.97

682.38

783.98

726.81

271.93

430.48

145.24

170.56

164.87

CS US (f+l)

491.31

845.20

965.17

372.20

1257.50

1444.72

1339.36

501.11

793.30

267.50

314.30

303.83

CS World (f+l)

1005.29

1774.22

2148.18

818.52

2697.24

3201.73

2986.22

1156.21

1971.80

680.76

847.33

825.88

(b.2) CS—NPV—float vs. baseline

(3)

NPV‐CS/yr.

160.95

282.43

329.04

129.48

446.46

523.60

495.60

189.35

306.17

105.53

126.62

125.09

(4)

NPV‐Cons. Surp.

(5)

NPV‐CS/yr.(flat+lam)

296.60

520.46

606.35

238.60

822.74

964.89

913.30

348.90

564.22

194.47

233.34

230.52

NPV CS (f+l).

(6)

NPV‐CS‐World/yr.(flat+lam)

606.89

1092.53

1349.56

524.72

1764.71

2138.35

2036.28

805.12

1402.40

494.63

629.03

626.60

CS world (npv62, 92$bn)

CS world (npv92, 92$bn)

(c) Full Consumer Surplus Calculation—No float

(c.1.1) Shipments projections—No float—(GDP growth rate: plate 1969–92; sheet 1974–92)

(14)

US GDP (1987 $)

2877.1

2875.8

2959.3

3107.1

3268.6

3248.1

3221.7

3380.8

3533.3

3703.5

3796.8

3776.3

(10)

Plate (proj.)

221135.2

221035.2

227453.1

238813.1

251226

238813.1

251226

249650.4

271571

284652.6

291823.7

290248.1

Sheet (proj.)

123534.5

102392.8

112108.4

113565.7

103552

102902.5

102066.2

107106.6

111937.9

117330

120285.8

119636.4

Total (proj.)

344669.5

323428

339561.5

352378.8

354778

352552.9

349687.4

366956.3

383508.9

401982.6

412109.5

409884.4

Plate (proj.)

684.6

698.9

734.5

787.7

846.3

858.9

870.1

932.5

995.4

1065.6

1115.7

1133.4

Sheet (proj.)

1161.0

1069.7

1188.8

1197.5

1125.9

1141.4

1154.9

1236.4

1318.2

1409.6

1474.3

1495.9

Total (proj.)

1845.6

1768.6

1923.3

1985.2

1972.2

2000.3

2025.1

2169.0

2313.6

2475.2

2590.0

2629.2

(c.1.2) Price projections—No float—(plate 1969–92, sheet 1974–92)

avg (3yr)

(11)

rate plate %

avg (3yr)

109.43

104.17

98.81

94.95

93.71

rate sheet %

−1.98%

(8)

Plate (proj.)

323.00

316.26

309.65

303.19

296.85

290.66

284.59

278.64

272.83

267.13

261.55

256.09

(9)

Sheet (proj.)

106.41

95.72

94.31

94.84

91.97

90.15

88.37

86.63

84.91

83.24

81.59

79.98

(12)

Inc price redn‐plate

5.18

6.74

6.60

6.47

6.33

6.20

6.07

5.94

5.82

5.70

5.58

5.46

Inc price redn‐sheet

2.86

1.82

1.78

1.75

1.71

1.68

1.65

1.61

(c.2) CS projections—undisc.—no float vs. baseline

CS‐Plate

81.53

106.20

101.54

99.74

99.73

99.67

93.83

87.90

86.81

85.06

83.21

79.22

CS‐Sheet

0

0

0

0

68.57

38.88

36.61

34.29

33.87

33.18

32.46

30.90

CS‐Total

81.53

106.20

101.54

99.74

168.30

138.55

130.44

122.19

120.68

118.24

115.68

110.12

CS‐Total (f+l)

150.25

195.71

187.12

183.30

310.15

255.33

240.37

225.17

222.40

217.90

213.17

202.93

CS‐World (f+l)

307.43

410.82

416.46

404.19

665.24

565.85

535.93

519.55

552.78

554.22

574.69

551.61

(c.3) CS projections—NPV—no float vs. baseline

CS/yr‐plate

49.22

65.40

63.79

63.94

65.25

66.57

63.98

61.21

61.74

61.80

61.78

60.10

CS‐plate

CS/yr‐sheet

0

0

0

0

44.86

25.97

24.96

23.88

24.09

24.11

24.10

23.45

CS‐sheet