Hill 1111 312930010 ‘V V__ «WW 1 111 z. 1.11- i’ .5! ’11! '11 3": 3.1.. 9.. : 1 I 3 - Y. A ’5 l; ". ig‘c -. 9!. ,4.-. This is to certify that the thesis entitled PRODUCT INNOVATION AND DIFFUSION IN THE WOOD-BASED PANEL INDUSTRY presented by Larry Alan Leefers has been accepted towards fulfillment of the requirements for Ph. D. degreein Forestry Z2163: 2f Qfiawfig‘yr‘ Major professor 0-7639 OVERDUE FINES: 25¢ per du per item RETURNING LIBRARY MATERIALS: Place in book return to remove charge from circulation records Ifgvglyz golm INNOVATION AND PRODUCT DIFFUSION IN THE WOOD—BASED PANEL INDUSTRY By Larry Alan Leefers A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Forestry 1981 ABSTRACT PRODUCT INNOVATION AND DIFFUSION IN THE WOOD-BASED PANEL INDUSTRY By Larry A. Leefers The diffusion of particleboard and southern pine plywood is exam- ined in this study. The effects of diffusion on wood requirements are analyzed along with the roles of various factors affecting the diffusion of panel product innovations. Particleboard wood requirements are projected to increase from 5.4 million tons in 1979 to between 9.3 and 10.5 million tons in 2000. Over the same period, southern pine plywood wood requirements are pro- jected to increase from 4.5 million tons to between 6.8 and 7.7 million tons. The projections are based on (1) an aggregate panel consumption submodel, (2) a logistic function submodel, and (3) average wood require- ments per unit of output. Though aggregate panel consumption is a function of expected econ- cnnic activity, diffusion of specific panels is based on many interacting :factors, many of which are not amenable to quantitative analysis. Among true more important factors are: panel characteristics, building codes arui standards, process innovations, economic variables, raw material pxnice and availability, and competition from other wood—based panels. Two diffusion submodels, the logistic function and the Gompertz (nirve, are used to estimate potential market shares and growth rates for the panels. However, these models are not useful for explaining the diffusion process. Multivariate models were utilized to provide a partial explanation for this process. Southern pine plywood's rapid diffusion was enhanced by (1) its similarity to existing low quality softwood plywood, (2) its accepted standardization coinciding with initial production, (3) its low price relative to western softwood plywood, and (4) its regional advantage. The large amount of raw material available for processing promoted the establishment of the southern pine plywood industry. Factors which will eventually slow its diffusion include: (1) almost total reliance on one end-use, sheathing, (2) rising stumpage costs, and (3) competi- tion from new structural panels. Particleboard's diffusion was aided by its many end-use applica- tions, its declining relative price, and the large amounts of residues available for processing. The slow evolution of standards and particle- board's atypical characteristics initially hindered its diffusion. As new specialized products capture portions of particleboard's end— use markets and as competition for raw materials increases, its mar- ket share growth will continue to slow. ACKNOWLEDGEMENTS This study was prepared for the U.S.D.A. Forest Products Labora— tory. Without the funding they provided, this undertaking would not have been possible. I am especially indebted to Dr. Christopher D. Risbrudt who acted as Principal Scientist for FPL and whose disserta- tion on technological change provided the impetus for this study. I am grateful to Dr. Robert S. Manthy who provided insight and advice over the past few years. I also thank Dr. John Burde for whet- ting my interest in forest economics and guiding me to Michigan State University. I thank Judie Shepherd for typing the rough draft and Jill Manthy for typing the final copy. To my wife, Becky, Je t'aime et merci. To Karin, our daughter, goes my thanks for being so entertaining. ii TABLE OF CONTENTS LIST OF TABLES ............. . ..... ..... .......... . ............... vi LIST OF FIGURES .. ...... . ................ . ....................... vii SUWARYANDCONCLUSIONS 0......0.0.0.0.00.0.........OOOOOOOOOOOOO 1 INTRODUCTION .. .............. ..... ............ . .................. 17 The problem ..... ....... . ..... .............. ........ ........ 17 Objective of the study .. ........ ... ........................ 19 Scope of the study ......................................... 19 Definitions and concepts ................................... 22 METHODS OF EXAMINING INNOVATION, PRODUCT DIFFUSION, AND ITS EFFECTS ....... . ....................................... 23b Interacting factors ........................................ 23b Characteristics of goods ............................... 27 Process innovations ................................... 28 Building codes and standards .......................... 29 Economic and marketing factors ........................ 30 Raw material situations .............. ........ ... ..... . 32 Models .. .......... ..... .................................... 33 Aggregate panel market submodel ....................... 34 Market share submodels ................................ 40 Estimation procedures .......... ... ................ .... 45 Methods of estimation ..... ...................... . 45 Completeness and identification .................. 48 Autocorrelation ................. ....... . ......... 49 iii Multicollinearity .............. ..... . ........ ..... 50 Significance tests ................................ 51 Specification errors .............................. 51 PAST TRENDS IN INNOVATION AND PRODUCT DIFFUSION .................. 53 Product characteristics ..................................... 53 Particleboard characteristics .......................... 54 Southern pine plywood characteristics ....... . .......... 55 Process and product trends .................................. 56 Particleboard process and product trends ............... 56 Southern pine plywood process and product trends .. ..... 63 Product standards and building codes ........... ... .......... 68 Particleboard standards .................. . ............. 69 Plywood standards ...................................... 70 Building codes ..... .................................... 71 Product consumption, price and end-use trends ............... 72 Particleboard consumption, price and end-use trends .... 72 Southern pine plywood consumption, price and end- use trends ........................................... 76 Raw material situation ................................... ... 79 Particleboard raw material situation ................... 79 Southern pine plywood raw material situation ..... . ..... 85 MODELLING PRODUCT DIFFUSION ...................................... 92 Aggregate panel market submodel ............................. 92 Market share submodels ...................................... 103 Logistic function and Gompertz curve submodels ......... 103 Multivariate diffusion submodels ....................... 108 PROJECTIONS AND NEWER PRODUCT INNOVATIONS ........................ 118 iv Comparison of logistic function and Gompertz curve for projecting market shares ....... . ...................... 119 Projections of panel consumption and wood requirements ...... 127 Future particleboard trends ....... . ..................... 134 Future southern pine plywood trends .................... 136 New product innovations ..................................... 138 Waferboard or structural flakeboard .................... 140 Medium density fiberboard .............................. 146 Composite panels ....... . ........ . ...................... 149 Other factors influencing panel diffusion ...... ... .......... 151 APPENDIX: DATA USED IN ECONOMETRIC AND DIFFUSION MODELS ......... 155 REFERENCES CITED ................................................. 202 Table 1. Table 2. Table 3. Table 4. Table 5 Table 6. Table 7. Table 8. Table 9. Table 10. Table 11. LIST OF TABLES --Projected total wood requirements for particleboard and southern pine plywood at high, medium, and low levels of disposable personal income, 1990 and 2000 ...... 2 —-WOod-based panel product market share (diffusion) and price trends, 1969 and 1979................. ......... 8 --Estimated parameters of the multivariate, temporal diffusion submodel for particleboard and southern pine plywood using seemingly unrelated regression analysis ..... . ..... . ....... . ..... ............... ........ 10 -—Summary of factors influencing particleboard and southern pine plywood consumption and diffusion.........13 .--New capital expenditures for plant and equipment in the particleboard (SIC 2492) industry, 1972 and 1977 (in millions of dollars)... ............. . .......... 60 -—U.S. resin consumption by type used in the manufac— ture of other platen type particleboard, selected years (in millions of pounds)... ..... ...................62 --New capital expenditures for plant and equipment in the softwood veneer and plywood (SIC 2436) industry 1972 and 1977 (in millions of dollars) ........ . ......... 65 --Southern pine plywood producers, capacity and plant startup years by city and state, 1963 to l968.........66-7 --U.S. particleboard production, imports, exports, apparent consumption, and price index, 1955-1979 (in millions of square feet, 3/4-inch basis).... ...... 73-4 -uPercentage of total particleboard industry produc- tion by end—use, 1973 (in percent) ....... ....... ........ 75 -Particleboard consumption in manufacturing in the United States by product group, 1960, 1965, 1970, and 1976 (in millions of square feet, 3/4-inch basis)...77 vi Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. —-U.S. southern pine plywood production and price index, 1964-1979 (in millions of square feet, 3/8-inch basis)... .78 --Type and percentage of raw material used by the particleboard industry, l973........................81 --Delivered costs of resins and total materials, containers, and supplies cost to the particle- board (SIC 2492) industry, 1972 and 1977............83 —-Production workers, hours worked, and average hourly earnings in the particleboard (SIC 2492) industry, 1972 through 1977.........................84 --Softwood sawtimber net annual growth to annual removal ratio on forest industry and all come mercial timberland in the South 1962, 1970 and 1976.00.00.00. ..... 00.... 0.0.0.0... ...... 86 ——Comparison between southern yellow pine plywood production and log consumption, l964-l978.........88-9 --Wood-based panels and selected end-use markets ...... 93 --Estimated linear and log-linear structural demand equations based on product prices and income........ --Estimated parameters of linear time trend for southern pine plywood and particleboard............104 --Estimated paramters of the logistic function and Gompertz curve for particleboard and southern pine pIYWOOdOOOOOOOOOOOOO0.0.0.00000000000105 --Estimated structural demand and supply equa- tions for particleboard............ ................ 109 --Estimated structural demand and supply equa- tions for southern pine plywood....................110 --Estimated diffusion parameters of the multivariate, temporal submodel for particleboard...............ll4 --Estimated diffusion parameters of the multivariate, temporal submodel for southern pine plywood.......115 --Market share projections based on nonlinear diffusion functions, 1990 and 2000 (in percent)....120 --Estimated parameters of the logistic function and Gompertz curve for particleboard, selected years-0.00.000... ..... cocoon 00.00.123-4 vii viii Table Table Table Table Table Table Table Table Table 28. 29. 30. 31. 32. 33. 34. 35. 36. --Estimated parameters of the logistic function and Gompertz curve for southern pine plywood selected yearSOOOOOOOOOOO......OOOOOOOOOO0.0......00.000.00.000125-6 --Comparison of particleboard's market share poten- tial to the last market share achieved in selected samples..... ..... .....IOOCOOOOOOCO ........ 00.0.000000000128 --Comparison of southern pine plywood's market share potential to the last market share achieved in selected samles.........O0..........0.0.0.0000000000000129 --Projections of panel consumption and total wood requirements for high, medium, and low levels of disposable personal income, 1990 and 2000...............131 --Comparison of particleboard consumption projections made in this study with projections by the U.S. Forest Service (in millions of square feet, 3/4-inch basis.....l32 —-Comparison of southern pine plywood production pro- jections made in this study with projections by the U.S. Forest Service (in millions of square feet, 3/8-inch basis).........................................133 --New mills and modernization projects scheduled for completion in 1979 and announced specific projects requiring capital expenditures........................141-2 --Medium density fiberboard production, 1975-1979 (in millions of square feel, 3/4-inch basis)............l48 —-Examp1es of products substituting for wood-based panels in various end-uses.. .............. ..............154 ix LIST OF FIGURES Page Figure l.--Roles of qualitative factors in promoting (+) or hindering (-) particleboard and southern pine plywood diffusion over time. . . . . . . . . . . . . .12 Figure 2.--Product life cycle . . . . . . . . . . . . . . . . . .25 Figure 3.-—Submodels requiring exonometric analysis and the purpose for using the submodels. . . . . . . . . . . .35 Figure 4.--Total panel, non-southern softwood plywood, southern softwood plywood, and particleboard consumption, 1955-1979. . . . . . . . . . . . . . . . . . . . . .97-8 Figure 5.--Total panel, hardboard, hardwood plywood, insula- tion board, and medium density fiberboard consump- tion, 1955-19790 0 o o o o o o o o o o o o o o 0 099-100 Figure 6.--Market share of particleboard as a percent of total panel consumption, 19.55-1979 o o o o o o o o o o o o 102 Figure 7.-—Market share of southern pine plywood as a percent of total panel consumption, 1964-1979. . . . . . . . 102 Figure 8.--Logistic function fitted to particleboard market share data, 1955—1979. . . . . . . . . . . . . . . . 106 Figure 9.--Gompertz curve fitted to particleboard market share data, 1955—1979. . . . . . . . . . . . . . . . 106 Figure.flL-—Logistic function fitted to southern pine ply- wood market share data, 1964-1979. . . . . . . . . . 107 Figure llu—43ompertz curve fitted to southern pine plywood market share data, 1964-1979 . . . . . . . . . . . . 107 Figure 12.--Partic1eboard market share projections based on the logistic function, 1980-2000 . . . . . . . . . . 121 Figure 13ouParticleboard market projections based on the Gompertz curve, 1980—2000. . . . . . . . . . . . . . 121 Page Figure l4.--Southern pine plywood market share projections based on the logistic function, 1980-2000. . . . . .122 Figure 15.—-Southern pine plywood market share projections based on the Gompertz curve, 1980-2000 . . . . . . .122 xi SUMMARY AND CONCLUSIONS The primary objective of this study is to analyze the effects of product innovations and diffusion on raw material requirements in the wood-based panel products industry. In addition, the roles of economic and other factors in determining the direction and dif- fusion rate of panel product innovations are examined. Particleboard and southern pine plywood are the primary panels analyzed in this study. Total wood requirements for these panels are projected to increase substantially by the end of this century (see Table 1). When compared to 1979 levels, particleboard wood require- ments and the corresponding consumption levels are projected to increase between 49 and 61 percent by 1990 and between 71 and 93 percent by 2000. Similarly, southern pine plywood wood requirements and consumption are projected to increase between 33 and 44 percent by 1990 and between 53 and 72 percent by 2000. Particleboard and southern pine plywood consumption and wood requirements were projected based on: (1) an aggregate panel con- sumption submodel, (2) a logistic function submodel, and (3) average wood requirements per unit of output. U. S. disposable personal in- come and time were used as exogenous variables in the submodels. The projections were based on extrapolation of past diffusion trends and on expected economic activity. 2 Table 1. Projected total wood requirements for particleboard and southern pine plywood at high, medium, and low levels of disposable personal income, 1990 and 2000. Year Income Wood requirements level Particleboard Southern pine plywood 1/ (thousand tons)— 1990 High 8710 6452 Medium 8388 6213 Low 8045 5960 2000 High 10464 7703 Medium 9844 7245 Low 9270 6823 1/ -In thousands of tons, dry weight (ovendry)basis for particle- board and in thousands of tons, based on ovendry mass and volume at 12 percent moisture content for southern pine plywood. Source: Table 31. 3 The aggregate panel consumption submodel was estimated for projecting total panel consumption in 1990 and 2000. The model included all wood-based panels for which consumption and production data were available and covered all panel end-uses. Secondary time series data were not available for examining year-by-year consumption of panels in specific end-uses such as new housing, residential upkeep and improvement, manufacturing, and new nonresidential con— struction. This is a limitation of the submodel since penetration into different end-use markets is likely to vary for a given panel. Total panel consumption was estimated to be a positive function of U. S. disposable personal income, an important determinant of wood- based panel products, demand household formation, and furniture con- sumption. The submodel fits the long-term trend very well, but performs poorly during periods of great market fluctuations experi- enced during the 1972 to 1979 period. The strong reliance of wood— based panels on the construction industry led to this cyclic pattern of panel consumption. The submodel is not structured to project business cycles; instead, it is used to estimate long-term trends in panel consumption based on different levels of disposable personal income. The impacts of changes in family structure, family formation, and possible sub— stitution of other products for panels are not incorporated in this univariate submodel. The diffusion of new products can most readily be seen in terms of market shares. Market share, that is, a given panel's percent of the total panel market, was used in this study as a measure of panel diffusion. Both univariate and multivariate diffusion models were estimated to provide some quantitative information about panel diffusion. Though no two panels will follow the same diffusion path, the diffusion paths of particleboard and southern pine plywood evinced S-shaped patterns. This pattern is supported by diffusion theory which hypothesizes that diffusion follows a sigmoid pattern. That is, panel diffusion starts slowly, accelerates through an inflection point, and slows as the process nears completion. Two nonlinear models, the logistic function and Gompertz curve, were used to estimate this pattern, but yielded different results regarding potential market shares and growth rates. Higher potential market shares and longer periods to grow from 20 to 80 percent of the potential resulted when the Gompertz curve was estimated. Results comparing the two models were more consistent for the southern pine plywood data than for the particleboard data. The growth rate of southern pine plywood was approximately twice as rapid as the particle— board rate. For example, the period of time required to grow from 20 to 80 percent of the potential was 7 years for southern pine ply- wood and 14 years for particleboard based on the logistic function. Several assumptions are implied when the logistic function and the Gompertz curve are used to estimate past diffusion trends and project future trends. First, the functions imply that new products will gain in market share up to some maximum or potential level. Second, the proportion of market share already gained is an important determinant of future levels. Finally, the increase in market share is a function of supply, demand, and other factors. These assumptions highlight some of the major limitations of of simple nonlinear models. Though intuitively a panel will increase in market share only up to a maximum level, there is no independent means for estimating that level. Therefore, market share projections are based on a nonlinear extrapolation of past trends. A number of factors that will influence the potential are examined in this study; however, it is not possible to satisfactorily estimate the extent to which a panel will diffuse in a complex market. Similarly, the models provide no explanation of factors affecting diffusion. In order to assess the usefulness of these functions for pro- jecting trends and potential market shares under different conditions of data availability, five time periods of varying length for particle- board and southern pine plywood were fitted with the logistic function and the Gompertz curve. For most data sets tested, the logistic function yielded more conservative estimates of market share potential than did the Gompertz curve. Thus, model selection affects projections. In all cases where the standard error of the potential was small, over 85 percent of market potential had been reached by the last year of the data series. Therefore, these diffusion functions should be used for projections only when the diffusion process is substantially underway as was the case with particleboard and southern pine plywood. Despite the shortcomings of the nonlinear functions, they are useful for studying diffusion. For example, the S-shaped models represent the diffusion pattern better than simple linear trend models. In addition, they provide an estimate of potential market share and a measure of the rate of panel diffusion. As a result, they can be used to estimate the relative magnitude of future particleboard and southern pine plywood diffusion. While the univariate models are useful in depicting temporal diffusion trends, they are not very helpful in understanding the role supply and demand factors play in determining panel consumption and diffusion. The supply and demand factors influencing particleboard and southern pine plywood diffusion were screened by means of structural supply and demand equations. 0n the demand side, product price and multifamily (3 or more units) housing starts were inversely related to quantity demanded. One and two unit housing starts, mobile home shipments, and furniture manufacturing exhibited a positive relationship in the particleboard case. In the southern pine plywood case, western softwood plywood price and one and two unit housing starts were significant variables with positive signs; increases in these variables are expected to lead to increases in southern pine plywood demanded, all things being equal. In the particleboard supply equation, adhesives price and wages had negative signs while lumber production and energy price had posi- tive signs. The negative signs were expected since increases in input prices lead to shifts in the supply curve, ceteris paribus. Lumber production was used as a proxy for raw material (wood furnish) avail- ability and had the expected positive relationship to quantity supplied. For the time period analyzed, the net effect of energy prices on the quantity of particleboard supplied was positive indicating that com- peting products were adversely affected by price increases to a greater degree. The most significant supply factor in the southern pine ply- wood supply equation was mill productivity. As productivity increased, quantity supplied increased holding other variables constant. 7 The same factors which influence supply and demand for a panel also influence its diffusion. Table 2 presents data relating market share and price trend data for major wood-based panels since 1969. Apparent domestic consumption of western softwood plywood, hardwood plywood, and insulation board remained relatively stable over 1969 through 1979 period indicating they are mature products. Hardboard, particleboard, and southern pine plywood consumption more than doubled over the same period; they are substantially into market share growth stage and are approaching maturity. The increased price of western softwood plywood and competition from newer panel products contributed to its decrease in market share. Lack of suitable domestic wood supplies, competition from other panels, and fast growth of newer panel segments led to a decrease in hardwood plywood's market share. The relative flexibility and greater diverstiy of end-uses for competing panels accounts for the decreased market share for insulation board. Particleboard and hardboard increased in market share and decreased in relative price over the 1969 through 1979 period. This inverse relationship corresponds to the expected demand price-quantity rela- tionship and is further corroboration of the role demand factors can play in panel diffusion. Southern pine plywood's market share-price relationship is the same when viewed relative to western softwood plywood, a substitute in the sheathing market. Lagged market share variables along with the supply and demand variables discussed above were tested as explanatory variables in a temporal multivariate market share model. The lagged market share variables were included to show the accumulated effects of supply, xfiwaoma< "mousom .ooHumomH .xmccfi mowua umoavoua moHunoSEoo Ham ou o>wumaom.l \N .Ucouu mmmfi nwsounu momfi\fl mcfimmmuoom wcfimmouoma campy mm m.m mead OOH N.mH acma memos coaumasmcH wcflmmouoon wcammouoon wcopH an q.m memfi ooH q.¢fi mead ooo3maa woo3wumm wcflmmmpoma mcfimmouocH ocouH mm N.mN mnmfi OOH m.m~ moofi @umononmm wcflmmouooc wcfimmouocH ozone me o.¢ mnmfi OCH N.o momfi wumonoaofiuumm wcwumSuosam wcfimmouocH wcouh cm m.o~ mmmfi col m.os moms ceasefia mafia anmnosom mcwmmmuocH mcfimmouoon vcmuH omH m.m~ memH oofi ~.mm :moma woozkaa poozumom :uoumoz Apnoea mam xmvcfiv chouu paw unmouoav \Moofium oumcm noxumz \Mficouu\ummw maxu Hoamm .mmofi cam mcmfi .mwcouu mowua mam AcowmdMMfiov mumsm Dozens uoswoua Hocma ommmnloooz .N magma 9 demand, and qualitative factors on product diffusion. The seemingly unrelated regression results presented in Table 3 were not signifi- cantly different from the ordinary least squares results. The analysis indicates that past market share was a significant variable in the diffusion of product innovations. For southern pine plywood and particleboard, end-use market demand factors were also significant explanatory variables. Particleboard market share growth was a positive function of furniture and mobile home market growth. Inter- estingly, southern pine plywood market share growth was greatest during periods when housing starts declined. The spatial aspects of diffusion were examined for southern pine plywood in relation to western softwood plywood. A model relat- ing the percent of softwood plywood shipments to 29 Rand-McNally Major Trading Areas was formulated. As the distance from a Pacific North- west supply point (Portland, Oregon) to the Major Trading Areas increased and as the number of southern plywood plants near the Major Trading Areas increased, the southern pine plywood market share in- creased. Transportation cost is one of the main factors leading to these results and to southern pine plywood's regional advantage in the eastern United States. A number of qualitative factors that are not amenable to sta- tistical modelling play an important part in promoting and hindering product diffusion. Among the more important fattors are: character- istics of the good, building codes and standards, process innovations, and raw material price and availability. Other factors such as non- wood product competition, product promotion, research and development activities, and government policies, also influence product diffusion, .mN “Sum QN mMHflmn—w "mQUHDOm 10 .Ho.m u .3.o .om. u mm .mmmH nwsou£u «omfi Baum mumv co momma moumafiumM\M .om.~ n .3.9 .mm. H mm. .mmmfi nwsounu «omfi Scum mumv do momma moumafiumM\fl Hmmwoo. Nomfimo. ucmumcoo mooooo. mmoooo. I N w H .mupmum wcfimsom loumnm uoxuma moshaa Omnooo. oqqqoo.H oumnm uwxuma wowwmg mafia aposusom oeeooo. NNQNHO. I ucmumcoo mooooo. mmoooo. mucoaawnm was: oHHnoz «moooo. «mmooo. wafiuauommscma mouSuwa cam opsueausm mo woocH \Ioumnm poxums ammoko. ommmsk. spasm umxpma umwwmo mumonmauwucmm a mo pouuo Aav uaofiofimmooo Axv moanmfium> va oHanum> oumvamum coamwouwmm uuoumcmaaxm unopaoaoa .mfimhamam scammouwou omumaoua: hawafiaoom wch: woosham mafia aumzunom one vumonmaoeuuma now Houseman scamsmwwo Hmuoaaou .oumaum>fiuasa osu mo muouoamuma woumEaumm .m manna 11 but are beyond the scope of this study. There are similarities and differences in these factors and their roles in the diffusion process. The qualitative and quantitative factors presented in Table 4 provide a basis for explaining differences in particleboard and southern pine plywood diffusion. In addition, the factors provide a framework for examining other new products. Figure 1 summarizes the effects of qualitative factors on the diffusion of particleboard and southern pine plywood over the product life cycle. Notable similarities between the panels were the testing and modifying of production processes and the large amount of raw materials potentially available for processing before and during the product introduction stage. These factors will undoubtedly be important to future product successes. Southern pine plywood's rapid diffusion was enhanced by (1) its similarity to other existing low quality softwood plywood, (2) its accepted standardization coinciding with initial production, (3) its low price relative to western softwood plywood, and (4) its regional advantage. To a large degree, these factors led to southern pine plywood's broad acceptance in the marketplace. Factors which may eventually slow its diffusion include (1) almost total reliance on one end-use, sheathing, (2) rising sawlog costs, and (3) competition from new structural panels. Particleboard diffusion was somewhat slower than southern pine plywood's market share growth. Particleboard's unique properties have led to wide acceptance over time, but during its introductory stage the uniqueness was a hindrance. A multitude of end-use appli- cations and declining relative prices have aided particleboard's 12 mafia cuonusom mam wumonoaofiuumm AIV mcfiumocwz no A+V wcfiuosoua ca muouomm o>aumufiamow mo moHom||.H ouswflm .oEHu uo>o comeMMHw woosmmm wHHMDHoaca mmoooum wooskam mafia cuonusom wumonoaoauumm mwoom mo mofiumwuouomumao 13 eweuae>ve :oHu leooH HeSOfiweu .hufi>fiuoe uexuea emnlwae .mmeoodm hufi>fluoe uexuea eealoae muouoem Homes nexuea veueasasoo< .mmeoosm nexuea veueaaaboo< Imam GOHmSMMHQ va hufi>fiuo=woua .meofiua undue .hufi>fiuoe >uwafiaeafie>e movemep .meofiua nexuea emulwce .mooflun uaacfi .hua>fiuoe uexuea omslvso muouoew Howoa euaufiumnsm .eofiua nonvoum .meoHua ousuaumnnm .eoeum uoswoum waeaew wee haaa5m Amy waaeafiwfiuo eHQeHHe>e meOum haaecwwfiuo cowuesuam nonawuaem mo meanao> owned maneawe>e mesvfimeu e>Hmmez Heeueuez Bem on wafiuesuonam waflafiaoev kaaepecoo eofium Amv uaoa%eauewas wcwfiueegm .xooumeuoo eunufiausm mew: >ueawum Aqv fioueemeu :oueeeou useaaue>ow one huumsocfi uceacue>ow ode >uumsoafi ”onMHwoa one veuuanH “wefimacoa one weuuomaH meemeooum Amv coaueofimwmmeao euefiweeaH maowueoflMfivoE Heaveuu mvuewaeum ANV voo3maa hufiaenw haaecfiwfiuo Heuoepeso 30H pesuo ou uceHe>Hsvm meeuuoaoum enced: nonvoum AHV wooahaa mafia cuesusom vueoneaofiuuem uouoem .aOHmDMMfiv wce coaumasmcoo woo3kaa mafia cuesusom one vueoneaofiuuea waaocesawcfi euouoem mo hueaesm .s wanes 14 diffusion. In general, if a panel has product characteristics similar to those of an existing product, diffusion occurs more readily. This phenomenon is due to decreased consumer resistance to a familiar pro- duct. For example, waferboard, medium density fiberboard and com- posite panels are more acceptable because particleboard and hardboard had gained widespread approval prior to their introduction. Eventu- ally, the physical characteristic limitation of a panel will slow its diffusion in the total market. That is, it will diffuse only through end-use markets for which it is suited. The development of process innovations is important in all stages of a product's life. Major innovations may lead to the formation of new industries, and "nuts and bolts" innovations during the market growth and maturity stages may keep a panel competitive with its substitutes. Overall, process innovations promote the diffusion of products. Formal standards promoted the acceptance of particleboard and southern pine plywood. They were especially helpful in the case of southern pine plywood for which standards existed when the first panels were introduced. More rapid acceptance occurred as a result. For particleboard, the slow evolution standards initially hindered its diffusion. In general, early development of standards and code approval will aid in the diffusion of newer panel products. The availability and price of wood inputs are also important supply factors influencing the past direction of the panel products industry. Large quantities of relatively inexpensive manufacturing residues spurred the establishment and growth of the particleboard 15 industry. Likewise, the high price and reduced availability of Douglas- fir peelers contributed to southern pine plywood's establishment and growth as a panel product. The volume of southern pine sawtimber grow- ing stock was equally important. Over time, increased cost and decreased availability of wood inputs will hinder the diffusion of a given panel. However, if the capital is available for investment, the existence of large quantities of raw materials can promote the establishment of panel industries and lead to the introduction of new products. In conclusion, projections are difficult to make when large amounts of pertinent information are available and even more difficult when little is available. One example of a projection by an industry expert based on few trend data is the following: ...talk about dozens of new plywood plants in the South during the next decade does not seem justified. Land ownership prob— lems, availability of raw material on lands dedicated princi- pally to other products, and the as yet unfelt price competi- tion from the massive western industry all seem to militate against over-rapid expansion. A total of perhaps 8 or 10 mills during the coming decade may be a more realistic estimate (Fassnacht, 1964). This projection was surpassed by the end of 1965 when the twelfth plant started production. Due to the complexity of the diffusion process and lack of data on the newest products, estimates on the magnitude of their diffusion would only be speculative. However, some general statements can be made regarding the direction of product innovation and its effects on wood requirements based on the review of qualitative factors. First, new products will utilize more diverse raw materials and in greater quantities than has been the case in the past. Waferboard and composite board are ample evidence of this trend. Roundwood en- gineered into flakes, particles, and strands by equipment, such as 16 the chipping lathe headrig, is becoming more widespread. Increased production, particularly of waferboard, indicates new panels can be competitive even during periods of low market activity. Second, many products that will play a significant role in satis- fying panel product demand at the turn of the century are just becom- ing known or may still be on the drawing board. The gains by particle- board and southern pine plywood are cases in point. Raw material avail- ability was important to the success of those products, and presently underutilized resources may be the key to future successes. Third, the evolution of standards and the awareness of their im- portance has opened the door for future product innovations. As stand— ards become based on performance criteria rather than composition, new panels will be engineered to compete with existing panels. Stand- ards being developed by the American Plywood Association are examples of this type of evolutionary and innovation-promoting standard. Finally, new product and process innovations expand our resource base as they diffuse through the market. The new products, in turn, slow the diffusion of existing, traditional products. INTRODUCTION Technological change is a major factor influencing economic growth and productivity. Technological change and its subsequent diffusion through an industry can shift supply functions through process innova- tions and demand functions through product innovations (Bentley, 1970). The initial effect of a process innovation is a new relationship be- tween inputs and outputs, whereas, the initial impact of a product innovation is the substitution of one product for another. Over time, innovations lead to significant changes in raw material and other input requirements and extensive substitution of new products in various end-uses. The interaction of social, political and economic factors influ- ences the rate and magnitude of diffusion. In order to get more insight into the complex process of the diffusion of product innovations, the relationships between these factors must be analyzed in their historic context . The Problem In 1974, Congress enacted the Forest and Rangeland Renewable Re- sources Planning Act. This act requires the United States Forest Ser- vice to periodically assess the present and anticipated uses of our Nation's renewable forest resources. Research Work Unit 4151 at the United States Forest Service Forest Products Laboratory is responsible 17 18 for projections of timber requirements and trends in harvesting, pro- cessing, product design and end—use technology of wood products. As such, they are interested in trends of advancing technology in the wood-based panel products industries and their effects on the relation- ships between wood product consumption and timber use. The impact of product innovation on raw material requirements has not been ade- quately assessed in the past. The purpose of this research study is to investigate factors influencing product innovation and diffusion in the wood-based panel products industries and the effects of these changes on raw material requirements. Presently, the Resources Planning and Assessment Staff Unit uses historical data as a base for resource demand and supply projections. This implicitly assumes a continuing stream of technological changes and innovations for the future based on little specified knowledge of factors affecting the diffusion of innovations in the past (U.S. Department of Agriculture, Forest Service, 1977). By identifying forces that have historically influenced the diffusion of innovations, the forces can be incorporated in future models. Two major forest product innovations requiring wood inputs are particleboard and southern softwood plywood. Particleboard was an obscure panel product in the early 1950's, but over 6.3 billion square feet (3/8-inch basis) were consumed in 1976 (U.S. Department of Agri- culture, Forest Service, 1980). Southern pine plywood comprised zero percent of United States softwood plywood production in 1964; yet, in 1976 it had captured over 36% of U.S. output (Anderson, 1979). Fac- tors influencing these product innovations and their diffusion will be useful in assessing newer panel products such as medium density 19 fiberboard, composite board and waferboard. Objective of the Study The primary objective of this study is to analyze the effects of product innovations and diffusion on raw material requirements in the wood-based panel products industry. The secondary objective is to hypothesize the future roles economic and other factors will play in determining the direction and diffusion rates of panel product in- novations. Framework for Analysis Product innovation and diffusion are complex phenomena involving both economic and noneconomic factors (Warner, 1974). Economic factors include product prices, input prices, substitute prices and end-use market activity. Product characteristics, standards and building codes are examples of noneconomic factors. Due to this complexity, several approaches for examining innova— tion, diffusion and their effects are useful. First, process innova- tions, product innovations and end-uses (existing and potential) can be described. Second, economic, social and physical input trends that may influence the rate and direction of innovation and diffusion can be identified and evaluated. Finally, mathematical models can be used to relate trends in these factors to innovation, diffusion and raw material requirements. The first two approaches are qualitative in nature. Major process innovations can be described and classified as to (1) year of adoption, (2) adopting company, (3) location, and (4) origin of process. By 20 examining these traits, a greater understanding of innovation and dif- fusion can be gained. This information is enhanced when coupled with trend data on inputs, production, prices, product standards and other important factors. Though there is presently insufficient knowledge of important factors to adequately forecast future technologies and their effects in the wood-based panel products industry, mathematical models can be used to quantify some effects of existing product innovations. Pro- duct consumption, price and raw material requirements are some major effects of concern to forest economists. The analytical approach most commonly used by economists is to model supply and demand, separately or simultaneously (McKillop, 1967; McKillop, Stuart, and Geissler, 1980; Mills and Manthy, 1974). These models provide estimates relating product consumption and production to various factors including price, income and population, but they do not directly yield estimates on the rate of adoption and potential market share of new products. Econometric models for particleboard and southern pine plywood supply and demand are used in this study to corroborate the role of economic variables in product diffusion. Mathematical diffusion functions, on the other hand, have been widely used to model the diffusion process. The modified exponential function, the Gompertz curve, and the logistic function are examples of diffusion functions used by researchers (Lekvall and Wahlbin, 1973). The functions are used to fit empirical observations and to make pre- dictions concerning developments in diffusion processes already underway. Selection of the appropriate diffusion model or functional form poses a problem and is decided on the basis of observed data (Maddala, 21 1977). A considerable amount of empirical research supports the use of the logistic function to describe the diffusion process for new products. Since observed market share data for particleboard and southern pine plywood follow an S-shaped pattern, the logistic model will be utilized in this study because it provides information on the rate and magnitude of product diffusion. Results based on the Gompertz Curve are also presented. The models assume that diffusion follows a sigmoid pattern starting slowly, accelerating through an inflection point, and slowing as the diffusion process nears completion. The diffusion models are used to compare the diffusion of particle— board and southern pine plywood. The purpose in studying these products is to compare their rates of adoption, potential market shares, and wood requirements along with differences and similarities of factors affecting the diffusion process. This information will be used to hypothesize on the roles these factors will play on other existing and future wood-based panel products. Scope of the Study A number of important product innovations have occurred in the wood-based panel products industry since Werld War II. This study focuses on two product innovations so that the study will be manageable and because time series data are available only for a limited number of innovations. Rather than studying the products of individual firms, aggregate product forms and their related wood requirements are ana- lyzed. Data utilized in this study are primarily from various Census of Manufactures, Current Industrial Reports, and forest products 22 publications. The product forms studied are particleboard which corre- sponds with Standard Industrial Classification (SIC) 2492, Particle- board, and southern softwood plywood which is a component of SIC 2436, Softwood Veneer and Plywood. The time frames are years 1955 to 1979 for particleboard and 1964 to 1979 for southern pine plywood. These product forms have shown significant growth in recent years. The number of plants producing particleboard in North America increased from 23 in 1955 to 72 in 1979. Particleboard production increased from an estimated 140 million square feet (3/8-inch basis) to over 7.2 billion square feet during this period. Three plants were responsi— ble for the production of 80 million square feet (3/8-inch basis) of southern pine plywood in 1964. Seventy-two southern plywood plants produced approximately 7.6 billion square feet in 1979. Southern pine plywood accounted for 41.2 percent of United States softwood plywood production in 1979. Definition and Concepts Economic and management literature are replete with definitions of technology and related concepts. For this study, the following commonly used definitions will apply. Technology is defined as the social pool of knowledge of the industrial arts (Schmookler, 1966). Technological change is regarded as the advance of technology whose rate is determined by new technology produced in any given period. Technique is a specific method of producing a good or service; change of technique is ...the employment of existing but unused technology" (Manthy, 1974). An innovation is defined by the National Research Council (1978) 23 as the introduction of new things or methods. Implementation by indus— try or some other group is required by this definition. In a given industry, the first enterprise to adopt a new technique is considered an innovator. Subsequent adopters are imitators. This study focuses on innovation and subsequent adoption or diffusion. Technological innovations, which are implemented technological changes, are classified in three categories: "nuts and bolts" innova- tions, major technological advances and creation of large new systems (National Research Council, 1978). "Nuts and bolts" innovations are exemplified by product differentiation and minor changes in production processes. Major technological advances, which are examined in this study, provide opportunities for new processes or products. The cre— ation of a large new system may or may not be based on a new technology, but it does require a new combination of techniques in a complex man- ner. Another often cited classification distinguishes product innova- tions from process innovations (Schmookler, 1966). New commercial products are product innovations, whereas process innovations are new methods used for creating products. Major process innovations in the particleboard and southern plywood industries are identified in this study. Products can be viewed as (1) general product classes (e.g. wood-based panels), (2) product forms (e.g. particleboard and south- ern pine plywood), and (3) brands (e.g. GP Particleboard and Plum Creek Fiberboard) (Kotler, 1976). This study analyzes the diffusion of pro- duct forms in the wood-based panel products industry. Market share is analyzed as the indicator of panel diffusion. Hereafter, product forms will be referred to as products. 23a Innovation and product diffusion have a variety of effects. For example, they may cause changes in the nature and quantity of wood inputs utilized. Other effects include factor and product substitu- tion, the creation of new resources, and changes in industry struc- ture. METHODS OF EXAMINING INNOVATION, PRODUCT DIFFUSION, AND ITS EFFECTS In order to assess the effects of product innovations on raw mater— ial requirements in the wood-based panel products industry, an understand- ing of the diffusion process and factors influencing successful diffusion is paramount. Historical examination of interacting factors such as process innovations, products characteristics, building codes and stand- ards, raw material situations, and economic data series are useful in analyzing product innovation and diffusion. In addition, econometric models can be used to supplement our knowledge and to measure some effects of product diffusion. Specifically, the models can be used to estimate future panel consumption, rates of diffusion, market shares, and wood requirements. The rationale for examining specific interacting factors is pre- sented in the following section. Since this is an aggregate analysis, these factors were selected on the basis of economic theory and a liter- ature review (Leefers, 1979). The final section of this chapter provides the basis for the models used and the estimation procedures employed in this study. Interacting Factors The interaction of social, political, and economic factors influ— ences the rate and direction of technological change (Rogers and 23b 24 Shoemaker, 1971). Following the first implementation is the diffusion of the innovation. In this case, the primary focus is on the diffusion of particleboard and southern pine plywood, factors influencing their diffusion, and the effects of diffusion in terms of wood requirements. The magnitude of diffusion is the primary determinant of wood require- ments for the new product. A concept frequently used to describe the path a product follows over time is the product life cycle (McCarthy, 1975). Figure 2 illus- trates this concept in terms of sales. There are four major stages in the life of a product: (1) product introduction, (2) market growth, (3) market maturity, and (4) sales decline. Introduction is a period of slow growth and is followed by a stage of more rapid growth. Matur— ity is a period when sales slow due to widespread acceptance by most buyers. The final stage marks the decline of product sales (Kotler, 1976). This decline is often caused by new products replacing the old. Support for the life cycle concept is based on the hypothesis that new products must overcome customer reluctance or resistance to established purchasing patterns (Buzzell, 1966). The idealized S- shaped product life cycle curve, of course, does not hold for all pro- ducts, and the period of time for each stage may vary dramatically. For example, product classes such as wood-based panels may continue in a growth or mature stage for long periods of time. Since product innovation and diffusion are closely related to existing products and markets, it is important to view the introduction and growth of products in the context of the whole market. New products that will have a significant long-term effect can be expected to grow 25 .eHuNw.eMHH posooum .N euawfim mirh mzHeomm weHmseaz mezome oneoanomezH mmaem Hemmez Hemmez economm SBWVS 26 at a faster rate than the market. The diffusion of new products in the market can most readily be seen in terms of market shares. The product life cycle can be redefined using market share of consumption rather than sales as the variable of interest. Any demar— cation between the various stages is subjective. For this study, the product introduction stage is the period which begins with the product innovation and ends when the panel has reached twenty percent of its potential market share. Though the distinction between the introductory stage and the market share growth stage is somewhat arbitrary, one characteristic of the growth stage is the point of maximum marginal market share growth (inflection point). After this point, market share growth slows. The mature stage is a period when market share initially stabilizes indicating that the growth rate of product form consumption is equal to the growth rate of product class consumption. The mature stage can also exhibit a decline in market share with growing or stable product form consumption. Existing products are expected to exhibit this condition as new products are introduced. The decline stage is identified as a long period during which product consumption and market share decline. The market share associated with a particular product innovation is determined by the interaction of supply, demand and other forces along a path to maximum market share (Griliches, 1957). Since the data available on individual brand diffusion and intrafirm process adoption are limited, this study analyzes the impact of aggregate vari- ables on product form diffusion. 27 Characteristics of Goods Traditional demand theory permits economists to analyze the effects of different preferences on demand, but provides no mechanism for tracing the effect on demand of changes in physical properties of goods (Lancas- ter, 1971). Traditional microeconomic analysis begins with a preference map. At that stage, the characteristics of the goods, other than price, have already been absorbed into the analytical framework. Though the characteristics (e.g., weight, color, size, etc.) may change, this information is rarely used at a later stage in demand theory analysis. Product variants, model changes, and new goods are difficult to incorporate into traditional economic models. Either (1) the changes are ignored or (2) the variant is treated as an entirely new good. In practice, our ability to investigate physical changes in goods is lim- ited by the data available on the altered goods. When changes are ignored, the characteristics of the variant are implicitly absorbed into the aggregate product grouping. For instance, graded density particleboard may be viewed simply as particleboard. In the second case, the variant is treated as a new good which acts as a substitute for the original product. As a substitute, we explicit- ly acknowledge that certain comparable attributes or characteristics link the new with the old. Southern pine plywood, for example, can be treated as a new good, as a substitute for western softwood plywood, and not simply as more softwood plywood. Ultimately, the utility a good will provide establishes the extent to which it will be used vis-a-vis other goods. This utility, in part, is closely related to the characteristics and properties of the good. Theoretically, the demand for new goods could be predicted based on 28 observed behavior of consumers toward existing goods with similar charac- teristics (Lancaster, 1971). In principle technical data concerning product attributes will determine the closeness of substitution. Thus, the characteristics of a particular wood-based panel determine the range of uses to which it is suited and the extent to which it can substitute for other products. Though any good possesses a number of physical properties, not all properties are relevant to choice between goods. From a practical standpoint this situation is further exacerbated by the fact that indi- viduals react differently to specific attributes depending upon their own preferences. Therefore, in an aggregate analysis it is difficult to objectively ascertain which characteristics increase the probability of an innovation's successful diffusion. This is especially true for products that have many properties and a multitude of potential applica- tions. A useful method of incorporating characteristics into traditional economic analysis is to identify relevant physical properties emphasizing those properties that are most pertinent to specific end-uses. In some instances, a product's characteristics may remain unchanged over time. However, knowledge regarding the product and its characteristics will increase in a free market as time passes. This increased knowledge may be due to marketing efforts, word—of—mouth, and other factors. Process Innovations Within a product class, process innovations can chdnge the produc— tion function for existing products or serve as the basis for new product forms and products. The new processes can be developed within or outside 29 the adopting industry and can be imported from other countries. Regard- less of the origin, process innovations have played a major role in defining the characterisitics of panels as they exist today. Many technological innovations have been adopted in the wood- based panel industry. These innovations range from ”nuts and bolts" innovations such as using forklifts for loading plywood into railroad cars (Cour, 1955) to the creation of large new systems such as the Georgia Pacific's Fordyce, Arkansas, plywood plant (Baldwin, 1977). The effects of adopting new processes are reflected in prices of finished goods, raw material utilization, and products available in the market. As is the case with characteristics of goods, process innovations are difficult to incorporate in an aggregate analysis. However, many important process innovations can be identified and their roles described. The speed of adoption of new techniques has been analyzed by Mans- field (1963). His study supports the hypothesis that speed of adoption is a positive function of firm size and expected returns. In more concentrated industries, process diffusion generally is more rapid due to the trialability and observability of the innovation and the high degree of intrafirm communication, ceteris paribus (Rogers and Shoemaker, 1971). Building Codes and Standards "'evolutionary documents' Building codes and product standards are that have been amended to accept new products and test methods as they have been developed (Pease, 1980)." The nature and flexibility of the codes and standards are factors determining how rapidly new products 30 will gain acceptance. If panels do not have code approval, they cannot be used for many potential end-uses (Brown, 1979). If standards are nonexistent, poor quality products will sometimes be produced. The result is bad publicity and fewer sales (Anonymous, 1966a). New standards are created to encompass new products, to recognize more wood species, and to account for changing log quality (Anonymous, 1966b). The standards, in effect, change to accommodate raw material situations, process and product innovations, and the producers' capabil— ities. Economic and Marketing Factors As mentioned previously, supply, demand and other factors influence the diffusion of new products. Product prices and consumption in various end-uses represent the interaction of supply and demand over time (Ris- brudt, 1979). Consumption data can be transformed readily into market share data once the total market and appropriate measurement units are defined. Marketing professionals view the period of growth in product consump- tion and in knowledge about the product as the initial stages of the product's life cycle (Kotler, 1976). The marketing mix of product, price, place and promotion changes as the product moves from the intro- ductory stage through the growth stage. The initial message for new products is a description of what they are and what they do. The characteristics, unknown to potential users, must be promoted. As knowledge concerning the physical attri- butes of a product increases, a marketer's activities must be shifted to other facets of the product mix. As time passes, most potential 31 consumers become aware of a product's characteristics and act according— ly. The eventual result is a slowing of market share growth. The availability of distribution channels determines whether or not products reach potential consumers. Since most new panel products are developed by companies with existing distribution channels, the "place" marketing factor is not included explicitly in this analysis. Panel products can be classified as either industrial goods or consumer goods depending on their end-use. Unfortunately, most end— use information is derived from production rather than consumption sta- tistics (Food and Agriculture Organization of the United Nations, 1976). The major end-uses in terms of volume for panels are new housing, resi— dential upkeep and improvement, manufacturing, and new nonresidential construction (U.S. Department of Agriculture, Forest Service, 1979). Since panel demand generally is derived from demand for final consumer goods (e.g., houses, furniture, etc.), changes in end-use consumption will influence the diffusion of new products. That is, if end-use con- sumption is growing, market penetration is more likely. All other things being equal, increases in product prices are ex- pected to hinder the diffusion of new products. Skimming price poli— cies (i.e., higher prices) will lead to slower gains in market share than penetration price politices (i.e., lower prices) (McCarthy, 1975). Changes in price may be caused by shifts in supply, demand, or both. The prices of competing products are important factors, too. As competing products become more expensive relative to new products, mar- ket penetration is enhanced for new products, ceteris paribus. Income and population are expected to influence market share growth since they also affect long—term demand. Disposable personal income 32 is an important determinant of demand for wood-based products and influ— ences household formation and furniture consumption (U.S. Department of Agriculture, Forest Service, 1979). Finally, the availability of products in the market is influenced by transportation costs. As the distance from the producing region increases, inplace cost increases. As a result, product innovations produced in a given region have a competitive advantage over imported products, ceteris paribus. Raw Material Situations The raw material situation has always played a key role in the development of the forest products industry. Timber availability and depletion historically caused the movement of the logging industry into the Lake States on to the South and the Pacific Northwest and now, back to the South. The panel products industry is also influenced by wood supplies. In the west, DouglaS"fir is important for plywood production; loblolly, shortleaf and longleaf pine are the main plywood species in the South; and an assortment of species in the form of roundwood and residues is utilized nationwide in the other panels. The availability and price of these materials will determine, in part, the competitiveness of the various panels and the direction of future panel developments. Labor, capital, energy and adhesives are other input factors affecting product innovations in the wood-based panel industry. The plywood industry has evolved from labor-intensive operations using animal glue and wooden presses into a capital-intensive industry with multi-platen presses and a variety of glues (Cour, 1955). Likewise, 33 particleboard plants have grown in size and complexity requiring new combinations of inputs. Energy inputs are becoming more expensive, and particleboard mills are now competing for residues with plants which may burn the residues for energy production. In total, these factors will lead to changes in processing and input utilization. Models There are many approaches to modelling the demand for an existing product and its substitution for other products. Some approaches used are: (1) Methods based primarily on the autocorrelation properties of the time series to be forecast, (2) Methods conceived on the basis of a Markov process which hypothesizes that demand is a random variable, and (3) Structural econometric models which attempt to provide causal explanations for the phenomenon studied (Quandt, 1964). The first approach includes time series decomposition in which trend, cyclic and seasonal components are identified. The second approach involves a brand switching matrix which can be used to calculate market shares. The last approach is based on economic hypotheses and can be subjected to statistical testing. This study utilizes the first and third approaches for analyzing the diffusion or product innovations, factors influencing diffusion, and the effect of diffusion on wood re- quirements. The model hypothesized for this study is composed of three sub- models: (1) the aggregate panel market submodel, (2) the market share submodel, and (3) the wood utilization submodel. The aggregate panel 34 market submodel is an econometric model based on relatively gross econ- omic variables (Mills and Manthy, 1974). The market share submodel is an extensive examination of trends in new product diffusion. Both simplistic trend models and multivariate econometric models are used and compared. The wood utilization submodel is based solely on published data relating wood inputs to panel output. The two submodels requiring econometric analysis and the purpose for the models are presented in Figure 3. Mathematically, the three submodels can be related to total wood requirements as follows: WRi(t) = C(t) x Mi(t) x Ui(t) (1) where: Wri(t) = wood requirements for panel products i at time t, C(t) = aggregate panel consumption at time t, Mi(t) = market share of panel product i at time t, and Ui(t) = average wood requirement per unit of panel 1 output at time t. Theoretically, this model can be expanded by considering panel consump- tion and market share by end-use category. However, in practice it is difficult to expand the model because published data are generally based on production rather than consumption statistics. To the extent possible, end-use trends are incorporated in the discussion. Aggregate Panel Market Submodel The aggregate panel market is equivalent to the general product class, wood-based panels. Product classes have long life histories because they are highly population related (Kotler, 1976). Three steps are required in modelling aggregate panel consumption. First, the 35 Submodel Purposes Aggregate panel market (1) Estimate past consumption and test model parameters (2) Project future consumption Market share (1) Estimate past market shares and test model parameters (2) Estimate past market share trends (3) Estimate diffusion rates (4) Project future market shares Figure 3. Submodels requiring econometric analysis and the purpose for using the submodels. 36 extent of the market is determined, that is, which products will be included in the analysis? Second, a plausible method of aggregating dissimilar products is developed. Third, an aggregate model based on the first two steps is hypothesized. Since the emphasis of this study is on intra-industry competition between wood—based panels, substitutes such as aluminum, concrete, plastic, and so on are not explicitly considered. Empirical results from a study by McKillop, Stuart, and Geissler (1980) support not in— cluding non-wood structural product prices as explanatory variables for wood-based panel products. Two complimentary approaches are used to determine which products are included in the analysis. One approach is to examine panels in terms of end-uses; those panels having similar end—uses can be included in the aggregate model. The second approach is based on economic theory. A demand function can be estimated in linear or log-linear form, and the coefficients associated with other panel prices will indicate which panels are economic substitutes and which are complements. The demand for a good is a function of tastes and preferences, the price of the good, income, and the price of the other goods (Green- wald and Associates, 1973). Since there is no prior knowledge regard— ing the appropriate functional form for the demand curve, two simple models were estimated for this study: the linear demand curve and the constant elasticity demand curve (Hirshleifer, 1976). Estimation procedures are discussed in the final section of this chapter. The linear demand function has the following form: n P +b21+ZbiPi+e (2) Y=b +131 0 y i=3 where: 37 Y = quantity of the panel demanded P&= price of panel demanded I = income, Pi = prices of substitute and complement panels bO = constant bi = coefficients estimating the change in quantity demanded associated with a one unit change of the variable, ceteris paribus e = unexplained variation of Y. The coefficients associated with Pi have the same signs as the cross- elasticities. The coefficient is positive if the good is a substitute for Y and negative if the good is a complement (Hirshleifer, 1976). The log-linear demand function is similar to the linear form, except (1) the variables are the logarithms of variables used in the linear model and (2) the coefficients, bi’ become estimated elastici— ties. The elasticities associated with the logarithms of prices of other goods are cross-elasticities of demand. As such, if the estimated cross elasticity is positive, the goods are substitutes. If the cross- elasticity is negative, the goods are complements. An important factor that must be considered in determining which panels will comprise the total market is the panel with which comparisons and statistical analyses are made. If a panel with relatively few end-uses is selected, the total panel market will be smaller and market shares for individual product forms will be larger. By selecting a panel with many end-uses, the opposite results are achieved. As long as the panels under the study, particleboard and southern pine plywood, are included in the analysis, the panel used as the basis for comparisons 38 is not critical. This condition exists because the total market can be disaggregated to individual panels. Markets can be defined in terms of dollar sales or physical units. Physical units are used in this study to allow estimation of wood re- quirements. As mentioned previously, physical consumption data for all panels in various end-uses are not available over time. Therefore, the mathematical models used in this study utilize data aggregated over all end—uses. Supplemental end-use data are analyzed where pos- sible. In order to determine the total physical consumption of panel products, a method of aggregation must be developed. Ideally, the total market would be defined as the total surface area of substitute panels consumed in a period of time. However, available data do not permit this approach. Two alternative approaches can be used. One approach is to select a standard thickness for all panels, convert all panel consumption data to that thickness, and add the new panel data to determine the total market. This approach was used by the U.S. Forest Service in determining panel consumption on a 3/8-inch basis (U.S. Department of Agriculture, Forest Service, 1979). A different approach is used in this study. Panel production, consumption, and wood utilization factors are generally reported in standard measures unique to each panel. These standard measures are closely related to average thick- nesses used in various end-uses. Therefore, the surface measures with various standard thicknesses have been aggregated to define the total market. In using this approach, disaggregation by market share yields the surface area in terms of conventional standard measures. 39 Having developed a method for selecting and aggregating substitute panels, an aggregate consumption model was developed. This model was used to project long-term future panel consumption based on projections of exogenous variables in the model (Stone and Marcin, 1978). The general form of the model is as follows: n C(t) = a + :a,x, + v (3) o . 1 1 i=1 where: C(t) = aggregate panel consumption at time t xi = exogenous variables that shift demand and supply functions a = constant a. = coefficients estimating the change in total panel consumption associated with a one unit change in the respective exogen- ous variables, ceteris paribus v = unexplained variation in C(t). The most commonly used and available exogenous variables are gross national product (GNP), disposable personal income (DPI), population and time. These variables are extremely collinear (average r = .989 for the 1955 to 1979 period) so each one was tested individually as an explanatory variable. Based on statistical results and its role in housing and furniture purchase, disposable personal income was se- lected as the exogenous variable for projecting future panel consump- tion. Low, medium, and high values for disposable personal income were used (U.S. Department of Agriculture, Forest ServiCe, 1979). Though the appropriate functional form for the model is unknown, Hair (1967) infers that the semi-log form with logarithms of indepen— dent variables has substantial empirical support. Both the linear model above and the semi—log model were tested in this study. The 4O semi-log model was selected for future consumption projections due to statistical performance and empirical support. There are several problems with this type of model: (1) the esti- mated relationship may not be stable over the period of estimation and the period of projection, and (2) the explanatory variables have to be projected, too (Maddala, 1977). In addition, there are statis— tical problems discussed in the last section of this chapter. Market Share Submodels Four market share submodels were used to estimate the diffusion of particleboard and southern pine plywood: two were univariate non- linear models and two were multivariate linear models. The univariate models, the logistic function and the Gompertz curve, were used for analyzing the diffusion trend and rate over time. One multivariate model was used to analyze factors influencing the market share of par- ticleboard and southern pine plywood over time. The other multivari- ate model was used to analyze factors affecting the market share of southern pine plywood over space. Supply and derived demand structural equations for particleboard and southern pine plywood were estimated to screen variables used in the multivariate diffusion models and to isolate factors associated with the interaction of supply and demand. Since projections of these variables are not available, only the nonlinear univariate models were used to project future diffusion of the panel products. The diffusion of successful product innovations, including most grades of paper and board, evinces an S-shaped pattern (Warner, 1974; Hair, 1967). The logistic function and the Gompertz curve have been 41 used to model this diffusion pattern (Griliches, 1957; Buongiorno and Oliveira, 1977; Mansfield, 1961). The general form for these univari- ate functions is M = f(t) where M is the market share for a product innovation and f(t) is a simple nonlinear function of the time that has elapsed since the introduction of the innovation. The specific form for the logistic function is: where: M = market share P = potential or maximum market share constant a = location constant (determines horizontal movement of curve) b = proportionality constant (determines rate of increase of market share) t = time (1955 = 1, 1956 = 2, ..., 1979 = 25) e = 2.71828, the exponential constant The function is monotonically increasing and lies between two asymp- totes, M = 0 and M = P. The function is symmetric with a point of inflection at P/2 (Oliver, 1964). The Gompertz curve has the form: M = P a(bt) (5) where: M = market share P = potential or maximum market share constant a and b = constants determining the growth rate of market share t = time (for particleboard: 1955 = 1, 1956 = 2, ..., 1979 = 16) This curve is asymmetric and approaches P asymptotically. The point of inflection is at P/e (Lekvall and Wahlbin, 1973). 42 The logistic function is one of the most widely used diffusion functions in empirical research. Mansfield (1961) developed a strong theoretical basis for using the logistic function to study the diffu- sion of innovations. His work has been modified here to show its appli— cability to market share diffusion. Mansfield's basic premise is that the proportion of firms that have not adopted an innovation at time t that will introduce the inno- vation by time t+1 is a function of (1) the proportion of firms that already adopted it by time t, (2) various economic variables, and (3) other unspecified variables. By making a number of restrictive assump- tions, he argues that the number of firms that have adopted an innova- tion over time can be approximated by the logistic function. A number of Mansfield's assumptions can be modified to show their relevance to diffusion in terms of market shares. First, the function implies that new products will gain in market share up to some maximum level. This level can range from zero to one hundred percent of the market. Statistical estimation procedures, however, do not impose any restriction on the maximum level. Second, the proportion of market share already gained is an impor- tant determinant of future levels. This assumption is based in part on the role knowledge and learning play in economic growth (Arrow, 1962). As use of the innovation becomes more widespread, uncertainty concerning the innovation is reduced, thereby increasing adoption of successful innovations. Finally, the proportionality constant, which determines the rate of increase of market share, is a function of supply, demand, and other factors. The constant also has been defined as the rate of imitation 43 (Mansfield, 1961), the rate of acceptance (Griliches, 1957), and the proportionate rate of growth (Maddala, 1977). In essence, this assump— tion acknowledges that a multitude of factors influence acceptance of new products. In order to gain insight into the usefulness of the univariate models under different conditions of data availability, the model para- meters were estimated and compared for twenty different time periods. In addition to the model parameters, two additional statistics were estimated: (1) the inflection point and (2) the number of years required for the product to increase from 20 percent to 80 percent of potential market share. The latter statistic was calculated as an indicator of the diffusion rate, and is equivalent to the market share growth stage discussed previously. The multivariate model utilized to estimate the market share trend over time consists of one identity and two equations. By aggregating over all end-use markets at a point in time, the sum of the market shares of substitute products is equal to one, i.e., n t Z Mi(t) = 1 (6) i=1 Mi(t) is the percentage of the aggregate wood-based panel market (in decimal form) of the 1th product at time t and nt is the number of products in the market at time t. The market share functions each contain one endogenous variable which is influenced by a group of exo- genous variables. The functional form was: n M1= a0 +§:‘_aixi +111 (7) i=1 m M2 = bo + Egibjxj + v2 (8) 44 M1 and M2 = market share of particleboard and southern pine plywood, respectively x1 and xj = exogenous variables affecting particleboard and southern pine plywood, respectively constants a and b o a. and b structural coefficients estimating the change in market share for particleboard and southern pine plywood associated with a one unit change of the vari- able, ceteris paribus u1 and v2 = unexplained variation of M1 and M2 The two equation model was estimated as a system of seemingly unrelated regression equations and by ordinary least squares. Annual market share data from 1955 through 1979 for particleboard and from 1964 through 1979 for southern pine plywood were used in the model. Portions of the study period were relatively stable while other portions were turbulent in nature. The influence of transportation costs and regional availability of a product innovation, southern pine plywood, was examined in the following model: n M = C0 + clN + c D + c x + e (9) 2 1;3 i i where: M = the percent of all softwood plywood shipments attributed to southern pine plywood in a Rand-McNally Major Trading Area (MTA) in 1978 N = the number of southern pine plywood plants in the state con- taining the MTA and in adjacent states in 1978 D = the distance from Portland, Oregon in miles 45 x additional exogenous variables 1 co = constant c1, c2 and c1 = structural coefficients e = unexplained variation of M Data for twenty-nine Major Trading Areas were included in the analysis. The sample was comprised of the top 20 softwood plywood markets and the top 20 southern pine plywood markets. The model provides a method for examining the spatial diffusion of southern pine plywood in 1978. Estimation Procedures Several statistical problems must be considered when applying econometric analysis for estimation, parameter testing and projection. The appropriate method of estimation, completeness and identification, autocorrelation, multicollinarity, and specification errors are of particular concern. These factors must be considered for each submodel. MethOds of Estimation Estimation methods were determined for five functional relation- ships: (1) the supply-demand models for determining substitutes and screening variables, (2) the aggregate panel market submodel, (3) the univariate nonlinear market share submodels, (4) the multivariate mar- ket share submodels, and (5) the southern pine plywood spatial diffusion submodel. The supply—demand models contain endogenous price and quantity variables that are simultaneously determined in the market. If ordinary least squares (OLS) is applied to a single equation in a model and more than one endogenous variable is included in the equation, OLS 46 estimates will be biased and inconsistent due to correlation between endogenous variables and the disturbance term. The OLS estimating technique, however, provides an optimal solution in the case of recur- sive models (Johnston, 1972). For more general simultaneous equation models, several estimating techniques have been developed which yield consistent estimates for large samples. The techniques are indirect least squares (ILS), two- stage least-squares (ZSLS), limited-information maximum likelihood (LIML), three-stage least-squares (3SLS), and full information maximum likelihood (FIML). The first three techniques are classified as single— equation methods and the 3SLS and FIML techniques are systems methods. The systems methods were not considered for this study because they require large samples to be useful. The sample sizes for the simultaneous equation models were n = 13 and n = 11 for particle board and southern pine plywood, respectively. The sample size was restricted by the availability of comparable product price data. Both the ILS and the LIML methods require the use of all exogenous variables in the model. In addition, ILS requires exact identification (by the order condition) for unique solutions. Due to sample size and the number of exogenous variables considered in the supply-demand models, ILS and LIML were not used in this study. Two-stage least squares (ZSLS) was selected as the method of esti- mation for the simultaneous equation model because (1) it is a single- equation method, (2) it provides consistent estimates even with over- identification, and (3) it is characterized by low bias and small vari— ance. In any case, the selection of a given estimation technique for small sample simultaneous equation models is based on results of Monte 47 Carlo studies and personal preference. The instrumental variable method of ZSLS estimation was used in this study. When all exogenous variables in the complete model are used as instruments, the instrumental method is identical to ZSLS. Valid estimation can be based on fewer than all instruments when a complete model involves a large number of exogenous variables (Hall and Hall, 1978). That is, a demand function can be estimated even when the supply equation is not fully specified. A number of supply and demand variables were used as instru- ments in developing the final model specification (see Appendix for a list of variable names and their values). In order to prevent the first stage regression from being completely deterministic, a set of instruments not exceeding the sample size was used in each equation. The aggregate market and the southern pine plywood spatial dif- fusion submodels were estimated by means of ordinary least squares. This aggregate model included one endogenous variable and one endog- enous variable and several exogenous variables. Provided the OLS assumptions were not violated, the estimators will be unbiased and have minimum.variance for the class of linear estimators. The Gompertz curve and the logistic function parameters were estimated by nonlinear least squares. The statistical technique in- volves the minimization of the sum of squared residuals when the regression equation is nonlinear in its parameters. Gauss's method of estimation was used as the procedure for minimization (Hall and Hall, 1978). An alternative technique was used by Griliches (1957) to estimate model parameters. However, this technique was not used 48 because it was based on an independent and somewhat arbitrary estimate of potential market share. The multivariate market share submodels contain the appropriate endogenous market share variable and a series of exogenous supply and demand variables in each equation. Two techniques were used for estimation: ordinary least squares and seemingly unrelated regression. The OLS technique assumes implicitly the regression disturbance in one regression model is not correlated with the disturbance in another regression model. Since this condition is not explicit, most market share analyses do not account for the relationship between error terms (Stern, 1964; Weiss, 1968). The parameter estimates based on OLS are included in this study for comparison with seemingly unre— lated regression estimates. Seemingly unrelated regression, on the other hand, is used in models that have mutually correlated disturbance terms. This condi- tion is expected in equations (7) and (8) due to the market share identity (equation (6)). This subtle link allows estimation to pro- ceed using the maximum amount of information regarding the model. Completeness and Identification These concerns apply not only to the supply and demand models esti- mated by ZSLS. The particleboard and southern pine plywood models contain two equations (a supply and demand function) and two unknowns (panel price and quantity). Therefore, the models are statistically complete. Apparent consumption data for particleboard and production data for southern pine plywood were used as the quantity supplied and 49 quantity demanded variables in the functions. The wholesale price index for the particleboard industry (19678100) and the wholesale price index for southern softwood plywood (1969=100) were used as both supply and demand price. Price data was deflated by the whole- sale price index for all commodities. The equations in this study are all identified by the rank con- dition and overidentified by the order condition (Maddala, 1977). Therefore, the equations are estimable. The instrumental variable method of estimation requires that at least one instrument (predeter- mined variable) from another equation be used for each endogenous variable other than the "dependent" variable. Overidentification simply makes additional instruments available. Autocorrelation Autocorrelation among error terms causes several statistical problems. First, the estimated coefficients are still unbiased, but no longer have the minimum variance property. Second, the standard error estimates of the coefficients may underestimate the true standard error. Finally, tests using the t distribution are no longer strictly applicable (Neter and Wasserman, 1974). The Durbin—Watson statistic was used to test for the presence of autocorrelation. For sample sizes smaller than 15, the test ranges were extrapolated (Durbin and Watson, 1951). The test was inconclusive or showed no autocorrelation for all multivariate linear models. The Durbin-Watson test indicated autocorrelation in the univar— iate aggregate panel submodel and in the univariate nonlinear models. This was expected as the omission of one or more key variables in 50 business and economic regression equations often leads to positively autocorrelated error terms. These equations, when used for projections, will produce inefficient estimates (Johnson, 1971). However, this is not a problem in this study since projection confidence intervals are not used. Multicollinearity Though predetermined variables may be highly correlated, this does not inhibit obtaining a good fit to the data. Multicollinearity has several effects on statistical estimation. There is a loss in the precision of estimation, that is, it is difficult to disentangle the effects of the collinear variables. Also, coefficient estimates are very sensitive to particular sets of sample data. Multicollinearity always exists in economic variables. The degree of effects is the major concern. Two methods were used to reduce the apparent effects of multicollinearity in the models. First, if two variables were extremely collinear (simple correlation between the two of .99 or greater), one was dropped from the study and the other was retained to represent the pooled effect of the collinear variables. In the case of somewhat equivalent variables such as housing starts and residential construction expenditures, only one variable was re— tained in the final model though both were retained for test runs. Second, extensive test runs were used to examine the addition and dele- tion of variables for different sample sizes. Even with this two step process, the selection of the final equa- tion variables was based on judgment. This was especially true in the case of the supply and demand equations. 51 Significance Tests Variables were tested for significance using a one-tailed t-test. The test indicates the probability level at which the estimated coeffi- cient sign is its true sign. For the aggregate panel market submodel, the Cochrane-Orcutt procedure (Maddala, 1977) was used for re-estimating and testing parameters after the presence of autocorrelation was indi- cated by the Durbin—Watson statistics. This procedure was used to statistically correct for first—order autocorrelation. After correction, all parameters were significant at the 1% alpha level. Specification Errors Specification errors can result from: (1) omission of relevant explanatory variables, (2) disregard for qualitative changes in vari- ables, (3) inclusion of irrelevant variables, (4) inappropriate mathe- matical form used in estimation, and (5) incorrect assumptions concern- ing the disturbance terms in the regression (Kmenta, 1971). Initially, a large number of variables were selected based on economic theory and a literature survey. Many variables were later dropped due to high collinearity and the results of test runs. The unavailability of data on important variables, such as, wood residue price for particleboard, would have some effects on final model parameters. When possible proxy variables were used to account for the hypothesized relationship between the dependent and excluded vari- ables. In an aggregate analysis, qualitative changes are ignored. In the case of wood-based panels, the lack of available data on panel quality changes inhibits incorporating changes into the analysis as 52 an independent factor. A number of functional forms are used in the analysis. Justifi— cation of particular mathematical forms was presented on a case by case basis in previous sections of this chapter. Standard assumptions regarding the disturbance term were made with the exception of the seemingly unrelated regression. In that case, correlation between the disturbance terms in the individual equations was assumed. PAST TRENDS IN INNOVATION AND PRODUCT DIFFUSION Qualitative and quantitative factors interact as new products diffuse through the market. By examining past trends in these factors, some insight into their relationship to successful diffusion is gained. Many of the factors are intertwined; however, they are presented sep- arately to emphasize their importance. Particleboard and southern pine plywood are successful product innovations. They have gained widespread acceptance in end-use markets. Their product characteristics play a large role in this acceptance. In addition, product standards, building codes, process innovations, raw material price and availability, and various economic factors have been instrumental in this diffusion process. Product Characteristics Panel products possess three general characteristics: physical properties, strength properties, and working properties. Physical properties include density, moisture content, water absorption, and fire resistance. Examples of strength properties are bending strength perpendicular to the plane, deflection under load, impact resistance, and screw holding. Accuracy to size, machining properties and surface quality are the primary working properties (Akers, 1966). Property comparisons of different panels are made with consideration 53 54 of end-use requirements. Due to its properties, particleboard is used extensively in furniture and construction, whereas, southern pine ply— wood is used only in construction. Particleboard Characteristics Particleboard is: A generic term used to describe panel products made from discrete particles of wood or other ligno-cellulose material. Other materials can be added to the production process to improve the board. Thermosetting resins are added to the particles to serve as a binder. The particles are bound into a solid board when the particles and resins are placed under heat and pressure (Dickerhoof and McKeever, 1979). Particleboard is produced in a variety of densities and thicknesses. For brevity, only the general panel characteristics are presented here. Particleboard is a popular core panel in furniture because it has excellent gluing qualities on all planes and superior working prop- erties. The surface quality is especially important due to the increas- ing use of thin veneers as laminates. The panel surface is uniform and the panels stay flat making it a good product for furniture manufac- turers (U.S. Department of Agriculture, Forest Products Laboratory, 1974). Particleboard characteristics are well suited for construction, too. After manufacture, the board is constant and uniform in its prop— erties. Its use as floor underlayment is attributed, in part, to its impact resistance and flatness. However, particleboard is more suscep- tible to damage by excessively damp and wet conditions than plywood panels. The ability of manufacturers to create large panels enhances particleboard's acceptance as mobile home and prefabricated house deck- ing. 55 In general, the flexibility of manufacturing processes have per— mitted particleboard to be tailored to specific end-use characteristics. A key factor in particleboard's future success will be the continued modification of its characteristics. Southern Pine Plywood Characteristics Plywood is defined as: ...a flat panel built up of sheets of veneer called plys, united under pressure by a bonding agent to create a panel with an adhesive bond between plys as strong as or stronger than the wood. Plywood is constructed of an odd number of layers with grain of adjacent layers perpendicular. Layers may consist of a single ply or two or more plys laminated with grain direction oriented parallel to the long dimension of the panel. The odd number of layers with alternating grain direction equalizes strains, prevents splitting, and minimizes dimensional change and working of the panel (U.S. Department of Commerce, National Bureau of Stands, 1974). The primary species utilized in southern pine plywood are loblolly, longleaf, and shortleaf pine. Slash and spruce pine are used to a lesser extent. Plywood characteristics vary considerably. The veneer grades used as plys determine the quality of the final product. CD sheathing is the mainstay of the southern pine plywood industry. However, plyform, underlayment, 303 Siding, and AC/Ad panel production is increasing (Baldwin, 1977). Structural plywood panels such as CD sheathing are engineered for end—uses where tension, compression, shear, cross-panel flexural properties and nail bearing are important (U.S. Department of Commerce, National Bureau of Standards, 1974). Grade C face veneer restrictions are well suited to southern pine species since small knots, discolora- tion, and synthetic repairs are permitted. Grade D back veneers have 56 fewer restrictions. Process and Product Trends Processes adopted by the U.S. panel industry often are imported from other countries. As the number of processes increases, products become more varied and specialized. The result is greater acceptance of wood-based panels in end-use applications. Particleboard Process and Product Trends Particleboard was developed in this country to utilize waste wood and to compete with established wood products (Reid, 1958). The indus- try has grown as new processes and products have emerged. The major processes used in the manufacture of particleboard are the extrusion process and the platen press process. Product characteristics vary according to the process used. Early attempts to produce particleboard in the United States during the 1920's failed primarily due to the lack of suitable adhesives. During the 1930's and early 1940's, several breakthroughs in resin production and application enabled particleboard production to proceed. The resin shortages during World War II gave way to reindustrializa- tion following the war; the particleboard industry has grown drama- tically since that time (Moslemi, 1974). A number of types of particle- board and production techniques have been developed over the post- war years. The alternative production methods often were developed to avoid infringement on Swiss and German patents (Akers, 1966). Akers (1966) identified four general types of particleboard: (1) single layer, platen pressed, (2) three layer platen pressed, 57 (3) graded density, platen pressed, and (4) extruded board. These boards can be produced in a variety of thicknesses and in a range of densities. The final product is determined primarily by market require- ments. The single—layer, platen pressed board was the first type developed. The boards were homogeneous and utilized planer shavings, industrial wood residues, and flax shives as raw materials. The particles are in a random criss-cross pattern parallel to the surface of the board. Early single layer boards were produced in batches. The first domestic continuous press process was designed by Ralph Chapman and installed in the Chapwood plant at Philomath, Oregon in 1957 (Anonymous, 1957a). The three-layer board is similar to the single—layer type, but has a sandwich construction with courser materials placed between flakes or other material for smoother surfaces. This process allows for greater bending strength and was first used in furniture as coreboard. Fred Fahrni, a Swiss engineer, developed the first 3—layer type of board which became known as Novoply in the United States. The United States Plywood Corporation began producing Novoply in 1951 at their Anderson, California, plant. Edwin Behr developed a similar board in West Germany for use in furniture. The Behr process was first intro- duced to the United States in 1957 by Roddiscraft, Inc. at Arcata, California (Anonymous, 1957c). Though it was the first of its type in the U.S., twenty-one similar plants had been built previously world- wide. The graded density board, originally manufactured in 1958 by Baehre Metallwerke in West Germany, was designed with a high density surface and a low density core. Finer particles are near the surface, and 58 unlike three-layer boards there are no abrupt changes between particle sizes. The boards are used in furniture and building construction. The Bahre—Bison process is used widely in the United States. Edwin Behr also developed a graded density board. Other platen press particleboard manufacturing systems adopted in the formative years of the industry were designed by major companies, such as Industrial Development, Miller Hofft, and Columbia Engineering and by various smaller companies (Mottet, 1962). The processes were usually tailored to the raw material available and to specific end- use needs. Extruded boards are made by forcing the wood and resin mix through two parallel metal platens. The particles are perpendicular to the surface of the board. As a result, the board has low bending strength and is used as a corestock for furniture. The three extrusion processes most used in this country are: (1) the Kreibaum process, (2) the Lanewood process, and (3) the Chipcraft process (Mottet, 1962). The Kreibaum vertical extrusion process was developed by Otto Kreibaum in West Germany and first used in the United States by the Jasper American Manufacturing Company at Henderson, Ken— tucky in 1954. The Lanewood process for horizontal extrusion was in- stalled initially by the Lane Company at Alta Vista, Virginia in 1953 (Anonymous, 1957b). One of the earliest Chipcraft horizontal extrusion process was installed by Berkline Corporation, Mbrristown, Tennessee, in 1953 (Anonymous, 1957b). Three recent developments in product or process design are wafer- board, oriented strand particleboard, and Mende process particleboard. These newer panels are discussed in final chapter along with newer 59 process innovations. The production of new panels is another indica- tion of continuing innovation in this industry. The raw materials for particleboard are industrial wood residues, forest residues, roundwood, resins and additives. The wood materials are prepared by hammermill or flaker, dryed and screened, mixed with adhesives and additives, formed into boards, and pressed. In addition to major process innovation and diffusion, evidenced by new production systems, improvements in many aspects of the production process have occurred. Recent capital expenditures presented in Table 5 indicate that refinement in processes rather than construction of new plants is occurring in the particleboard industry. The traditional method for preparing particles is to use an assort- ment of attrition mills, hammermills, chippers, ring flakers and drum flakers. Refinements in this equipment and in screening continues to improve the uniformity of materials used in manufacturing (Food and Agriculture Organization of the United Nations, 1976). The shaping lathe headrig and "fingerling" production are among the newest develop— ments in flake preparation (McSwain, 1977). Particle dryers have increased in size and efficiency. In some cases, wood residues are being utilized as an energy source for drying. Particle moisture, once measured by hand sampling, is now monitored and controlled by infrared spectrometers and devices measuring electri- cal resistance. After drying, the particles are stored or blended with resins and additives. Short—retention-time blenders have replaced many long-retention-time blenders over the past 15 years (McSwain, 1977). Urea-formaldehyde and phenol-formaldehyde are the primary resins 60 Table 5. New capital expenditures for plant and equipment in the particleboard (SIC 2492) industry, 1972 and 1977 (in millions of dollars). Year Capital Expenditures New buildings and New machinery other structures and equipment 1972 $2.7 $32.1 1977 3.2 29.3 Sources: U.S. Department of Commerce, Census of Manufactures, 1972, 1975b; U.S. Department of Commerce, Census of Manufactures, 1977, 1980b. 61 blended with the wood particles. Interior particleboard is made with urea-formaldehyde resins and exterior particleboard is bonded primarily with phenol-formaldehyde. Both are water soluble and thermosetting. Urea-formaldehyde is used more extensively in the industry according to available data (see Table 6). Improvements in adhesive cure time and tack control have occurred over the years. Experimentation is increasing with alternative binders such as isocyanates, cement, mala— mine based resins, and tannin formaldehyde adhesives as industry concern for resin cost and availability continues (Coppens, Santana, and Pastore, 1980; McSwain, 1977). Additives are also blended in with the adhesive and wood particles. Common additives are rosin and/or paraffin emulsion as water repellents, phenol formaldehyde for added strength, penta-chlorophenol and arsenical compounds for fungus and insect protection, and borate for fire retar- dancy (Akers, 1966). By changing the additive mix, new board character- istics can be attained. The next stage in the production process, board forming and trans- port, has undergone several changes as the industry has evolved. The improvements are concentrated in platen press systems which account for the majority of particleboard produced in the United States. Some of the most notable advances in this area are the increased size and sophistication of matforming equipment, the development of fiber allign- ment systems, and the introduction of caulless transport systems. These innovations allow for a broader range of wood inputs and create product variants for end-use markets (Food and Agriculture Organization of the United Nations, 1976; McSwain, 1977). Platen-type and continuous prepresses have been developed for ‘ 62 Table 6. U.S. resin consumption by type used in the manufacture of other platen type particleboard, selected years (in millions of pounds). Year Quantity of resin consumedl/ Urea Phenol formaldehyde formaldehyde and other types 1966 262.0 6.2 1969 631.7 25.0 1972 1215.2 22.6 1975 1025.6 11.6 -l/Dry basis prior to 1969; wet basis beginning in 1969. Source: U.S. Department of Commerce, Bureau of the Census, selected years(a). 63 caulless transport systems. The prepresses reduce the maximum press opening needed which leads to faster press closing. The result is a reduction in surface precuring, increased core density, and better internal bonds. Particleboard presses are either continuous types or discontinuous single-opening or multi—opening types. Modifica- tions in single-opening presses have improved their competitive posi- tion in recent years, but multi-opening presses are still dominant. Shorter press cycles, larger board sizes, and better thickness control are the main advantages of single-opening presses. The development of continuous presses for thin (Mende) particleboard production is another example of recent improvements in board pressing (MCSwain, 1977). Board finishing and secondary processing is the last step in the production process. Notable advances in this processing stage include the extensive adoption of wide-belt sanders and improvements in lami~ nating and curing. These advances lead to broader acceptance of par— ticleboard due to increased quality. Southern Pine Plywood Process and Product Trends The southern pine plywood success story is based on adopting tech— niques for peeling small diameter logs, developing suitable resins, and utilizing substantial southern timber resources. The small sized low quality peeler logs are used primarily for construction grade ply- wood. A substantial number of process innovations have been incorpor- ated in the rapidly growing southern pine plywood industry. Many of these have been installed during plant construction. Recent trends 64 in nationwide capital expenditures are presented in Table 7. Approxi— mately $26 million of the 1972 total expenditures and $45 million of the 1977 total occurred in the South (U.S. Department of Commerce, Census of Manufactures, 1972, 1975a; U.S. Department of Commerce, Cen- sus of Manufactures, 1977, 1980a). Before the first southern pine plywood plant was opened in Decem— ber, 1963, considerable industrial research had taken place. Promis- ing yield trials of southern pine logs in western mills provided an impetus for interest in the South (Fassnacht, 1965). Though consider- able government and industry experimentation in an attempt to find suitable adhesives was ongoing, Georgia Pacific Corporation's develop— ment of the first adhesive capable of satisfactorily laminating south— ern pine was a major breakthrough (Koch, 1972 and Anonymous, 1980d). With the adoption of high-speed automatic lathes, the southern plywood boom was underway; thirty-four plants began operation in the first five years of the industry's existence. These plant startups indicate which companies were innovators and early imitators (see Table 8). The production process begins in the log yards used to store and sort logs in the South. Huge jib cranes often are used to stack and move the logs (Baldwin, 1977). Once inside a plant, the logs are peeled with automated, high-speed lathes. Only rotary-peeled veneer is produced (Koch, 1972). Back-up rolls, high-speed log charging, and geometric chuck centering increase the efficiency of the process (Food and Agriculture Organization of the United Nations, 1976; McSwain, 1977). Southern pine veneer is mostly sapwood. It is cut from logs averaging 14 inches in diameter with 4 to 6 inches left as core. Jet 65 Table 7. New capital expenditures for plant and equipment in the softwood veneer and plywood (SIC 2436) industry, 1972 and 1977 (in millions of dollars). Year Capital Expenditures New buildings and New machinery other structures and equipment 1972 $ 9.0 $56.5 1977 10.6 95.1 Sources: U.S. Department of Commerce, Census of Manufactures, 1972, 19753; U.S. Department of Commerce, Census of Manufactures, 1977, 1980a. 66 coo4 om m2 .e444>e4:o4 ofiweoemlefiwuoeo ocoH om m2 .ucoaseem .auoo voo3h4m eaHm eu4ea oom4 on n2 .hufio exoaooom .ou woo3>4m hem oxeeaeeeso oom4 004 <4 .aeHuOHM .oo voosk4m ue>sooce> coma won <4 .cmeHOm coaaaeeoneoosa4m .m.: oo¢4 co <4 .v4e4wca43 .ocH .umepx:440 oom4 ow <4 .comwoa .oo voo3z4m ecefimfino4 @004 cc <4 .sepcwz .auoo cumnunomlefinSDHOU cooH om <0 .neae>em ofiwfioemlefiwuoeu mom4 oo4 <> .eeuoaam oem4oemue4wuoeu moa4 Nu x9 .cfist4 mHocHHHHImcm3O mcm4 ow oz .nusoaz4m .ou nemaeesuexez mess on we .eane4mee44nm .oo tomameetmsmz mom4 004 <4 .m4emxeo .ou woo3h4m um>sooce> mo¢4 mm <4 .COOmsm cuenusomaaewucem moo4 om4 m< .Ne uuemeouo 04m4oemne4wuoeo mc¢4 om4 m< .4e uuemmOHU onHoemlekuoew mo¢4 co 4< .GOuadm .oo penas4 nouoom ecoss4m .m.: «oaH om x9 .440oen \.oo .un4 ween neeausom «oo4 co x9 .ee4m44m .ou nonas4 hnufim moo4 0N4 m< .eomeuom owmwoemlewwuoeu new» mauueum lkuwoemeo GOHueoo4 woodwoum \a ou mom4 .mueum wee hufio >4 mueo% anuueum uce4a use .huwoeaeo .mumoapoua .wom4 mooBm4a mafia cuesusom .w eHneH 67 .wmo4 .esoakcoc< "meadow .mHmen socelw\m .ueem eueavm 5044445 :4 weusmeea mueoeaeo.l \4 wom4 wom4 wcm4 w©o4 wom4 noa4 mom4 n©a4 nom4 n©m4 n©m4 mom4 ©©m4 oo oo oo om oo4 co mo om oo oo 00 0a Om Um .e444>4emmnm <4 .wcHHmea aHmHm COm¥UMW U4440emle4wuoeo muosvoum ummuom hcosuc< .auoo aEeo .m.: .auou aEeo GO4GD 4evo04m ce4442oez .auou mooB>4m e4wcewuH ofiwfioemle4uuoeo .oo ue4634 ucoeeua .ocH .%4mlnunom e003>4m auezuuoz ueeuo ofimfioemlewwuoeu .muoo 4e048e£u e4xu< .oo Heeseenummez A.e.ucoov w «Heme 68 dryers are the predominant method for drying the plys (Koch, 1972). Experimentation with various dryer configurations (baffle arrangements, temperature settings, drying times, and so on) continues in the South. However, drying is still a major bottleneck in the production process (Food and Agriculture Organization of the United Nations, 1976). Curtain coaters, spray lines, and soft roll spreaders are used for glue waste while cutting down on manpower requirements. Most southern pine plywood is glued with phenol-formaldehyde resins (Food and Agriculture Organization of the United Nations, 1976; McSwain, 1977). The mechanization of the production process has increased in recent years. One example of mechanization was the advent of auto- matic layup systems in 1968 (Mast and Pease, 1969). These systems provide an efficient means for continuous production (DeLess, 1975). Large plant size is one result of extensive automation. Faster cure time, automatic charging, and automatic discharging equipment are examples of increased mechanization in pressing. Other improvements include better blister detection, new patching and plugging materials, and more extensive use of wide-belt sanders and improved abrasives (Food and Agriculture Organization of the United Nations, 1976). Product Standards and Building Codes Product characteristics and product standards are tied to building codes and panel end-uses. Panel standards generally fall into two categories: general descriptive type standards and specific end—use standards (Frashour, 1961). The former are developed to encompass most producing companies, whereas the latter are more exacting and 69 restrictive. Government Commercial Standards and Voluntary Product Standards are examples of general descriptive type standards. Typical end—use standards are FHA (Federal Housing Administration) Interim Standards, Industry Association Standards, and ASTM (American Society for Testing and Materials) Standards. The Voluntary Product Standard (previously called the Commercial Standard) is voluntary, however, grademarking based on the standard is believed to promote product acceptance in the market (Anonymous, 1971b). The purpose of the Commercial Standard is "...to establish quality criteria, standard methods of testing, rating, certification, and labeling of manufactured commodities (U.S. Department of Commerce, National Bureau of Standards, 1966a)." Particleboard Standards Particleboard standards evolved from the 1956 FHA Interim Standard for Floor Underlayment, Kitchen Countertops to be Overlaid with Plastics, and Wardrobe Doors to the Commercial Standard 08236-66, Mat—Formed Wood Particle Board, issued in 1966. A number of standards for specific end-uses were developed between 1956 and 1966. Particleboard standards for specific end-uses have been promoted by the National Particleboard Association (NPA). Examples of these standards are: NPA-l, Standard for Particleboard for Mobile Home Decking; NPA-2, Standard for Particleboard Decking for Factory Built Housing; and NPA-3, Specification for Particleboard Floor Under- layment Coated with Wax-Polymer Type Hot-Melt Coatings (U.S. Department of Commerce, National Bureau of Standards, 1977). The 1966 Standard contains information on interior and exterior 70 panels including panel form, materials, dimensions, dimensional tolerances, and properties. Inspection, test methods, marking, and certification are described in the commercial standard, too. The property requirements for mat-formed particleboard are based on type of use (interior or exterior) and panel density (U.S. Department of Commerce, National Bureau of Standards, 1966a). C8236-66 was recently updated as ANSI (American National Standards Institute) A208.l (Anonymous, 1980c). The new standard consolidated existing particle- board standards such as ANSI Al6l.l for kitchen cabinets and ANSI Al6l.2 for countertops. Plywood Standards Standards applicable to southern softwood plywood have existed throughout the short life of the product. The first Commercial Standard was C5259-63, Southern Softwood Plywood, which was in existence when the first panels were produced. This undoubtedly paved the way for more rapid acceptance of this new panel. 08259-63 was combined with Commercial Standards for Douglas Fir Plywood and Western Softwood Plywood in 1966. Product Standard PS 1-66, Softwood Plywood, Construc- tion and Industrial, was the result (U.S. Department of Commerce, Nation- al Bureau of Standards, 1966b). PS 1-66 was updated and re-issued in 1974 as Voluntary Product Standard PS l—74, Construction and Industrial Plywood. PS 1-74 covers "...the wood species, veneer grading, glue bonds, panel construction and workmanship, dimensions and tolerances, marking, moisture content, and packing of plywood intended for construction and industrial uses (U.S. Department of Commerce, National Bureau of 71 Standards, 1974). Test methods are also included. Plywoods are classified by their exposure capability (interior or exterior) and grade. In general, higher grades have more restrictions on the veneer quality. The primary panel grade for southern pine plywood is interior CD sheathing with exterior glue. Interior, intermediate and exterior glues can be used with interior type plywood. Since they provide consistency, standards are desired by pro- ducers, distributors, specifiers, architects, home builders and con- tractors. The standards often are adopted in model, state, and city building codes. BuildinggCodes Building codes are the legal minimum criteria which must be met in new construction, alteration and modification of buildings. According to a 1976 study (Vogel, 1976), over 7000 building codes existed in the United States. The study covered three Model Building Codes, twenty State Building Codes, and thirty City Building Codes. CSZ36-66 and PS 1-74 were included in the study. The particleboard standard was referenced in two Model Building Codes (the 1975 Basic Building Code and the 1973 Standard Building CodeL,nine State Build- ing Codes, and six City Building Codes. The plywood standard was referenced in flrasame Model Building Codes, eleven State Building Codes, and seventeen City Building Codes. As expected, the more traditional product, plywood, had gained wider acceptance. Building codes can prevent or allow the utilization of new pro- ducts in construction. Therefore, acceptance of product standards by building regulatory agencies is necessary for the widespread 72 diffusion of panel innovations. Ventre (1979) studied the adoption of innovations in residential construction. He found that the regula— tory agencies acted as technological gatekeepers and in the aggregate adopted innovations slowly at first, then more rapidly until adoption finally leveled off. As more regulatory agencies accept innovations, adoption by the construction industry is facilitated and product dif- fusion is increased. Product Consumption, Price and End-Use Trends Total consumption and price trends reflect the interaction of supply and demand over time. An important factor in determining total panel product demand is economic activity in end-use markets. One indication of a panel's competitiveness in a given market is its abil- ity to penetrate the market successfully. Particleboard Consumption, Price and End-Use Trends Particleboard consumption in the United States increased from 70 million square feet (3/4—inch basis) in 1955 to 3586 million square feet in 1979 (see Table 9). Available price data indicates that the wholesale price of particleboard relative to other goods has declined since 1967. Rising costs for transportation, energy, and adhesives along with increased competition for particleboard furnish may reverse this trend in the future. The declining price is one impetus for the increased consumption of particleboard; particleboard's diverse uses is another (see Table 10). Particleboard is used primarily as core stock, floor underlayment, and mobile home decking. For 1979, 48.4 percent of the particleboard 73 q.c¢ quN cm 54 mmmN mnm4 o.mm mmmN N44 N «o4m qnm4 m.w© qmqm mm N4 «men mmm4 w.No owom me «4 544m Nnm4 m.No Nme cN w «mmN 4nm4 «.mo mmm4 04 m «054 ohm4 m.mo4 «454 «4 N4 @454 mc¢4 m.w¢ oNq4 o 4 qu4 wom4 c.004 «444 N 4 m444 mom4 I mom I 4 mom oom4 I now I q mow mom4 I oqo I N wmo wom4 I wmq I 4 omq mom4 I woe I I woe Noo4 I on I I on 4oo4 I wON I I woN com4 I omN I I cmN ama4 I OmN I I omN wmm4 I mw4 I I mw4 nmm4 I 444 I I 444 omm4 I ON I I On mmm4 Acosueeo4v .Ixoec4 \MUHHQ m4eme4os3 .Icowugasmcoo cowuonwoua meueHMem \N uceueaa< wupoaxm muuoaaH o4umeaon new» ..lAwfimen noc4Iq\m .uoem whence mo mc044448 SHV mmm4 Immo4 .erC4 eoflpa wee .504uaanmcoo oceueamm .muuoaxe .eupOQEH .a04uosvoua vweone4owuuea .m.D .m o44ea 74 .mo4om4ueum none4 mo seeusm .pone4 mo Samsoueeeo .m.: mom? .mom4 .mm4enm use unw443 ”enema peuoe4ee .mUHBOGOUm 4e4uumswcH mo deeusm .euuee I500 40 uaeEuueamQ .m.D ”va whee» veuoe4mm .mSmceo ecu mo seeuam .moueEEou mo useEuuemen .m.: ”AUVmuemz wmuue4mm .msmceo ezu mo deeusm .moumasou mo uaeaupeamn .m.D mnn¢4 .ma4e4m mow¢4 .uum48e4 “meousom .msmceo emu mo deepen >4 3e4>eu peecD.l \m .u4m4e3 c4 weuuoaeu euev anomafi Eoum meueefiumm.l \q .kea4 o“ “04ta e4ae44e>e Soc Aoo4ueea4v mmaoaeoeeOu 44m toe xmece e04pa e4eme4043 man an weue4mee Aoo4umom4v xemcfl eowua e4eee4043 usauso %pum5@¢4 vueone4owuuem.l \m .muuomxe I muuoaafi + CO4uosmoua oflumeEoe u :04uaasmcoo uaeuema<\m .mueon emeooua mace: use wueonue444 muHecev Bowmee ev24054 uoc meoa\w o.Nm nwmm oo4 \m¢w4 40mm mnm4 m.mo «wen m¢4 \m¢m4 040m wma4 N.qm Neon mm \wcm4 Nmmm nmm4 w.qm mm4m om \m4m Non omo4 4.e.uc00v a e4aee 75 Table 10. Percentage of total particleboard industry production by end-use, 1973 (in percent). Intended end-use Percent of total industry Production Floor underlayment 27 Furniture core 27 Mobile home decking 15 General purpose 10 core stock General purpose 8 stock, furniture core, cabinets Cabinets 4 Door core 4 Siding 1 1/ Factory-built -— housing, decking Other 4 1/ —-Less than 0.5 percent. Source: Dickerhoof, 1976. 76 market was for industrial grades, 34.6 percent was for floor underlay- ment, 8.6 percent was for mobile home decking, 3.6 percent was for door core, and miscellaneous uses comprised the remaining 4.8 percent (Lambert, 1980). Activity in these end-use markets determines the particleboard product mix to a large extent. Changing styles in the furniture industry have influenced particle- board consumption. As the trend toward cleanline furniture continues, particleboard's properties become more desirable (Wilson, 1975). The furniture production index, which measures physical output, increased 185 percent from 1960 to 1976. Particleboard consumption in furniture increased 1533 percent during the same period (see Table 11). Though not as dramatic, increases in average per unit consumption of particle— board has also occurred in the housing and mobile home markets (Dicker— hoof, 1978; Wilson, 1975). Southern Pine Plywood Consumption, Price and End-Use Trends Southern pine plywood production increased from 80 million square feet (3/8-inch basis) to 7650 million square feet over the 1964 to 1979 period (see Table 12). Most production is consumed domestically. For example, less than seven-tenths of one percent of 1978 shipments were exported (Anderson, 1979). The deflated price of southern pine plywood has fluctuated in response to residential construction activity in recent years. During slack periods in construction, panel prices decline. As activity in- creases, prices rebound. Though stumpage prices are rising in the South, proximity to the major markets in the Eastern United States will continue to enhance southern pine plywood's in-place cost advantage 77 .Ow©~ .w0fi>hmm ummhom .m.D umUHDOm .wueon euwmoaaou use .mueoeexe4m .wueonuemeB .cueonuenflw zuemcemlaswmma mevs4ocH.l \4 N.Nm4 o4m4 oo4 ooq 0mm omm4 4.wc4 ocm ow omN com Omm4 4.mm one we @44 N4m mom4 w.4n 004 N4 em om oom4 Aoo4nmcm4v eusu4cusm xewcfi muoswoua 4ea04usu4umc4 eunufisusm coauoswoua penuo vce 4e404eaaoo w4osemsom ensuficusm 4eu0H msouu nonwoum Meow .4m4mmn euaaIs\m .umme museum no maoe44aa see eem4 eae .okma .moa4 .ocm4 .asouw acetone >4 meueum ueuHGD enu :4 NG4usuoemaceB c4 :OHumasecoo \flcneone4owuuem .44 e44eH 78 Table 12. U.S. southern pine plywood production and price index, 1964-1979 (in millions of square feet, 3/8-inch basis). Year Production Deflated wholesalel/ price index— (1969=100) 1964 80 _ 1965 402 - 1966 1140 - 1967 1779 - 1968 2373 - 1969 2875 100.0 1970 3315 81.1 1971 4410 89.8 1972 5319 102.8 1973 5560 105.1 1974 5130 84.0 1975 5676 80.4 1976 6814 101.0 1977 7447 116.0 1978 7898 115.2 1979 7650 94.0 -l/Southern pine plywood wholesale price index (1969=100) deflated by the wholesale price index for all commodities (1969=100); not available prior to December, 1968. Source: Anderson, 1979; U.S. Department of Labor, Bureau of Labor Statistics, selected years. 79 relative to Douglas fir plywood (Holley, 1969). Unlike particleboard, southern pine plywood production is almost exclusively for CD sheathing. The product mix for southern region shipments in 1978 was: 91 percent sheathing, 5 percent sanded, and 4 percent specialties (Anderson, 1979). The majority of shipments were to the southern region. According to one source (Baldwin, 1977), the product mix will shift to higher value added products, such as plyform, underlayment, and siding, after mill operations stabilize with efficient production of sheathing. However, until that time, consumption will be predominantly in the form of sheathing in the hous- ing and mobile home markets. Raw Material Situation The price and availability of inputs into the productive process influences the competitiveness of any product. The inputs can be in the form of wood, adhesives, labor, energy, or capital. As price in- creases or as availability becomes restricted, there will be a response in the price and production level of the panel product. Particleboard Raw Material Situation As mentioned previously, one factor that spurred the growth of the particleboard industry was the availability of residues from wood processing plants. Examples of wood inputs used in particleboard are: planer shavings, plywood residues, furniture and veneer manufacturing residues, and millwork residues. In 1956, approximately 50 percent of particleboard wood inputs came from sawmills. Veneer plants and furniture plants accounted for 80 30 and 14 percent of the inputs, respectively, with other wood proces- sors supplying the remaining 6 percent (Reid, 1958). The situation has changed since 1956; the sources of material now include pulp-type chips and roundwood (see Table 13). Planer shavings were the primary residue used in the West and South in 1973, and roundwood was the most important source in the North (Dickerhoof and McKeever, 1979). More recent data indicates planer shavings are declining in importance rela- tive to other residues (Food and Agriculture Organization of the United Nations, 1976; U.S. Department of Commerce, Census of Manufactures, 1977, 1980). The availability of wood furnish is dependent on productive acti- vity in supply industries. Also, the trend toward non-residue sources is due to: (1) new sawmill technology that improves sawlog utiliza- tion, (2) new pulp and paper industry technology for utilizing resi- dues in production, and (3) increased use of residues as an energy source (Wilson, 1975). In addition, other composition board products are competing for many of the same raw materials. Though a large portion of wood inputs are now purchased, there is no satisfactory measure for the price trend in wood furnish for particleboard. However, as industry is forced to utilize more non- residue furnish, the furnish cost will increase considerably (Wilson, 1975). In order to estimate wood requirements, some measure of physical input to product output is needed. According to data collected by Dickerhoof (1976), 1.51 tons of wood raw material (dry weight basis) were used to produce one thousand square feet (3/4-inch basis) of par- ticleboard on average during 1973. Approximately 88 percent of the 81 Table 13. Type and percentage of raw material used by the particle— board industry, 1973. Type of Percent used raw by material total industry Planer shavings 65 Plywood mill waste 10 Sawdust 9 Roundwood 7 Chips (pulp—type) 5 Slabs, edgings, and 3 trimmings Veneer core 1] Other 1 l/Less than 0.5 percent. Source: Dickerhoof, 1976. 82 furnish was softwood. By assuming the input mix does not change appre- ciably, this factor can be used to estimate wood requirements given panel consumption. Several measures have been developed for converting material requirements to cubic foot measures, but they are not used in this study (Risbrudt, 1980; Wright and Phelps, 1967). As mentioned earlier, the end of resin shortages after World War II made more resins available to manufacturers. The price of adhesives and resins relative to all commodities (1967=100) was 100.2 in 1979 (see Appendix). However, this does not reflect the trend that has persisted since 1972 which shows adhesive prices have increased 34 percent more than all commodities. The delivered cost of resins in- creased at the same rate as the cost of all materials used in the par- ticleboard industry between 1972 and 1977 (see Table 14). Continued materials cost increases will exert upward pressures on particleboard prices. Average hourly earnings of production workers did not increase as rapidly as adhesive prices from 1972 through 1979 (see Table 15). The number of production workers and the total hours they work is deter- mined by economic activity in end-use markets and the degree of mechan- ization in the industry. Particleboard output per production hour worked (3/4-inch basis) was 226 square feet in 1972 and 330 square feet in 1977; labor-saving process innovations being adopted by in- dustry are responsible for part of this increase. Energy costs have risen sharply in recent years. As the cost of energy enters in at each stage of wood harvesting and processing, final product prices are expected to rise. However, energy also influ- ences the cost of competing materials. Therefore, the net effect of 83 .nowm4 .mnm4 .meueuuemeeez mo memeeo .moeefifioo mo uemEueeemQ .m.D “mousom o.m4N m.o4 4.5m nmm4 q.wm4m N.m w 4.Nmm Nmm4 Amue44oe c0444eev umoo mm44meem Amee44ow e044448v Amee44ow e04444EV wee .muoefieueoo meHmeu wfioe eeu meemeu .m4e4eoue8 eeeuo wee 044oeeem eefiae4ea wee ewe: 4euOH umoo weee>44en new» .eeme use Name .ANmsN oemv enumaeae eemoe Ie4o4ueee one On umoo me44eaem wee .memeHeueoo .m4e4eouea 4eu0u wee meemmu mo mumoo weum>44mn .q4 o4neH .4owa4 .mnm4 .meeeuoeweeez mo memeeo .eoeeEEoo mo uemaueeeea .m.D “moeeom 84 mm.m m.o4 o.m Nmm4 mo.m m.m o.c emm4 oo.q ¢.w m.q m>o4 oo.q w.N4 4.0 cum4 Nm.N w.m4 m.w m>m4 mm.mm w.m4 «.0 NN¢4 Awee44owv meexeo3 Ameow44fiev Aooo4v eowuoewoue meexeos meexeos mo mwefieeee eOHuoswoea >4 eOHuoewoue >4unoe eweeo>< wexuoa mueom mo penaez uee> .Nmo4 eweousu Nma4 .ANoeN UHmv >uumewe4 wue04e4oflupee mew e4 ewewewee >4eeoe emeee>e wee .wmxu03 mueoe .muoxuo3 eowuoswoum .m4 m44eH 85 the rising energy cost is difficult to determine (Wisdom, 1977). In the short run, rising energy prices will have two effects on the particleboard industry. First, the industry will attempt to reduce consumption of purchased energy. This is borne out by available data on purchased energy per unit of particleboard output and by industry efforts to become more energy self-sufficient (Egan, 1980; TenWolde, 1979). Second, other wood processors will begin competing for residues as a source of energy (Wilson, 1975). Though competition will cause residue prices to rise, significant quantities of unused manufacturing residues still exist. For example, there were 547 thousand cubic feet (solid basis) of unused manufacturing residues in 1976 while only 330 thousand cubic feet of particleboard furnish were consumed (U.S. Depart- ment of Agriculture, Forest Service, 1979). As presented earlier, recent trends in capital expenditures have been primarily in the area of new machinery and equipment. However, considerable capital has been invested in the form of new plants over the years; 72 plants were operating in 1979 as compared to 23 in 1955. Southern Pine Plywood Raw Material Situation As was the case with particleboard, raw material availability has played an important role in the growth of the southern pine plywood industry. In the South, the net volume of softwood sawtimber on commer- cial timberlands increased 39 percent from 1962 to 1977 despite the growth of the plywood industry (U.S. Department of Agriculture, Forest Service, 1979). Volume increases were recorded in every diameter class. Net softwood sawtimber growth to removal ratios, however, have been declining, especially on forest industry lands (see Table 16). 86 Table 16. Softwood sawtimber net annual growth to annual removal ratio on forest industry and all commercial timberland in the South, 1962, 1970, and 1976. Year Net growth to removals Forest All southern industry commercial lands timberlands 1962 2.14 1.65 1970 1.29 1.42 1976 .91 1.28 Source: U.S. Department of Agriculture, Forest Service, 1979. 87 The result may be increased competition for softwood sawtimber on other lands and increased stumpage costs. In fact, average 1978 stumpage prices (in 1967 dollars) for southern pine sawtimber sold from National Forests in the Southern Region were 33 percent higher than prices paid in 1969 (U.S. Department of Agriculture, Forest Service, 1979). South- ern softwood plywood prices in 1978 were only 15 percent higher than the 1969 level. Table 17 presents Bureau of the Census data on log consumption for plywood in the South. The data are based on local log scales and are not strictly comparable over a long period of timecl/ In general, there is no discernable trend in the log requirements per unit of out- put. By assuming an average specific gravity of .51 for southern soft- wood species and an additional 5 percent of veneer volume that will be compressed into a 3/8-inch panel, approximately 1,168 pounds of wood (ovendry mass and volume at 12 percent moisture content) are re- quired per thousand square feet (3/8-inch basis) of southern softwood plywood (U.S. Department of Agriculture, Forest Products Laboratory, 1974). Bolt diameter, core diameter, and processing efficiency will deter- mine wood requirements in terms of softwood bolts. As bolt diameter decreases and core diameter increases, more bolts are needed to pro- duct 1,000 square feet of 3/8-inch plywood (Koch, 1972). By maintain- ing a stable bolt diameter and reducing core diameter, greater yields per bolt and improved lathe productivity are possible. 1/ —-Comparability problems resulted in discontinuing this data series in 1979 (Ambler, 1980). 88 mom. 44m.mq>.N qwm.wmq.> mnm4 mom. mmN.an.N oww.¢w>.w o>m4 mmm. w>w.omm.4 an.wmq.m mmm4 4mm. emo.me.4 emm.m4m.q e>m4 mom. «M4.qoo.N mwo.>me.m mnm4 «mm. www.mqm.4 4¢m.ooN.m N504 mom. emq.mmm.4 owm.N4m.q 4mm4 Nmm. mow.qMN.4 oqw.m4m.m Oma4 mum. mw>.mmo.4 «mo.Now.N oom4 Nwm. mqo.mmw mw0.mqm.N wom4 4mm. moo.Nmo mmm.mo>.4 nom4 4mm. om4.omq 40>.mmo.4 wwm4 mum. 4mm.4q4 moo.mnm m©m4 04uee Ae4eom w04 AmHmen noeHIw\m .umem eOHuoewoum On .ueew weemsoeuv eeeevm weemeoeuv eOHDQEemeoo mo4 eowuaaemeoo wo4 coeuoewoea wooB>4m pee> .wnm4Iewm4 .e04ueaemeou w04 wee e04uoewoua wooB>4e eefim 3044e> eeenueom eeeaumn eomfieeeaoo .m4 044eH 89 .Anvmeee> weuoe4em .msmemo men 40 eeeeem .eoeeEBoo wo ueeEuueeen .m.D "moeeom 4mm. qu.w>m.N 0>4.mm>.n w>m4 4.e.ucoov 44 e4nee 90 The net effects of increases in phenol-formaldehyde resin prices, labor costs, and energy prices on the consumption of plywood are diffi- cult to determine a priori. While southern pine plywood utilizes more expensive resins, resin content per panel is lower in plywood than in composite panels. On the other hand, Douglas fir plywood requires less expensive urea—formaldehyde resins to obtain comparable interior grade qualities; therefore, southern pine plywood is at a cost disad- vantage in that respect. Lower transportation costs lessen the effects of the resin cost differential though. Historically labor costs in the South have been below the national average. The annual average rate of increase in weekly earnings, how- ever is greater than the national average (Irland, 1972). Over time, this will tend to reduce southern labor cost advantages. Labor wages are influenced by productivity. In the veneer and plywood industry, output per production hour worked has been above the average for all manufacturing. Process innovations have contributed to this trend (Farris, 1978). Hence, productivity gains may offset rising labor costs. The effect of increases in energy price on the southern pine ply- wood industry differs from the effects on the particleboard industry. For example, residues are now being used as fuel for dryers reducing dependence on purchased energy as is the case with particleboard. Com— petition for wood raw materials, though does not increase as a result. By 1979, seventy-two softwood plywood plants were operating in the South. This represents a huge influx of capital into an industry that began in late 1963. This capital has created large, efficient mills with modern processing equipment. Innovations continue to be 91 incorporated in new plants as they are built. As a result, the south- ern plywood industry competes successfully with other panel products industries. MODELLING PRODUCT DIFFUSION Prior to analyzing diffusion in terms of market shares, the pro- ducts to be aggregated were selected. First, general wood—based panel end-use markets were determined for panels with available consumption or production data. Second, structural demand equations were estimated for particleboard. Finally, panels were aggregated to form the total panel market. Aggregate Panel Market Submodel Table 18 presents an overview of panels and their end-use markets. In many cases panels can act as substitutes for one another, but they can also be complements and inputs into other panel products. For example, hardboard and particleboard are substitutes as floor underlay- ment. Insulation board and hardwood plywood can be complements in the sense that a wall system may have a finished interior wall hardwood panel backed with insulation board sheathing. An example of a panel as an input, is the use of particleboard as a core material in plywood sheets. Thus, the relationships among panels are complex. Since particleboard is one of the panels being examined in this study and because it has numerous end-uses, it was selected initially as the panel on which to base the total panel market. Table 19 presents the estimated parameters of the linear and log-linear structural demand equations. As expected, the positive signs associated with hardwood 92 .nmo4 .>oeeeo mowo4 .eomemwe< "meoeeom 93 .oeeuHeuew .Ame4w4mv eoeuoeeumeoo 4e4ueewwmee 3oz weeonee44w >u4meew Eefiwez .Amwooe .m44e3v eOHuoeeumeoo 4e4ueew4mee 3oz weeon eOHue4eee4 .menufieeew mmeOHunwe wee me4eeee .eeexe: 4e4ueew4moe “Am44e3 .me004mV e04uoeuumeoo 4e4ueew4mme 3oz wueoeweem .eeeuweeem mme04u4wwe wee eeHeeee .eeexee 4e4ueew4mee unwe4>4eem .me004mv e04uoeeumeoo 4e4ueew4mee 3oz weeone404ueem .eeeu4eeem ”meowufiwwe wee me4eaee .eeexee 4e4ueew4mee "Am44e3v e04uoeuueeoo 4e4ueew4mee 3oz w003>4e woo3weem .me04u Iwae wee meeeeee .eeexe: 4e4ueew4mee mAmMOOM .m44e3v e04uoeeueeou 4e4ueew4mee 3oz wooa>4e ween eeeeueom .e04uoeeumeoo 4e4ueew4meeIeoe “mumxeee 4e4eueewe4 meeeu4eeem mmeowuewwe wee meHeeee .eeexe: 4e4ueew4mee ”Awe4>4eem .m44e3 .mmooe .weHme .me004wv eowuoeeumeoo 4e4ueew4mmu 3oz wooB>4e e44 ee4weoa muexeea emeIwem 4eeem .muexeea emelwee wouue4em wee m4eeea womenlwooz .w4 e44eH 94 e0w NN e44eH mom .mmm4 emeoenu mwo4 Eoem euew SD43 weuuem meowueevm . omen eo ANV e04ueeve eem\fl co.n o~.mI ueeumeou am. om. Aqxv e4 em. Ho.N Amxv :4 mm. mq.4 ANNV e4 a... N3- 25 a E 5 \mamaees as 04.4mMN ow.NqomI ueeumeou Aexv esooefi o4.N aw.m 4eeomeea e44emoemww .m.D Amxv eo4ee no.e am.em ecossea eoosoeom ANxv e04ee cw.q4 mm.mm wooS>4n4 woozweem A>v >e4ueeev em.w wN.emI Alec manna eeeone4oeneem eeeopm4ueueee \Meemane 41V a Any Axe Awe mo eoeee ueefiowmmeoo me44e4ee> e4ne4ee> weeweeum eoemmeemem >eoueee4exm ueeweeeea eOHueevm leeooew wee meofiea uoewoee eo women meOHueeve weeEew 4eeeuoeeum eeee44lw04 wee eeee44 wmweawumm .m4 e44e9 95 .emem uoeeeoo eeu we4>ee mo >u4444e40ea peeoeee om ee umee4 ue e>e4 mueeHUHMMeoo wmueawuee 44< “euoz BA .35 .8. MI II M .309 aqao aq.N .mueeaeuumeH >4eaem A.w.ueoov m4 e44eH 96 and softwood plywood price indicate that they are net substitutes for particleboard. Hardboard and insulation board prices were included in test runs of the model, but were deleted because of (1) their col— linearity with each other, (2) hardboard's collinearity with hardwood plywood, and (3) little additional variance being explained by their inclusion. In order to accomodate the entire range of end-uses and the com- plexity of product substitution, all panels listed in Table 14 were included in the total market. Hence, the substitutability of hardboard and medium density fiberboard for particleboard in furniture as well as the use of insulation board for sheathing can be enveloped in the model. Apparent consumption data were developed for particleboard, hard- board, insulation board, and plywood other than southern pine plywood. For southern pine plywood and medium density fiberboard, production data were used in the model. This poses few problems since southern pine plywood is primarily a domestic product at this time and medium density fiberboard comprises a minute portion of the total market. Hereafter, both production and consumption data will be generalized as consumption data. Figures 4 and 5 present consumption data for the panels in the total market. The total market is growing over time but is very cyclic in nature. This is due primarily to the strong reliance of wood-based panels on the construction industry. Non-southern softwood plywood which is comprised mostly of Douglas-fir plywood is a mature product maintaining a high level of the total consumption. It is also charac— terized by a declining market share. Hardwood plywood and insulation board are also mature products with declining market shares. 97 Figure 4.——Total panel, non-southern softwood plywood, southern softwood plywood, and particleboard consumtion, 1955-1979. Legend: Totoal panel consumption (standard measures) Non—southern softwood plywood consumption (3/8—inch basis) Southern pine plywood consumption (3/8—inch basis) ——--—-——-——-—————. Particle board consumption (3/4—inch basis) —————————————————— Figure 4. 98 I 1976 1980 1970 .1965 YEHR 1960 1955 l 0000? T I 00008 ‘00003 I 00001 (311 N01111u1 NellaunSNoa °Isso 99 Figure 5.-—Total panel, hardboard, hardwood plywood, insulation board, and medium density fiberboard consumption, 1955—1979. Legend: Total panel consumption (standard measures) Hardboard consumption (l/8-inch basis) Hardwood plywood consumption (3/8-inch basis) —————————————— Insulation board consumption (l/2—inch basis) ——-——-——————__.__.___ Medium density fiberboard consumption (3/4-inch basis) —————————— 100 Figure 5. o co '2 L 1' i“ \ 1 E- \ 1 1'2 I “m / i 7 \ J . \l‘l :o \‘.I —§ Xi: : l _,, a: 1 ~15 1" - >- h \ . I -D I'l '3 r ...-a 1‘ . '17 I1 '3 a: [1. L. o to I ' I ' I ' I 4 ‘33 00007 00008 00003 00001 0 (31:1 NOI'I'IIH] NOIldelSNOIJ 101 Southern pine plywood, particleboard, hardboard, and medium den- sity fiberboard are increasing in terms of consumption and market shares. The first three panels are substantially into the growth stage while medium density fiberboard is probably in its introductory stage. The diffusion of particleboard and southern pine plywood through the market is illustrated in Figure 6 and 7, respectively. An aggregate panel consumption submodel was estimated with ordinary least squares using total panel consumption as the independent variables. The final form of the model was: C = -184025 + 32220.1 x ln(DPI) R2 = .95, D.W. = 1.21 (9496.55) (1464.30) where: C = total panel consumption (standard measures) DPI = U.S. disposable personal income in billions of 1972 dollars Since the presence of autocorrelation is indicated by the Durbin—Watson statistic at the 5 percent alpha level, the equation was re-estimated using the Cochrane-Orcutt technique. Using the one—tailed t-test, both coefficients were significant at the .0005 alpha level. Estimated coefficients changed relatively little so the above model was retained for projection purposes. The model fits the long-term trend data very well, but performs poorly during periods of great market fluctuations. This was evident when estimates for total panel consumption were compared to actual consumption during the 1972 to 1975 period. 102 HRRKET SHRRE iPERCENT) l i Q I B a .1 . 1050 Y Y‘V’ 1 ff 7 V ' V I I V I V V V V I V V V ‘ I 1980 1005 1070 [075 1000 YEHR . .4r . 1955 Figure 6.-Market share of,particleboard as a percent of total pgnel consumption, 1955-1979. HHRKET SHRRE (PERCENT) 31 0 124 0"° 4} 0 9: ° 0 0’ In- ‘0 6 ‘ad-w-r-HF-‘L I . A. . v , v r .4 I r . arij 1900 1908 1070 1078 ' 1000 YERR Figure 7.--Market share of southern pine plywood as a percent of total panel consumption, 1964-1979. 103 Market Share Submodels Four market share submodels were estimated to examine different aspects of the diffusion. The logistic function and Gompertz curve submodels were used to estimate the time trend in market share dif- fusion, the growth rate of panel diffusion, and the inflection points. These trend models were used to project future market shares in the next chapter. Two multivariate models were estimated to examine temporal and spatial aspects of diffusion. Logistic Function and Gompertz Curve Submodels The market share data in Figures 5 and 6 were fitted with a simple linear trend model using ordinary least squares and with the nonlinear submodels using nonlinear least squares. The regression results are reported in Tables 20 and 21. The fitted nonlinear functions are illustrated in Figures 8 through 11. Several general conclusions can be derived from the regression results. First, the S-shaped models represent the diffusion process better than the linear models. Second, the logistic function is slightly superior to the Gompertz curve for the particleboard data, and the opposite is true in the case of the southern pine plywood. Third, as espected given the model formulation, the potential market share (P) and the period of time (growth rate) required for a panel to grow from 20 percent to 80 percent of its potential are greatest for the Gompertz curve. Finally, the growth rate of southern pine plywood was approximately twice as rapid as the particleboard rate. In fact, the logistic function proportionality constant (b) which determines the diffusion rate is substantially greater for southern pine plywood than for particleboard. 104 .N.>4 weueemeeeme m4 eeecm uexees e4 eoeuewee> wee4e4axeeb .mN n m>m4 .... .N u cmm4 .4 u mmm4 me weeemeee m4 e849 .> + Aeefiuv m + < u oeeem uexeez ”weeeu mafia neme44 \N .eemeeueeeea e4 meoeee weeweeum \fl 4Nooo.v Amoo.v omo. es. «soc. moo.OI eemoae4unneme mean I mmoe Ammo.v Ammo.v «mm. as. «no. mee.I eooae4m meme I some oenmeomom 4.3.av m a Na eomuezleweeen meeueaeeem ee>u 4eeem meee> .weeone4o4ueee wee wooB>4e eewe enmeueom pow .Iweeeu mafia eeee44 40 \N MmeeueEeeee weueEHDem .oN e44eH \ 105 .Amv 4e4ueeuoe mo ueeoeen om Cu ueeoeee ON Eoum zoom Ou wee4m nexeea ecu weefieuoe meee> wo eo4aez.l \N .o4nmnm4 .... .Numwm4 .4uewm4 ee43 e>ueo Noumeaoo mean eee4ueom e4u eom neeoxe mNummm4 .... .anmm4 .4nmmm4 we weeemeee m4 e849 .memeeueeuen e4 eeoeee weeweeum\w aemo.v 4N4o.v Aoeo.v w wwm4 mm. mm. «mm. «mo. mNN. eueeeEoo mmm4 I qwm4 Asso.v Amoe.v Amoo.v .eooseea m @494 mm. mm. mum. oqw.ml 44N. o4um4mo4 ween eee4ueom Aeeo.v Awoo.v Ammo.v m4 nwm4 no. me. Now. ONo. 4q4. eueeano m>o4 I mmm4 A4No.v Amm4.v Awoo.v .weeo4 an meme mm. mm. 4mm. Aeo.mI 4mm. geomewoe Ie4ueueee Ameee>v Auee>v leuee uefioe oflumeueum 4 e m meee> mmzoeo e04uoe4we4 NM eomuezIe44ueo meeueaeeem eOHOUeem \uoewoem .wooB>4n ween eee4ueom wee .4N o44ea weeo4e4owueea now m>eeo wueeeEoo wee eowuoeew o4umwm04 e40 mo \flmuoueEeeem weueafiumm 106 NRRKET SHRRE (PERCENT) 0 firY r‘rT' 1080 . t '1385 YEAR 0 I j I I l I I I I 1 I I V V I I' I 1050 1055 1070 1078 1000 Figure 8.--Logistic function fitted to particleboard market share data, 1955-1979. HRRKET SHRRE (PERCENT) 0 r 1 T f fi'j’ T I I l I I I I V I I I l 7 I F" I T I I I r 1 1050 1055 1000 1005 1070 1075 1000 YERR Figure 9.--Gompertz curve fitted to particleboard market share data, 1955-1979. 107 31 A h- 53‘ (J M u: a.m. u: m: C :0-4 m— .— u: g .1 mo ,1: a rfirg' Y I I I '1 Y I I I Y rwfij £980 £988 1970 1978 1900 YEHR Figure 10.--Logistic function fitted to southern pine ,plywood market share data,,1964-1979. IO- N i- z O- “J N U C LIJ e .. LIJ m C I O-I a) .- .— DJ x '1 $10 I: a I —r A I T I I I ‘ V I ‘I‘ T r I V _f I I 1000 1005 1970 1070 1900 YERR Figure 11.--Gompertz curve fitted to southern pine plywood market share data, 1964-1979. 108 Based on the Durbin-Watson statistics at the 5 percent alpha level, autocorrelation is present in all models. For projection purposes, this does not pose a problem. It is interesting to note that the 1979 actual marekt share has surpassed more than 80 percent of the estimated potential in all cases except the Gompertz curve for particleboard. The implications of this result on projections are discussed in the next chapter. Multivariate Diffusion Submodels While the univariate models are useful in depicting temporal diffu- sion trends, they are not very helpful in understanding the role supply and demand factors play in determining panel consumption and diffusion. Two multivariate models were used to gain insight into the effects of supply and demand factors. Time series data were used in the first model and cross-sectional data were used in the second model. Estimated structural supply and demand equations were used to screen variables for the time series model. Tables 22 and 23 present the structural demand and supply equations for particleboard and southern pine plywood, respectively. Though the results must be interpreted with the power of the t—test as a re- striction, the signs of the coefficients were generally the expected signs. For particleboard demand, quantity demanded is a positive function of one and two unit housing starts, mobile home shipments, and furniture manufacturing. These relationships are expected since they are the primary end-use markets for particleboard. Particleboard price and housing starts with three or more units exhibited the opposite 109 .e04ueede >4eeem e40 e4 ueeumeoo e4u ueeoxe ewfie uoeeeoo e40 we4>e4 mo >u4444e4oee ueeouee cw ee umee4 ue e>e4 eueeHUfimmooo woueawume 44< neuoz mm N I 3 a me I NIexm sm.~ I .3.9 .am. I I NI \N .aeae peaches aeoe some meme eons emuuee meannesumxw mm.wmwm mm.4wmI ueeumeou w¢.wo4 4m.mqm eowuoewoea ee4ee4 cm.e4m4 mn.m4wNI eewez mw.m wm.e4 eofiea >weeem mw.>4 mm.mNI mowue ee>4me4w< >uwueeev N4.44 m4.NmI eofiee weeo4e4o4ueem weeo4e4o4ueem \m>4eeem ANV mo.oom cq.wmmNI neeumeoo wefieeuoemeeea >>.4 «4.wN eeeufieeem mo xeweH Nm. No.0 mueoae44m e804 e444oz mm. ww.mI mowed +m .muueue weemeom 4N. ww.4 mDHee N44 .mueeum wefimeom >ufiueeec mm.4 mm.mI eoeee weeo4e4oeueem weeo4e4040eem \Mweeemm 44v 4 43 cc CC 40 eoeue ueeeoHMMeoo me44e4ee> e44e4ue> weeweeum e04meeewem >uOueee4exm ueeweeeea eowueevm .weeo4e4owuuee e04 meOHueeve >4enem wee weeaew 4eeeuoeeum woueaeumm .NN e44eH .I c C . ......L. .em4e uoeeeoo e40 we4>e4 mo >04444e4oem ueeoeoe om ee ueee4 ue e>e4 eueefiofimweoo weueSHume 44< “euoz llO cm 4 I 3 4 .mm I NIm\m OHNI SQ mm INI\M .meme canoes“ some acne meme eons eeooee economasmxw >¢.wmq4 om.NNmm4I ueeueeoo me.oe Na.¢oe su4>nuoseoea 44es eooss4e eoflea eweaeeum 4N.m4 mq.m4 w043em ween eee4ueom e04en >ueueeev wooB>4n w>.m4 em.mm woos>4m ween eee4ueom eefia eee4unom \m>4eeem ANV 4q.wmn mm.4omI neeumeou we. 4m.NI muwee +m .euueum weemeom ow. «4.e muHe: N44 .mueeum we4meom eofiee em.44 oq.wm wooB>4e wooaumom eeeumez eofiee >u40eeec wooa>4e oq.q4 nw.NwI woos>4e ween eee4ueom mean eee4ueom \MweeEeo 44v a 43 co :0 mo eoeee ueeHUHMMeoo we44e4ee> e44e4ee> weeweeum eOHeeeemem >e00eee4exm ueeweeema eowueecm .l.wooB>4e eeHQ eee4ueoe pom me04ueeve >4enem wee weeEew 4eeeuoeeum weueEHDmm .MN o44eH \e lll relationship. As price increased, ceteris paribus, quantity demanded decreased. This is the expected demand price-quantity relationship. Multifamily units utilize only a small portion of particleboard rela- tive to other end-use markets (U.S. Department of Agriculture, Forest Service, 1974). The direct competition of multifamily units with mobile homes and one and two family homes leads to this statistical result. As expected, particleboard quantity supplied declines as wages and adhesives prices increase. Lumber production was used as a proxy for planer shavings availability and has the expected positive relation- ship. This result must be viewed with caution since residue price would be a more appropriate supply variable and because solid wood products production is positively correlated with residential construc— tion. The net effect of increasing energy prices on particleboard was positive over the test period. In supply functions, increases in board price are associated with increases in quantity offered in the market. The only unexpected sign was the negative sign on particle- board price. This perverse relationship may be due to the general negative trend in particleboard prices (see Table 9) and the predomin- ance of demand in determining price—quantity relationships. Sheathing is the primary market for southern pine plywood, and as a result, one and two unit housing starts are positively correlated with plywood quantity demanded. The role of western softwood plywood as a substitute for southern pine plywood is indicated by the positive sign on the former price. As was the case with particleboard, multi- family housing starts have a negative effect on plywood quantity demanded. The southern pine plywood supply equation had the expected signs on price and productivity variables. As either price or productivity t1 9C. St of at b1; 00 re 9X St in dw Va 1 112 increases, producers offer more plywood in the market. As the price of inputs increases, a negative effect on quantity supplied is expected. The sign on stumpage price was positive; however, it was only signifi- cant at the 20 percent alpha level and the power of the test was very small. Despite the statistical result, stumpage price was retained in the equation because the cost of stumpage (l) is an important factor in the total cost of producing the panel and (2) is expected to increase in importance as competition for stumpage intensifies. The positive correlation between product price and stumpage price and the small sample size may have led to this result. The final supply and demand equations were derived through an iter- ative process in which many variables were examined for inclusion in the regressions. Additional variables considered in the particleboard equations were: weighted price of substitutes, total U.S. dwelling starts, disposable personal income, index of all manufacturing, value of residential construction, value of residential additions and alter- ations, value of private nonresidential buildings, value of public buildings, percent of all housing starts that were multifamily units, output per plant, softwood plywood production and railroad freight revenue per ton—mile. Other variables, in addition to those above, examined for inclusion in the plywood equations were: one unit housing starts in the South, percent of total U. 8. one unit housing starts in the South, five or more unit dwelling starts in the South, total dwelling starts in the South, personal income in the South, plywood mill wages, adhesives prices, and energy prices. The variables in Tables 19, 22, and 23 and lagged market share variables were tested as explanatory variables for the models described 113 by equations (7) and (8). The lagged market share variables were included to show the accumulated effect of supply, demand qualitative factors on product diffusion. In short, they were used as proxies for past market success and the knowledge gained by users as a result of that success. The regression results presented in Tables 24 and 25 corroborate the importance of past success and accumulated knowledge in the diffu- sion of the new products. In fact, the adjusted R2 for particleboard and southern pine plywood were .94 and .98, respectively, without the end-use market variables included in the model. In the particleboard submodel, economic activity in the furniture and fixtures industry and the mobile home industry are important vari- ables for particleboard diffusion. This was an expected result since particleboard made substantial inroads in both markets. However, the combined effects of those two variables were small when compared to the importance of the lagged market share variable. The effect of using seemingly unrelated regression estimates was a slight shift in the estimated coefficients and a reduction in the magnitudes of the standard errors. Based on that result, the ordinary least squares particleboard model was estimated for the 1956 through 1/ 1979 period. The resulting regression—- was: M= —.011552 + .719247 x M(—l) + .000216 x IMF + .000027 x MOBILE (.005083) (.073876) (.000085) (.000007) where: M= particleboard market share (in decimal form) M(-l) = particleboard market share lagged one year 1/ — Standard error in parentheses. . 140.4:3334044L1 . - .h— Lnyh. F Av 14:5» , ~44m flung—IT. .. I~—WL~—L —.~.\Ln~n~:_.I~L IlsquL—.>.— a - I. ,I. . -—E n..P—~ Kr. .J. I r LAIVL.:=FNI~n.A~ . ., Tufluh ~2:.~. HJ..~ a .. QN 0447.4 E .eme uoeeeoo e4» we4>e4 mo >04444e4oee ueeoeee ow ee umee4 ue e>e4 mueefio4mweoo weueawueo 44< "euoz .me04ueeve eme4u e4 eoeoeemmew ueeUHMHeme oe e4 eee49.l 114 \N .om.N u .3.4 .mm. u NIm .mnm4 4weoe4u eom4 aoew euew eo wome4 eeueafiumM\fl Onnwoo. NNON4o.I .w4wmoo. nqu4o.I ueeueeoo euee8e44m mooooo. NNoooo. mooooo. NNoooo. e604 e444oz wefieeuoemneea «moooo. qNNooo. wo4ooo. qNNooo. eeeueeeem mo xeweH Nwmmmo. ommN4n. nqNNmo. mNmN4N. eee4e nexeee wewwe4 4 mo 44V nee4o444eoo 4 mo 44V ueowowmmeoo noeee weeweeum eoemeeewem eoeee weeweeum e04mmeewem \Meeuea4ume .IeeueEHDme e044e4ee> eowemeemee weue4eeee >4me48eem eeee%Wm ueee4 >eee4weo >e00eee4exm . lweeo4m4ofiueea now 4eon4em eowmeww4w 4euomaeu .eueflee>4u4ea Q40 40 memumaeeea weueawumm .4N e44eH I 115 .ewHe uooeeoo e4u we4>e4 mo >04444e4oee ueeoeen ow ee ueee4 ue e>e4 mueewofimmeoo weueEHume 44< neuoz .meOHueeve eme4u e4 eoeeuewmww ueeUHMHeme oe m4 euo4H.l \N .4o.N u .3.4 .am. u NIm .mmm4 4weoeeu ewm4 Boem euew eo wome4 eeueefiuem\fl 4newoo. Nwm4mo. 4mmmoo. mwm4mo. ueeumeoo mu4ee mooooo. mmoooo. I o4oooo. mmoooo. I Nw4 .mueeum we4eeom moemmo. osqsoo.4 smeemo. oeseoo.e steam smegma eemwme a mo 43v oceeoneeeou a no Ape nemeuneemou eoeee weeweeum eOHemeewem eoeee weeweeum eoeemeewem Axv Mmeueefiume \INImeueEHuee me44e4ee> eOHmmeeuwe weue4eeen >4we48eem me4e54uee meeenvm ueee4 >eee4weo >eoueee4exm .I.woo3 I>4e eefie eee4ueom e04 4eon4em eowm3444w 4eeoeaeu .euewee>404ea e4u mo meeueeeeee wemmafiumm .mN m44eH 116 IMF= index of furniture and fixtures production (1967=100) MOBILE= manufacturers shipments of mobile homes (in thousands of units) The adjusted R2 was .99, and the Durbin—Watson statistic was 2.29. The results were very comparable to those derived from the shorter time period. In the southern pine plywood market submodel, the lagged market share variable was again the most significant variable. The negative sign on one and two unit housing starts was contrary to expectations. However, closer examination of housing starts and market share trends shows that the greatest increases in market share were attained during periods when housing starts were declining. This was due in large part to southern pine plywood producers' ability to cut plywood prices to a greater extent than western producers could. The interaction of factors influencing diffusion over time, in general, do not incorporate spatial aspects of diffusion. For example, local products have limited potential because knowledge of products is restricted to a given geographic region. By 1978, knowledge regard- ing southern pine plywood was fairly extensive. Therefore, a spatial diffusion model was examined to gain insight primarily into the effects of transportation costs on spatial diffusion. Distance from Portland, Oregon (DIST) in miles and the number of southern pine plywood plants in adjacent states plus the number of plants in the state containing the major portion of Rand McNally Major Trading Areas (PLANTS) were used as explanatory variables. The sample was developed by crossing the top 20 U. S. softwood plywood markets and the top 20 southern soft- wood plywood markets. The resulting sample size was 29. 117 The a priori expectation was that the percent (PERCENT) of total softwood plywood shipments attributed to southern pine going to a Major Trading Area would increase as the distance from Portland increases and as the number of nearby southern pine plywood plants increases. This model provides a rough indication of the regional advantage panel products can develop. The results of the regressionl/ were: PERCENT= -10.60 + .023 x DIST + 1.37 x PLANTS R2 = .83 (6.78) (.003) (.21) As expected, the regional proximity of the panel's producers and the distance from the major competing panel's supply region were positively related to the market penetration of the new product. Transportation cost is the main factor leading to these results. Two additional variables were tested for inclusion in the model. Total plywood shipments to the Major Trading Area in 1978 and total housing starts in the Trading Area's major metropolitan areas were used as proxies for market activity. Though factor market penetration may be expected in large, dynamic markets, neither variable was significant at the 20 percent alpha level. Inclusion of the variables did not improve the explanatory power of the model. Thus, southern pine plywood attained spatial diffusion which was determined to a large extent by its proximity to Eastern markets. 1/ -— Standard errors in parentheses. PROJECTIONS AND NEWER PRODUCT INNOVATIONS The projection model formulated for this study was that future wood requirements for a given panel project would be a function of future aggregate panel consumption, the paneksmarket share, and the average wood requirements per unit of panel output. Disposable per- sonal income was selected as the exogenous variable for the aggregate panel consumption submodel time was the exogenous variable in the nonlinear diffusion submodels. Wood inputs per unit of output, on a weight basis, were assumed to remain constant over the projection period. Bolt diameter, core diameter, and processing efficiency will determine wood requirements in terms of roundwood volume; one study assumes that wood requirements per unit of output, on a volume basis, will fall ten percent during the 1976 to 2030 period for lumber and plywood (Adams and Haynes, 1980). For this study, however, wood requirements wane projected on a weight basis. Rather than project consumption for particleboard and southern pine plywood based on both nonlinear submodels, the models were examr ined more closely to determine which best represented the diffusion pattern for each panel. In this examination, insight was gained into the usefulness of these models for projections under different conditions of data availability Following selection of the market share submodels is the pro— jection of panel consumption and wood requirements. Finally, 118 119 interacting factors that will influence future diffusion are dis- cussed, including newer product innovations. Comparison of Logistic Function and Gompertz Curve for Projecting Market Shares Market share projections based on the logistic function and the Gompertz curve are presented in Table 26 and in Figures 12 through 15. The Gompertz curve market share projections are always greater than the logistic function projections for the complete data sets. In the case of particleboard, the disparity between estimates will lead to great differences in projected wood requirements. For example, the particleboard projection for the year 2000 would be 23 percent greater if the Gompertz curve rather than the logistic function were used for projections. Since it is desirable to assess the reliability of these func- tions for projecting future panel market shares under different conditions of data availability, five time periods of varying length for each panel were fitted with the logistic function and the Gompertz curve. Regression results are presented in Table 27 for particle— board and in Table 28 for southern pine plywood. Preliminary results from these tables indicate that the logistic function yields a better fit than the Gompertz curve for particleboard and that there is little difference in the case of southern pine plywood. The adjusted R2, Durbin-Watson statistic, and standard error associated with mar— ket share potential (P) were the basis for the preliminary conclu- sions. By comparing the actual market share in the last year of a 120 Table 26. Market share projections based on nonlinear diffusion functions, 1990 and 2000 (in percent). Panel Function Year product 1990 2000 Particle— Logistic 11.00 11.07 board Gompertz 12.83 13.64 Southern Logistic 21.06 21.07 pine plywood Gompertz 22.19 22.28 121 .— z 4 u: U 1 K M L U (z (I J: (D I.— U K K (I I: o 1 I ' r ' l ' 1 ' '1 1950 1900 1970 1900 1900 2000 YERR Figure 12.--Particleboard market share projections based on the lggiStic function,_1980-2000. h 2 U U g U 0- U I z I I: (D h U x z m z o v T v I v If I I 1 I 1950 1900 1070 1900 1090 2000 YEHR Figure 13.--Partic1eboard market share projections based on the Gompertz curve, 1980-2000. 122 10 N ha Eiu U a: m 52 In a: a: to ma 1'- In K g. I: 91 .' r ' I I I ' fl 1930 1970 1900 1990 2000 YERR Figure 14.--Southern pine plxgood market share projections based on the logistic function, 1980-2000. I l I I I I I IIIII 3| '1' 3 I I ! ! I— . z:° ' U “' O u a: . In L a . . In a: a: I: o m c- .— o In a: gIo I: o I .* I I I ' I I I 1980 I970 1900 1990 2000 YERR Figure 15.--Southern pine plywood market share projections based on the Gompertz curve, 1980-2000. 123 Aoma.v Aomc.v Amoo.v mm. N©.N mmm. moo. NNo. Noumano Aomm.v Aooq.v Ammc.v no. mo.m Nmo.a ooo.~- omo. onumfimoa ommHImmmH Aomo.v Anqo.v Aaoo.v mm. Ho.a Nam. men. «mo. nomeEOo Ammn.v Amn~.v Aoao.v I um. um. mmm. omm.nu mmo. unumfiuoa «coaummma Ammo.v Aooo.v AoqN.v no. ma.n ego. mac. Nam. NoquEOQ Ammo.v Ammn.v Aqmo.v mm. HN.H awn. mmo.m- men. umumfimon meannmmon Aoao.v Awoo.v Ammo.v us. No. Nos. omo. Hem. NoquEOU Anmo.v Amma.v Aooo.v ma. No. HNN. Noo.mu can. ofiumflwoa mmmaummmn owumfiumum n m m N m comum3ICwnu3a mumumamumm :oHuoasm mumm» .mummm wmuomamm .vumonmaofiuuma new m>uso uuumaaou mam cowuocsm oflumwmofl msu mo twpmumamuma commafiumm .NN manmy \a 124 mm. mm. OH.H om.~ Ammo.v mow. nomo.v mum. Awfioo.v mooo. Anmm.v mow.MI .wmmmSucmumm CH muouuo wumvamum I \H Amoo.v mfifi. Nuumaaou Aqoo.v «ca. aflumfimoa mmmaumoaa A.u.uaoov RN magma Id in. .1151: r (46 H1 E : Iiiil‘t‘iaklltc T<;r£~1:h 3m Rik PK hurfi 125 ammo.v Awoo.v Amoo.v ma. om.H new. mac. «on. unconsoo ammo.v Ammm.nv Amoo.v mo. co.a owe. qmq.w- mm“. odomfimoo Nuaanoaa Aemo.v Aqno.v Aeao.v ms. ma. man. nmo. Hmm. Noumaaoo Ammo.v Ammn.v Amno.v so. be. mum. wwn.m- NnN. odomfiwon Neanuqoofi Amqo.v Anso.v Ammo.v om. wn.n mam. mmo. mmm. Noumaaoo Aemo.v Ammw.v Ammo.v om. mN.H “mm. okq.qs mNN. unomflwou memaumomm Ammo.v Ammo.v Aono.v mm. mm. «an. «mo. mNN. NumeEOu Aqqo.v Amoo.v Amoo.v ”a. me. mum. oem.mI Ham. unomfimon mnonlqomn UHumfiumuw n m m N M comumBIcwnusa mumumfimpmm coauocsm mummw .wooB>Ha mafia sumnusom MOM m>u50 NuquEoo cam cowuucsm owuwfluoa mzu mo .mpmmh vmuomamm Imumuwamuma wmumafiumm .wm manmfi \H «.3. at: U c mm ninth 126 .mwmmfiucwhmm CH kaHHw UHQUCNUW I \H Ammo.v Aqoo.v ono.v mm. mo.N mum. moo. Hmfi. Noumaaoo Aqmn.v Aqu.~V Amoo.v mm. ON.H gem. wmm.nnu man. usumflwoo okmanqoma A.w.u:oov mm magma 111 PU' pe St fi di fu C0: GOI ti. re; the 10; th. He: 80: to 801 3011 Squ 127 sample set with the estimated potential market share, the limitations of the nonlinear functions become evident. The come putations presented in Tables 29 and 30 show that: (1) over 85 percent of market potential was already achieved in cases when the standard error of the potential was small and (2) the potential market share was consistently underestimated in samples based on the first half of the available data. Therefore, the deterministic diffusion functions should be used for projecting only when the dif- fusion process is substantially underway. For most data sets tested, the logistic function yielded more conservative estimates of market share potential than did the Gompertz curve. After comparing the statistical results, the logis- tic function was selected for particleboard projections. Additional regression tests made on the southern pine plywood data indicated that the estimated market share potential associated with the logistic function was more stable than the potential associated with the Gompertz curve as the sample size approached 16 (total sample). Hence, the logistic function was selected for projecting future southern pine plywood market shares. Projections and factors expected to play a major role in slowing the diffusion of particleboard and southern pine plywood are discussed in the remainded of the chapter. Projections of Panel Consumption and Wood Requirements Based on high, medium, and low estimates of U. S. disposable per- sonal income, total panel consumption is projected to range from 48.5 to 52.5 billion square feet in 1990 and from 55.5 to 62.6 billion square feet in 2000. Historically, the greatest annual consumption ll‘ IN‘ ( ‘PC (‘0‘!!! t ’l“ (I [1 ‘11!!! “(I 111‘11‘1IFIE HMO (filfifls'b Cw 1<:C~J<< (1:4: IIJIII JEIF <44 <4 rI~4 1 1 a 1 . .1 J p w 128 .Hmwucmuoa wmumafiumm seas vmumfioommm uouum upmvnmum mwumA.I \m .Hmfiuamuoa mum:m umxume mp wmww>ww mumnm umxuma Hmsuom amok ummA\M .wOHuma mafia magamm mnmfilmmmfi mnu mo pmm% umma m£u ma mmmfi ..w.m .vowuma mafia mamamm mo pooh ummg\fl mm. o.m m.- uuuwanu mm. o.m q.oa uaumawoa mamaumoma mm. m.H N.N Nuuoaaoo mo. a.a o.N uaumawoa mmmaummma mm. o.m q.m muummaou as. o.m m.m usumawoa soaanmmaa \M¢N. o.w ~.om Noumssoo \mco. o.w N.qH UHumeoa manImmmH \mwo. o.m H.qH uuumaaoo om. o.m H.Ha uaumawoa mkoaummma wofiumu mum m umxuma Aucmoumav Auamoumav Hmwuamuoa ou .Impmnm umxume Hmfiucmuoa mumcm Hmauum ummm ummq Hmruom ummz ummq umxumz cowuuasm mummw .mmaaamm omuumamm cw wm>mfiaom mumnm umxume ummH onu ou Hmfiucwuoa mumsm umxums m.vpmonmaofiuuma mo somfiumgaoo .mm manna tea/cwficr. atria ...C—LCE to"; die r... Fevecctcc Extra t..§.2...E 0:25:57. cric :L..:::C.:. mc :CvgrrdECC -Cm. 3:295 129 .Hmwucmuoa mumnm umxume zn vowfl>fin mumcm umxuma Hmsuom amok ummq.l \N .vOfiumm mafia oHaEmm mnmfilqomfi osu mo pooh umma mnu ma mmmfi ..w.o .moauoa mafia mHaEmm mo umm> ummq\w mm. o.- H.m~ Nupanoo Ho.a o.Na m.aa anomawoa ohmauqoma mm. o.m~ «.0H Nuumasoo No.a o.ma m.ma uaumswoa msoanoma mm. m.oN H.mm Nuponeoo mm. m.om N.HN oflumfiuoq nnoHIqomH em. m.om m.m~ Nuumasoo am. m.om m.NN oHumHuoq mmmHImomH No. m.o~ m.~m nonmaaow no. m.o~ H.HN oaumfiwoa meaauqoma Wofiumu mumbm umxuma Assoopmav Aucmoummv Hmfiucotoa cu Imumnm umxume Hmfiucouon mumnm Hmsuom pooh ummq ammuom ummh ummq umxumz cowuocsm mummy .mmfiqamm wouomamm ca wm>mflnom mumsm umxuma umma mcu cu Hmfiucmuoa mumsm umxume m.w003%HQ mafia cumsusom mo :omaummfiou .om mHan an ar. of je pa pr 83 Se f( of 130 was in 1978 when 39.3 billion square feet of panels were consumed. Projected particleboard and southern pine plywood consumption and wood requirements are presented in Table 31. The projections are based on the logistic function diffusion trends and estimates of disposable personal income. The 2000 medium income level pro— jections are 82 and 62 percent greater than the 1979 volumes for particleboard and southern pine plywood, respectively. The same percentages apply to wood requirements on a weight basis. The projected average annual growth rates are well below those experienced in the industry during the 1960's and 1970's. The U.S.D.A Forest Service (1979) projected higher levels of particleboard consumption than were estimated in this study (see Table 32). However, panels other than particleboard were included in the Forest Service estimates. The Forest Service projections were based on price trends that have prevailed during the last two decades. The southern pine plywood production projections estimated by the U.S.D.A. Forest Service (1979) are similar to those developed in this study for 1990 (see Table 33), but all 11 percent lower in 2000. The Forest Service projections for plywood are based on stumpage prices that will equate stumpage quantity demanded to stumpage quantity supplied (U. S. Department of Agriculture, Forest Service, 1979). Other estimates of southern pine plywood production for 2000 are 9.9 and 10.2 billion square feet, 3/8-inch basis (Anonymous, 1981c; Haynes and Adams, 1979). The trend model estimates are derived from simple projections of consumption and an extrapolation of market share trends given m~o>v~ 30H tam . E: v 108 . ._Ma _ ”— hCL I -mu. UCLFCUL w 3 COL \N ~X9. v3 —G.LCL Tc: .1Cf~* Is foE-avhrhlkfx \a b 11.15%: I... II 1111.»! Ida] u 1.1 R— 131 . m m mm .I HN< an H m\m .ooos%aa mafia cumnusom so“ acoucoo mudumwoe unmoumm NH um manao> ocm moms huoam>o so comma .mGOu mo mommmSOSu ca ocm oumonmaoauuma MOM mammnAhuoco>ov uanmB >uo .mcou mo moammsonu :H.l \N .ooo3kaa mafia Ghmfiuflom How mammn LUCHIw\m 65m wumoanUHuqu How mflmmfl LUCHI¢\M .ummw mumdvm we mUGmeOSU CH\IHI mmmo comm amass comes 304 woossam meme mama magma mmooa answmz mafia mess Nose memH seems swam oaumfimoa comausom oemm meow ammo wmmm son sews wwmm mane mmmm assess women «seem oasw ammo woam awn: ouumawoa -maoauumm oooN Oman oooN oaoa \MAm>ma gaseous mufimfimhfidwmm @003 COHUQEDmGOU Hmcmm mEOUSH SOHUUGDm Hmfimm mHm>mH 30H mam .Esfioma .Smfiz “em a. .oooN mam omafi .maoocH Hmaomuoa mfinmmoamfio mo .Icofiumanmcoo Hmong mo mcoauomnoum .Hm manoH ucmEmufisumu ooo3 Hmuou oam \a 132 Table 32. Comparison of particleboard consumption projections made in this study with projections by the U.S. Forest Service (in millions of square feet, 3/4-inch basis). Year Logistic function trend Base level price . and medium income and medium incom level projection level projection— 1990 5555 7625 2000 6519 9020 l/ — Includes waferboard, flakeboard, composite board, medium density fiberboard, and similar products. Sources: Table 31; U.S. Department of Agriculture, Forest Service, 1979. 133 Table 33. Comparison of southern pine plywood production projections made in this study with projections by the U.S. Forest Service (in billions of square feet, 3/8—inch basis). Year Logistic function trend Equilibrium level and medium income level price projections projection 1990 10.6 10.5 2000 12.4 11.1 Sources: Table 28; U.S. Department of Agriculture, Forest Service, 1979. 134 market potential. The other projections involve estimates of future price and quantity relationships in the stumpage and final product sectors. The trend in the price-quantity relationships is implicit in the former and explicit in the latter. The panel consumption and wood requirement projections vary considerably depending on which model is used and what assumptions are made regarding model variables. Though the trend model is simple in design, it does indicate the direction and relative magnitude of future particleboard and southern pine plywood consumption. As was shown in the preceding chapters, many factors have influenced the diffusion of these panels. The ex- pected effects of these factors on future consumption and wood require- ments are discussed in the remainder of this section. Newer substitute panels are discussed in the final section of this chapter. Future Particleboard Trends Particleboard's unique properties and the flexibility of its manu- facturing processes will continue to make it an attractive panel for various end-uses. One example of particleboard's diversity is Mende Process board or thin panelboard. The Mende Process board was developed by Wilhelm Mende and Company in West Germany (Anonymous, 1971a). The process has been modified over the last 10 years, but the primary com- ponent is a continuous press. Like batch produced particleboard, Mende Process board is used as furniture core stock and in construction (Food and Agriculture Organization of the United Nations, 1976). It also is used for door skins and wall paneling (Bryan, 1978). There were eight domestic thin panelboard plants in production during 1980 (Anonymous, 1981a). 135 Improvements in manufacturing processes and the diffusion of exist- ing process innovations will lead to more efficient production in the future. Computerized process control systems, electron beam curing, and new binder developments are examples of process innovations that can make significant inroads in the particleboard industry (Blackman, 1978; Pease, 1977). Process developments will continue to play and important role in maintaining particleboard's competitive position. Declining relative price has been an important factor in particle— board's diffusion. Particleboard price and consumption are correlated with the prices of substitutes, disposable personal income, end-use activity, input prices, and raw material availability. These factors will determine future consumption levels. Traditionally, substitute prices have risen faster than particle- board prices. This trend is expected to continue in the future for softwood and hardwood plywood which rely on higher quality wood inputs. As income and end-use activity increase, particleboard consumption will also increase. Supply factors may constrain future particleboard consumption. Adhesive prices and availability are major concerns for producers. Though the net effect of energy prices was estimated to promote particle- board consumption, long-term effects may be negative. For example, multiple unit dwellings requiring less particleboard per unit could be the new housing norm in another 20 years. Increases in other input prices are expected to hinder future consumption. The availability and price of furnish is the problem most widely discussed by industry researchers. New sawmill production techniques are reducing planer shavings and sawdust per unit of output (Dicker- hoof and McKeever, 1979). Also, lumber production is increasing more 136 slowly than panel production and the available residues are more in demand as an energy source and for papermaking. Unused manufacturing residues, bark, logging residues, and roundwood are potential sources of future wood materials (Dickerhoof, 1976; Lehmann and Wahlgren, 1978; Lewis, 1965). Regardless of the source, greater handling and furnish preparation costs are expected (Wilson, 1975). The net effect is a slowing of the diffusion process as has been the case in recent years. Thus, the traditional particleboard product is nearing the apex of its market share. Both supply and demand factors have been instru- mental in its growth and will retard its future diffusion. However, newer panel products, such as, waferboard, medium density fiberboard, and composite panels are just beginning the diffusion process and will be the growth sector in the panel industry in the foreseeable future. They will play a major role in closing the projected consumption gap identified in Table 29. Future Southern Pine Plywood Trends Southern pine plywood manufacturers are diversifying their product mix. The characteristics required in CD sheathing were met by southern pine plys, and now higher quality plys are being marketed. Examples of higher value added products made from southern pine are: plyform, underlayment, 303 siding, and AC/AD panels. The addition of these products helps manufacturers offset rising log costs (Baldwin, 1977). Higher grade face plys also will be utilized in composiuapanels. Expan- sion into other grades is expected to help southern pine plywood cap- ture a greater market share in the future. Process changes will continue in the South so that producers can 137 utilize more of the softwood resources. Productivity gains reflect the modern, efficient processes incorporated in southern plants(Farris, 1978). In fact, capacity is still expanding in the South despite econ- omic conditions; Federal Paper Board recently announced plans for a new plant and International Paper has begun construction on another plant (Lambert, 1981). These plants are being built in anticipation of future demand. Substitution of other products for softwood plywood may dampen future diffusion of the product. In a recent survey (Carney, 1977), builders mentioned hardboard, aluminum, and phenolic waferboard as products replacing plywood in new residential construction. Shifts to other products in nonresidential construction may be due to econom- ics, fire codes, and other factors (Anderson, 1978). Siding and All-Weather' Wood Foundations are end-uses in which plywood may make significant progress. However, the concrete slab is the primary floor system in single-family homes in the South where southern pine plywood has its greatest acceptance (Carney, 1977). There- fore, North Central and Northeast markets have the greatest potential for southern pine plywood floor systems. The southern pine plywood price trend is projected to increase at about the same rate as western softwood plywood prices until 2000 (Haynes and Adams, 1970). As a result, southern pine plywood will maintain its in-place cost advantage in the eastern United States. Higher value added products, rising population in the South, and in- creasing end-use activity should combine to provide larger markets for southern pine plywood in the future. Supply factors will influence the diffusion of plywood, too. The 138 most notable factor is the price and availability of softwood stumpage in the South. Hair (1980) has estimated that stumpage price in the South will rise at an average annual rate of 2.5 percent between 1976 and 2030. Though this is slightly below the rate experienced between 1969 and 1978, it is still a significant increase. Upward pressure on stumpage price is predicted due to increasing demand for all southern forest products, expected declines in softwood roundwood inventories, and low levels of regeneration on private non-industrial lands (Anony- mous, 1981b and Hair, 1980). Eventually, the stumpage price increases will cause southern pine plywood to lose its competitive edge. Newer panels, such as waferboard and composite panels, will gain as a result. Newer Product Innovations Newer panel products that have been successfully introduced in the market give an indication of the direction of product innovation. As has been the case in the past, the new direction has important rami- fications with regard to wood inputs. The introduction of Douglas fir plywood in 1907 as a door panel has had remarkable consequences (Cour, 1955). Likewise, a faulty press valve in 1924 that resulted in hardboard first being produced during William Mason's lunch hour has had a significant impact on wood utilization (Maloney, 1977). Par- ticleboard and southern pine plywood are other examples of panels whose acceptance has altered perceptions of wood resources. Rather than forging a new direction, the most recent product inno— vations simply are continuing a process that has been underway for years, namely, the utilization of more diverse wood materials. From an economic standpoint, there is a dynamic shift from scarce materials 139 to less scarce materials. The increasing importance of southern pine plywood vis-a—vis Douglas fir plywood is a case in point. The use of manufacturing residues in particleboard is another example. Mere recently, waferboard, medium density fiberboard, and composite panels have added to the panel diversification trend. A conceptual starting point for future wood-based panel products is Dr. George Marra's (1969) "Non—periodic Table of Wood Elements." He identified 14 elements beginning with logs and ending with cellu- lose. Veneer, chips, flakes, particles, and wood flour are among the elements. By combining wood elements with other panel inputs, such as, binders and chemicals, boards with a variety of characteristics can be developed. End-use requirements and processing technology will be key factors in determining how these elements will be used in the future. The product innovations discussed in this section are not new in the sense that they were just recently introduced in the market, but they are new in the sense that they are only recently gaining wide— spread acceptance. Waferboard, for instance, has been manufactured in Canada since 1961 (Annonymous, 1962). Domestic medium density fiber- board production began in 1966, and composite panels were available as early as 1956 (Ananymous, 1980b; Elmendorf, 1956). Changes in standards, processes, panel prices, and the wood situation are factors leading to greater diffusion of these products. The newer panels primarily compete with existing wood-based panels. One indicator of the expected competitiveness of newer panels with existing panels is industry's capital expenditure commitments. New mills and modernization projects scheduled for completion in 1979 and announced specific projects requiring capital expenditures are listed 140 in Table 34. These projects accounted for over one—quarter of the committed $1 billion (Anonymous, 1980a). Waferboard was the most men- tioned type of new panel plant. The diffusion rates of these newer products will strongly influ- ence the potential diffusion of existing products, such as, particle- board and southern pine plywood. For example, waferboard is gaining acceptance rapidly in the Midwest. As a result, the southern pine plywood potential in that area may be reduced. On the other hand, waferboard and southern pine plywood may act in concert to reduce the market share of western softwood plywood. Since few quantitative data are available on waferboard, medium density fiberboard, and composite panels, a qualitative review is presented in the remainder of this section. The review will aid in assessing the potential success of these panels. Waferboard or Structural Flakeboard Waferboard or structural flakeboard characteristics can vary due to a number of factors. Some of the most important factors are: wood species flake geometry, flake quality, flake alignment, average density, density gradients, layer thicknesses, and resin content (Geimer and Price, 1978). Experimental flakeboard panels have been made with lodgepole pine, loblolly pine, sweetgum, southern red oak, Douglas-fir, sweetbay, red maple, black tupelo, white ash, post oak, yellow-poplar, and mixed species. Board properties for these panels are reported in General Technical Report WO-S, "Structural Flakeboard from Forest Residues" (U. S. Department of Agriculture, Forest Service, 1978). In general, 141 oususm osmonuomms muoxmn zosom .ou panama moawm oumSom coaumuficuoooa mama “swag eumonumwme m.z .smeumeo .eua .maaaz wooseuuoz owmfi cumonuowmz omnmso .muumm ma .ouq .couuom xofiEMoz owmfi oumopuommz oonmso .uo.o Hm> .ouq .xouon museum osmonuoumz oaumuco .ou monasq mmcwm cumsom meH oumonummmz mcflmz .ocmaoooa .auoo ofiwflommImeuomu mmmfi oumonpmmmz .mflz .oum3zmm .muoo owwaommlmcmwmflooq swam eumonummms .ccaz .xooo .auoo nuumHuom swam eumonummmz .casz .HneaEmm .auoo noumauom mesons eumonummmz .ccuz .Hmeaamm .eua .maaaz sooseuuoz owafi osmonuowmz \M.ccaz .flmanom .auou cameommlmcmflmfisoq cofiumNchmooa owmfi woman osmonummmz .CGHE .mofiamm ocmuo .ouA .muosoonm umouom .o.m coaumHano ucmam mo maze COHumooa zamafioo uamaomam omucsoccm cam mmmfi CH cowumHano .mmunufiocmaxm Houwamo wcwuflsvmu muoomopa mow omaoomsom muomHOpa soaumNficumooB ocm mHHfiE 3oz .qm maams 142 .mome .msoE%coc< "mousom .GOfiumuoa m>HumauouHm HOW wafixooq.l \s @maH mmofl owmfi Hwofi mususm Hmcma muwmoaaou Hmcma muwmonaou Hmcma m>ufimoq IEou .oumonmxmam oumon ocmuum omucmwuc coauMNchmooz uamHm oumonummmz .mp0 .mHHH>mc3oum .mp0 .umxmm .o.z .smaeso .m.z .ucoEmumHU .muae .mema m>mam .auoo osmonmam .ou umnEHH commcfiaam .auoo cameommImeuomo .auoo oumom umuoocmeam .cua .memamo mo wooseama A.u.ucoov em canoe 143 a variety of species can be utilized to create boards with aCCeptable stiffness and strength for various end-uses. In order to gain wide acceptance and to facilitate diffusion, code approval is needed. Canadian waferboard for sheathing was first approved by the New Jersey State Building Code in 1972. Subsequently, Canadian boards were approved by the Building Officials and Code Admin- istrators International Conference of Building Officials (Jorgensen, 1978). Roof sheathing and wall sheathing are approved uses. The American Plywood Association has begun to promote end-use or performance—rated standards. The new standards do not require parti- cular types of wood materials to be used in the panels. Instead, they specify performance characteristics. For example, sheathing can be conventional veneered plywood, composite panels, or unveneered panels such as structural particleboard, waferboard, and oriented strand board, so long as the performance standards are met (Lambert, 1980). As a result, waferboard and other newer panels will penetrate the market more readily. Waferboard production began in 1961. The first product was pro- duced by Wizewood, Ltd. at Hudson Bay, Saskatchewan and was sold under the trade name of "Aspenite." The process used in the first plant was developed by Dr. James d'A. Clark to utilize green poplar logs in the production of a weatherproof phenolic-binded board (Anonymous, 1962). Though waferboard production started in the early 1960's, con- struction on a second waferboard plant did not begin untill969, and the first United States plant did not come on-line until 1973. By the end of 1980, eleven plants in Canada and the United States with 144 1.25 billion square feet of capacity (3/8—inch basis) were expected to be operating (Pease, 1980). As indicated in Table 34, continued expansion of waferboard capacity is planned. The unique aspect of the waferboard manufacturing process is thick, manufactured wafers are used rather than traditional wood materials found in particleboard or hardboard. Since the wafers are large in size, less resin is required in the manufacturing process. In the case of structural flakeboard, the flakes are thinner than wafers, but no specific delineation exists (Dickerhoof and Marcin, 1978). Several existing means for manufacturing wafers and flakes hold promise for the future. These include the spiralhead clipper, the shaping-lathe headrig, disk flakers, drum flakers, and ring flakers (Arola, 1978; Koch, 1978; Price and Lehmann, 1978). The best method for flaking will be dependent on the species and wood form (e.g., round- wood, chips, etc.) used. One flakeboard variant is oriented strand board. Mechanical ori- entation of strands yields improved flexural strength in the aligned direction. Several existing waferboard plants have the capacity to orient wafers. The Elmendorf Board Corporation plant in Claremont, New Hampshire, will be the first plant constructed in the United States solely for the production of oriented strand board (Anonymous, 1979b). Regardless of whether the panel is structural flakeboard, wafer— board or oriented strand board, delivered price to major market areas will play a critical role in the diffusion of these products. In east- ern Canada, waferboard's delivered cost advantage over softwood ply— wood produced primarily in British Columbia is an important factor in its market success (Dickerhoof and Marcin, 1978). The regional 145 advantage is only partially offset by the greater shipping weight of waferboard. Feasibility studies indicate that flakeboard production sites in the Eastern and Southern United States have the greatest potential for competing with softwood plywood. Higher plywood prices in the eastern markets, lower freight costs for flakeboard from the East and South, and lower labor and raw material costs are cited as factors contributing to this potential (Adams, 1978). The locations of exist- ing and announced plants in Table 31 corroborate the eastern advantage. Aspen is the primary wood species used by existing waferboard manufacturers. Aspen has been a little utilized species in the past, but competition for resources is increasing. In fact, one potential manufacturer decided to look for a new plant site after a competitor announced plans for construction in the same timbershed (Lambert, 1981). There are, however, significant volumes of aspen in the East, especially in the Lake States. From 1970 to 1977, aspen and cottonwood growing stock on commercial timberlands in Lake States increased from 9.0 to 9.3 billion cubic feet (U.S. Department of Agriculture, Forest Service, 1974; U.S. Department of Agriculture, Forest Service, 1979). Aspen and cottonwood accounted for a greater percentage of Lake States' growing stock volume than any other species, hardwood or softwood. By considering other species for which experimental boards were tested, the potential wood resources for structural flakeboard can be expanded significantly. The outlook for waferboard is promising despite current economic conditions. For instance, a Lake States' waferboard plant that started production in late 1979 is now producing panels on a seven day, 24 146 hour per day schedule even though most panel plants in the United States are operating well below capacity (Pease, 1981). Desired board char- acteristics, code acceptance, delivered price advantage, and large volumes of potential wood inputs will contribute to waferboard's rapid diffusion. Lack of consistent consumption data precludes quantitative estimates of waferboard's potential. Medium Density Fiberboard The second panel product currently gaining acceptance in the mar— ket is medium density fiberboard. In the United States, two types of medium density fiberboard are manufactured: one type is for exter- ior siding and the other is for interior applications, especially fur- niture (Food and Agriculture Organization of the United Nations, 1976). The panels are composed of compressed interfelted fibers, binders, and other material added to improve board properties. For exterior siding, the boards are finished and coated for dura- bility. Due to board surface properties, a high quality finish is obtained. Competing products include plywood, aluminum, and other hardboard siding. Dry, wet, and wet-dry processes can be used to man— ufacture this board (Food and Agriculture Organization of the United Nations, 1976). In furniture applications, medium density fiberboard has a very smooth surface for finishing and good edge properties. The surface has less tendency for showthrough than particleboard, and the edge properties allow finishing without edge banding (Maloney, 1977). Particleboard is the primary competing product in this end-use. Dry processing with supplemental high radio frequency heating is a common 147 means for manufacturing the furniture board (Food and Agriculture Organ— ization of the United Nations, 1976). Numerous rocessing improvements have occurred in machining, sanding, laminating, embossing, and finish- ing medium density fiberboard (Anonymous, 1980a). NPA 4-73 "Standard for Medium Density Fiberboard," is the standard promulgated by the National Particleboard Association for this product. In addition, medium density fiberboard is included in the industrialite class of Voluntary Product Standard PS 58-73, "Basic Hardboard" and in PS 60-73, "Hardboard Siding." Physical property requirements such as modulus of rupture, thickness tolerances, and screw holding ability are included in the standards (Maloney, 1977). As mentioned previously, the first medium density fiberboard plant began production in 1966 and was located in Deposit, New York. Between 1969 and 1976, ten plants began production (Anonymous, 1980b). Thir- teen plants were in operation by the end of 1979 in the United States (Lambert, 1980). Production figures for the 1975 through 1979 period are presented in Table 35. Market share increased from 0.75 percent to 1.3 percent during that period. Though the diffusion process has begun, too few data are available for a quantitative estimate of med- ium density fiberboard's potential. Both hardwoods and softwoods are used in medium density fiber— board production. Ponderosa pine, southern pine, Douglas fir, and mixed hardwoods are used as furnish. Pulp chips, which are later fiber- ized, are the preferred form of raw material (Lehmann and Wahlgren, 1978). Therefore, producers must compete with the pulp and paper, particleboard, insulation board and hardboard industries for supplies. Medium density fiberboard is more expensive than particleboard 148 Table 35. Medium density fiberboard production, 1975-1979 (in millions of square feet, 3/4—inch baSis). Year Domestic production 1975 215 1976 280 1977 441 1978 480 1979 490 Source: Lambert, 1981. 149 on a comparable per unit basis. However, some furniture manufacturers are finding that it is very competitive in end-use cost. This is due to processing advantages such as machining, finishing, and shaping (Anonymous, 1980e). Computerized financial analysis is one tool being used to estimate end-use cost. As computer analysis increases, medium density fiberboard diffusion is expected to increase. The outlook for this product is good in siding and furniture appli- cations. Substitution for particleboard in furniture manufacture is the most promising market. Though there have been no new plants opened in the last few years, an economic recovery may lead to the construc- tion of new plants and an erosion of particleboard's market share. Excellent panel characteristics, standardization, and a competitive end-use price are factors that will have a positive effect on medium density fiberboard's diffusion. Composite Panels Composite panels were made as early as 1955; hardwood veneer faces and a resin-impregnated, softwood fiber core were used (Elmendorf, 1956). However, little interest was shown in composite panels until recently. With rising wood costs, products utilizing more low cost wood and retaining a high quality appearance are becoming desirable. Composite panels, which have particleboard or oriented strand board cores and veneer faces, fit these requirements. The composition core and veneer faces make composite panels slightly stiffer than plywood. The solid cores also reduce moisture penetration and subsequent cracking of plys receiving exterior exposure (Blackman, 1980). Manufacturing processes vary by producer; the components can 150 be layed up separately or in the same process. Composite panels meet the standards promulgated by the American Plywood Association for sheathing and flooring applications. Code acceptance has also been gained for some panels (McSwain, 1977). In those end-uses, the panels compete primarily with conventional plywood. Though price and production data are unavailable at this time, it is expected that the panel will be competitively priced. The use of low cost wood in the composite core is a major factor aiding this panel's ability to compete. By making composite panels, manufacturers are able to utilize the entire veneer log in one product. As a result, increased yields from a given supply of logs are possible. Lehmann and Wahlgren (1978) estimated that a 40 percent yield increase from a veneer log supply is possible. By using small logs and culls for the composite panel, one manufacturer estimated that composite panel output would be twice the plywood output for a given volume of logs (Bryan, 1981). Either oriented strand board or particleboard can be used as the core panel (Blackman, 1980; Lehmann and Wahlgren, 1978). Therefore, composite panels offer flexibility with regard to core materials and stretch existing veneer supplies. The first modern composite plant started production in 1975 and produced panels with softwood veneers and oriented strand cores (Anon- ymous, 1975). The same plant closed in 1979 due to environmental con- trol costs and poor market conditions (Lambert, 1981). Two addition- al plants began production in 1979 and 1980. In addition, the U.S. Forest Service and the U.S. Department of Housing and Urban Develop- ment cooperated in a research and development program for composite 151 panels in the mid-1970's (McSwain, 1977). The outlook for composite panels is excellent. By early 1981, nine plants in the United States, including those listed in Table 31, indicated they were producing composite panels. Given the economic climate, this increase in capacity is quite remarkable. Code accep- tance, resource utilization, and similarities to conventional plywood are factors that will promote the diffusion of composite panels. Other Factors Influencing Panel Diffusion Though a multitude of factors other than those discussed in pre- ceding chapters and sections may influence product diffusion, only five have been selected for brief discussion in this section. The five factors are: (1) formaldehyde emission, (2) market conditions, (3) U.S. Forest Service policies, (4) research and development, and (5) nonwood product competition. Other influencing factors such as antitrust laws, patents, capital availability, and internal company policies are beyond the scope of this study. Formaldehyde emissions during the production and use of panels, notably particleboard, have been acknowledged for some time. The specific effects of long-term exposure are not known with certainty, but some individuals are affected by the emissions. If regulations are mandated that limit allowable emissions, particleboard diffusion could be adversely affected. Publicity regarding formaldehyde release may have a similar impact on diffusion (Lambert, 1981). Foam insulation, however, may be affected to a greater extent due to its formaldehyde content. Adhesives product innovations, e.g., isocyanate-based binders, provide one means of overcoming diffusion barriers due to emissions 152 problems (Ball, Redman, and Adams, 1979). A second factor that may influence product diffusion is poor mar— ket conditions (Cox, 1974 and Dickerhoof, 1977). Plant managers may be unwilling to try unproven products during periods when economic activity in end-use markets is low. However, if the new product has been successful as was the case with southern pine plywood and wafer- board, diffusion proceeds. In essence, the new products may gain in market share despite the poor performance of the entire wood-based panel sector. U.S. Forest Service policies influence product innovation and diffusion through research activities and harvesting practices. Pro- cess and product research performed or funded by the U.S. Forest Ser- vice may lead to new processes, new products, or the dissemination of information (Fassnacht, 1964; Lehmann and Walhgren, 1978; U.S. De- partment of Agriculture, Forest Service, 1978). A less discussed im- pact on product diffusion may result from harvesting policies. If harvesting is restricted in a manner that increases final product cost, the diffusion of competing products is enhanced, ceteris paribus. Un- fortunately, forest industry generally views harvesting policies (e.g., the allowable cut) as factors constraining their growth locally rather than promoting their growth nationally (Lambert, 1981). Research and development is cited as another factor affecting innovation and diffusion (Cox, 1974). Research and development expen- ditures in the forest products industries are traditionally below those in other manufacturing industries (Risbrudt, 1979). No research and development expenditure data specific to the particleboard, southern pine plywood, and other new panel product industries is available, 153 but it is likely that expenditure levels relative to other manufactur- ing industries are low. Research and development activities by govern- ment, universities, the equipment manufacturing industry, and the adhe- sives industry may help offset these low expenditure levels and mitigate any adverse effects on the diffusion process. A final factor affecting product innovation, diffusion, and consump- tion is competition from non—wood products (Zivnuska, 1963). The prices of most competing materials have increased little relative to all goods since 1967 (U.S. Department of Agriculture, Forest Service, 1979). With the exception of western softwood plywood, most panel products have declined in relative price over the same period. Therefore, from a price standpoint, wood-based panels may be improving their position relative to competing materials. However, the substitution of other products for wood—based panels (see Table 36) will continue to be a major factor limiting the size of the panel market and the diffusion of panels within the industry. 154 Table 36. Examples of products substituting for wood-based panels in various end-uses. End-use Substitute material Floors and foundations Roof sheathing Siding Wall sheathing Furniture Concrete, boards Boards in single-family roof decks; concrete and metal in multifamily roof decks Brick, aluminum, stucco, boards, steel, plastic Plastic foam, foil—faced paperboard, gypsum products Plastics Source: Anderson, 1980; Anonymous, 1979a. APPENDIX DATA USED IN ECONOMETRIC AND DIFFUSION MODELS 155 «HH.H N H mNH.H Roma wma fl mom coma Now q mow momfi owe N wmo «omfi wmq H omq momfi woe woe Nomfi on on Nomfl moN woN ooafi omN ooN mmmfi omN omN wmmfi mwfi mwfi mmmfi HHN HHH ommfi On On mmmfi .Ifloauaasmcoo coauosoopm \N ucmpmaa< muuoaxm muuoaeH ofiumosoo you» . IAmHmmn LocHIq\m .ummw mumsvm mo mGOHHHHE ch manImmmH .coHuQESmcoo ucmumamm ocm .muuoaww .muuoaaw .CONuosooua oumonmaofiuuma .m.: .H< mHQmH 156 Nwm.m qq~.m ch.m mm~.m qu.N mom.N qmq.m owo.m Nwm.N mmm.fi «HN.H oNq.H .unwfioa >3 omuuoamu memo upomEH aoum omumawumm .msmsmu mnu mo smmunm mnu an 3mfi>mu poocD.I 3 \m. .muuoaxm I muuoasw + aowuoswoua uwumoEoo u coauQEDmcou undamaa<.l \N .oumon mmmooua doom: mam oumonumpwm huflmcmo aswoms monaocfi uoc mmom.l OOH mm om ow NHH um mq ON OH «H NH «H NH Hom.m oHc.m Nmm.m NON.m mmm.N qofi.m qmq.m NHH.m «om.N cen.fi oHN.H qu.H 2 mmmfi wmofi mmod ommfi mmmfi qmmfi mmmfi Nmmfi Hmmfi Onmfi mood womd A.o.ocoov H< manma 157 .ncmH .mmamsm vow uLwfluz mowafi .mUNEocoom HmfiuumnocH mo common .mouoEEoo mo ucmauumnwc .m.D ”on mummh omuomamm .msmcmo one we smousm .moumaaoo mo ucmEuumama .m.: ”on mumm> omuomfiom .msmcmo mzu mo smmusm .mouoEEou mo ucmEuumamo .m.D mnmmfi .maamnm mowmfi .uumnEmA "mmounom A.e.ucoov H< magma 158 qu.m Ne mmq Noo.m NomH oom.m me Noe mmo.m ocmH HN¢.m mm mwm NHm.N momH onm.m om mHm Now.N «omH HqH.m mN mmq NON.N mooH «ow.N NH HNm on.N NomH qu.N NH NwN qu.N HomH wNH.N NH moN 0mm.H oomH on,N NH NNm HNo.N mmmH on.H NH qu moo.H wmmH qu.H oH NmH 0mm.H NmmH owo.H mH HHN qu.H omoH moo.H NH NmH oNq.H mmmH .lcoaumasmaoo coauonooua \H ucmumaa< \Mmuuoaxm \mmuuomEH owummEoo mom» .wammn nocHIw\H .ummm mumscm mo mcoHHHHE ch mNmHImmmH .COHuaanmcoo ucoumaam mam .muuoaxm .mupanH .GOHuosooua oumonoum: .m.D .N< mHan 159 .ome .moHEocoom HmHuumaocH mo smmusm .moumEEoo mo unassummoo .m.D MNNmH .um>mmMoz "mmousom .muuomxm I muuanH + GOHuosooua oHummaoo n GOHuQESmcoo ucmumma<.l \N .AonHV uw>mmxoz Eoum dump mCOu ou ummm mumnvm COHHHHB Honccm mSu wcwm: wmumEHumm.l \H mmo.w Now.m moo.N NNq.c mHm.m mqq.o mNo.N woo.o omo.m oNN.< ONw.q ooN.q oHH mm ooH owH NmH mNH omH HoH «w NN No Nm Hoo.H mNH.H coo Hoe HNN mNN @OH.H woo.H «mo mow ooN mNo mwo.N on.N wom.o mNH.o HON.m mew.m oqo.o HNo.m oNH.m oqm.q NmH.¢ moo.m mNmH wNmH NNmH onH mNoH quH mNmH NNmH HNaH onH mcmH womH H.e.ucoov N< «Hana 160 mmN.m «q we ch.m NmmH mmo.m we we mNo.m oomH NNm.m Hm Ho Nom.m momH mNN.m Hm Ne NoN.m womH ch.m Hm om amo.m momH Nww.N me oe www.N NomH wa.N Ne «N omw.N HomH mew.N om Hm wow.N OQNH NHH.m Nm oq «HH.m mmoH HNw.N mm mN «ww.N wmmH qu.N em oH mNo.N NmmH «em.N om om mNm.N ommH mHm.N em «N mqo.N mmmH lboHuaEDmcoo GOHuonooua \N ucmumna< \Mmuuoaxm \mmuuanH oaummaoc umow .Amfimmn LUGHIN\H .ummm mumscm mo mGOHHHHE cwv mNmHImmmH .GOHuQEDmsoo ucmuwaam vow .muuomxm .muuanH .GOHuosooua oumon cowudemcw .m.D .m¢ mHan 161 wcH.m MNm.m Hom.m HNm.m HHw.N qNN.m NNm.m qu.m ocw.m omN.m omo.m mmm.m .omma .muancoom HmfiuumsecH to ammuam .muumeaoo mo unassumamn .m.a “mums .um>mm¥oz "mmuusom .muuoaxo I muuoaafi + GOHuosooua UHuonoo n GOHuaasmcoo ucmumaa<.l .AoNoHV uo>moxoz Boom owumu mcou ou uoom mumsvm coaHHHE Hmsacm mnu wcwm: woumafiumm me me ow 0N 0N HOH wN No we om we Nm me mm on oq Hm me ow om mm Nm Hm HHH on.m NNm.m Hmm.m Noo.m www.N NwN.m «Hm.m mHm.m mmw.m qu.m MNo.m oN¢.m \N \w sham mass Hams sass mass «Nos mama mean Haas oNaH mesa wcmH H.e.uaoov me magma 162 NmH.m m qu.H ch.H NomH HNm.m w qu.H oNo.N oomH omo.m o Nqo.H mco.N momH aqo.N N oq0.H NH¢.H qoaH NHc.N H mmm mwo.H momH qoq.m N How on.H NomH wNo.N m NNN mom.H HomH «Hw.H N mHN NoH.H oooH HmN.N m wmm oqm.H mmmH NoN.H N moo HmH.H wmmH mNN.H H Nmm NNH.H NmmH qu.H H wmq qu.H ommH mmN.H N Nee mmm.H mmmH ItOHuaeamcoo SOHuosooua \H ucmumaa< muuoaxm muuanH oHumoEon snow .AmHmmH LocHIm\m .ommm mumswm mo chHHHHE GHV mNmHImmoH .:0Huaenmcoo ocmumaam can .wuuoaxm .muuoaaH .aoHuosooua oooS%Ha ooozoumn .m.: .q¢ oHan 163 .ome .moHEocoom HmauumaoaH mo smousm .mouoaaoo mo unmauumaon .m.: ”va mummz wouoonm .msmcmo msu mo :mmusm .oouosaoo mo unmauumama .w.D “Nov mumm> wouoonw .msmcmu mnu mo smmusm .moumeaoo mo unmauummma .w.D mwNwH .mow>umw ummuom .muduHsoHuw< mo unmeuumamn .w.D "mmounow NwH.w qmw.m NNm.w owm.m Nom.N NNm.N mmN.q mmH.m omq.q qu.m owm.w Nww.w oq om ON mN we we oq wN mH wm wH qH wwo.N qu.N qu.N wmm.N me.H qu.H NNm.N me.m qu.N Nqo.N NoH.N www.H mHH.H wNH.H NwH.H wwo.H Nmo.H Hoq.H Now.H owo.N «N¢.H me.H www.H moo.N .muuomxo I muuoaaH + cowuonwoua owummEow u COHuQSsmcoo ucmumaa< I. \H meH wNaH NNmH wNmH mNmH «NoH NNmH NNoH HNmH onH mme wwwH H.e.ucouv q< mHnma 164 me.NH mw m oew.NH Nme «cw.NH we m mew.NH wme Noq.NH Om m qu.NH mme qu.HH wN m mmq.HH «wwH Nwm.cH wH oH mNm.oH mon HHm.o NH mH mHm.¢ Nme mae.w «H mH wmc.w meH NmN.N mH HH mmN.N ome qu.N NN I wwN.N mmmH mine 2 \N :36 was mm: 2 \M 203 $2 qu.m mH I qu.m wmoH ENK m \N. «38 HS .IcOHuaasmcoo COHuoswoua \H ucmumaa< mononxm mouoan ofiumoaon poo» .AmHmmn nocHIw\m .umow mumswm mo mcoHHHHE CHV mNmHImmmH .COHuQEDmcoo acoumaam wow .mupoaxo .muuoaaH .cOHuoswoua voozuHm wooauwom .w.D .m< oHan 165 .ome .moHEocoom HmwuumnwaH mo ammusm .moumfiaoo mo ucmEuuman .w.D “Nov whom» omuomHmm .mSmamo mnu mo nmmusm .moumEEoo mo ucmauumama .w.D “Nov wummh wmuomHmm .mnmcmo map mo :mmpam .moumeaoo mo ucma qumaom .m.D “any mummh wouomHom .mamcmu msu mo amousm .moumfifiou mo unmauumamo .w.D “NNmH .mmHmnm "mmousow .uomm oumnvm ooo.oom cmnu mmmH.l .muuoaxm I muuanH + SOHuoswoua ofiummEow u GOHuaEnmcoo usmummaumw umouom .ousuHsopr< mo unmauummmm .m.D mome .uuermH mnNmmH .msoamco:< umNmH .comuoe:< "mmounow .mumHHoe Neon mo mcoaaafln cww NN NN w.¢mm mNmH He NN m.NNm wNmH mm Hw m.mNm NNmH Nm mN w.Hmw eNmH Nm NN N.mmw mNmH mm NN o.qu quH mm NN N.qmw NNmH Nm oN w.How NNwH we qe o.oeN HNoH oq «e e.HqN ONmH em Nm m.NHN memH mm em N.mae wemH H.e.u:ooe os< mHamH 178 H.mw m.me o.Nm m.mMN mNmH w.em o.mN m.we m.mON wNmH e.H¢ m.mN N.¢m N.qu NNmH o.ww w.HN w.qm o.me eNmH w.Nw w.Ne q.eq m.qNH mNmH e.mw N.mN o.mm H.oeH quH m.om H.wN m.we N.emH mNaH o.mm w.mw w.Ne H.mHH NNoH ¢.ooH N.ww w.Ne o.qHH HNmH H.ocH e.Nm «.me q.oHH oNoH N.NoH N.ww m.moH m.eoH memH m.ooH m.mm m.wm w.NoH wemH o.ooH o.ooH o.ooH o.ooH NemH \Meumon cowumHsmcH \meumoneumm \MeumoanoHuumm moHuHeoEEou HH< ummw .AooH u NemHv meHINemH .eumon coaumHnmCN ecm .eumoeeum: .eumoanoHuumn .mmHuHeoEEou HHm pow mmxmecH oofium nooseoum .HH< mHan 179 .mnmm% emuumHmm .moflumwumuw poemH wo smmusm .uoemH mo ucmEuumamn .w.: HNNwH .maHmsm “mmousow cw emm: EfiuwumwOH .mdoammouwou meow .mmHuHeoEEoo HHm new xmecH ooHua nooseoua he emumeme xmecw mowua umoaeoum.l \H A.e.uaooe Hs< mHamH 180 e.oNH o.Ne N.mHH w.qu m.mmH wNaH o.oNH w.me o.eHH e.qu w.NmH NNoH o.moH m.ee o.HoH w.NmH m.mmH eNmH m.em m.we q.ow o.eHH N.eHH mNmH w.wm m.Hw o.qw e.NHH N.eHH «NoH e.mHH N.ww H.moH o.qu o.qu MNaH H.HHH e.Nw w.NoH «.mNH H.0mH NNmH w.NoH N.ww w.ow w.oHH q.HHH HNwH N.ww w.Nm H.Hw w.NoH w.NoH onH N.qHH e.Nm o.coH w.NoH N.omH memH m.mHH o.wm I o.eNH o.eNH wemH o.ooH c.ooH I o.ooH o.ooH NemH .I-Immu:uHumesm .IeooBNHa \mmwmoanoHuuma .leoo3%Ha leooz%Ha \mooauHOm .IeooBNHQ emustmz \wooseumm ocmm sumsunow cumumoz \mooauwow ummw .AooH u NeoHv mNaHINemH . ImquuHumesm eumonoHoHuuma wounwwms ecm .eooBNHa eoozeumz . leoo3NHa mcHa cumzusom .e003NHa eoobwwom auoumms .eoosta eooaumom pow mmxmesfi mowua umoneoum .NH< mHemH I 181 .mummx emuuoHom .moHumHumum uoemH mo smmusm .uoemH mo unmauummon .m.: HNNmH .maHmsm "mmuunom .moumsm uoxuma Ne wouanms mmxoecfi moHua umoseoua eooBNHa eooseums ecm eoo3mHa eooBuwom .eumon :OHumHSmcH .eumonwumm.l \m .mcoflmmouwmu maom CH emm: ELuHumwOH .mmwuweoafiou HHm you xmecH moflua Hmoseoua Ne eoumHeme xmecw moaum umo:eoum.l \N 2: u $ma q.on w.HN o.qm m.mmH H.NmH onH A.e.ucooVHH< mHan 182 m.oNH mq.we N.wm we.N H.qu m.mo wNmH N.eNH No.mm w.Nm Ne.N e.mmH N.HoH NNmH m.mHH qe.om m.em we.N H.qu o.eoH eNmH e.mm HN.¢w e.Nw we.N H.qu «.moH mNoH N.ww oN.om e.qm qe.N H.omH w.mw «NNH m.eoH mw.mN e.wm eN.N N.ww ¢.we wNwH w.NoH we.wm N.Nm wN.N e.mm m.qN NNmH o.wm mN.wq e.Nw eN.N N.coH o.wN HNmH w.Nw we.Ne N.qw we.N m.m¢ H.Nw onH e.¢w oN.Hm w.mm Ne.N m.qm o.qw meaH m.HN I m.em Hm.N m.em N.ww weoH m.me I N.ww Nm.N o.ooH o.ooH NewH Iooaua .INuH>HuoCeoua mmmaewuw .IxmeCH moHCa .IxoeCH moHCC \e HHHE on3mm mCHC ICOHuoCeoua mu meoua emumHmC mmmmou oHummHC eooshHa Cumnusow \q umnECH \mmmme eCm mHmCm eCm mm>Hmm£e< Cmmw .onHINemH .mCOHumCUo NHCan CH lemma mama dH< mHan I 183 .mpmoh emuomHom .mUHumHumum CoemH mo Cmmusm .CoemH mo quEquamm .w.D mome .moHEOCoom HmHuumCeCH mo Cmmunm .monEEoo mo quEuumamn .m.D mmNoH .moH>umw ammuom .oCCuHCoHCw< mo quEquCmm .m.D HNNmH .mCHmcm qu< eCm e< moHan “mmoudom .mquHQ eoo3hHa mCHa Cumnunom mo ConECC mcu Np emeH>He COHuoseoum eoo3zHa mCHC Chonuaow\m .ooH n memH .mmHuHeanoo HHm Cow xmeCH mUHua noose IOCC msu %e emumHeme mummuom HmCOHumz Eoum mmoHHC mwmmECuw CmnBHuamm mCHC Cumcusow .mumHHoe CH\N .ummH eumoe Ho mCOHHHHn CH\w .ooH u NeoH .moHuHeoaaoo HHm COM xoeCH ooHCC Cooseoum men we emomHeme mpmxuos mooseoua e003 eCm CerCH Cow meHCCmo NHCCom.I \m .AooH u NemHv mmHuHeoEEOU HHm new xmeCH moHCC Cmoaeoua msu CuHB emumHmon\M .mume NuHquCv eCm moHCQ Cmnu Conu0\fl m.eoH em.Hm o.em wm.N m.MNH N.ooH wNmH A.e.ucoov mC< mHamH 184 ommH mNe mHmH NNwH wqu Hmm NHNH eNmH HNHH qu Nmm mNmH mmmH NHq ewm «NmH meN omw mmHH mNmH aNmN owm mmmH NNmH mwON eqw mwNH HNmH meqH wNm How onH oomH mHe Nww memH memH New mwm weoH NNmH eoe eHm NewH emHH Nmm wwN eemH onH wmq NmoH memH HemH mom emoH qeaH Imuumum ImuHCC +m ImuHCC NwH wCHAmmBe HmuOH .mmwmum wCHmCom .mmwmum wCHmCom Com» .AmuHC: mo meCmmCocu CHV mNmHIqemH .mCOHuvao eCmEme CH com: mume wCHmCO£ .m.D .qH4 mHan 185 .oHHeCC eCm mum>HCm.l \m .zHCo oum>HCm\M .muHCC +m .muumum wCHmCo: mCCHE mupmum wCHHHm3e Hmu08\fl quH mmq emNH mNmH wNON mNm wqu wNmH A.e.uCoov «HC mHan 186 e.me q.qqm.N q.e¢e.mN m.emH.¢ N.NNm.mm meH N.mNH m.eNe.w N.eme.ON H.mom.o m.eqH.me wNwH m.emH N.wwH.w m.wmm.wH m.omo.o ¢.meo.Nq NNmH w.NqH m.Hem.o m.HmH.wH N.mmm.w H.qmw.Nm eNmH e.me e.wwH.HH N.owo.mH w.mmw.N m.Hew.qN meH m.eNH w.qu.HH H.oew.mN q.oqm.e N.wNo.Nm «NwH N.on H.wH¢.HH m.eNm.mN N.Nme.e N.wNo.eq wNmH H.ooH e.wwq.HH o.NHo.qN e.NHq.N N.qmw.qq NNmH w.qm N.NNo.NH o.NHN.wN e.wwH.N m.mwm.em HNmH e.ww N.wNo.NH N.NNH.QN H.emo.N o.mmm.NN oNoH N.Nw N.mNm.mH e.owm.mN e.NHH.N e.Nem.Hm memH H.eN w.NHN.mH e.wew.mN e.oem.e m.eNm.Hm wewH «.NN w.NwN.mH N.qu.qN m.wew.N «.NNN.eN NeoH xmeCH ouHmoaaoo meHeHHCe meHeHHCe mCOHumumuHm ICOHuUCCumCoo muumaaoo mo oHHACC HmHqueHmmCCOC eCm mCOHuHeem \NHmHqueHmmH quEuumCoo mo mCHm> mum>HCQ mo oCHm> HmHqueHmmu mo mCHm> mo mCHm> Cow» .\flAmCmHHoe NNmH mo mCOHHHHE CHV mNmHINemH .mmumuw emuHCD mcu CH COHuuCCumCoo mo mCHm> Co mama .mH< mHan 187 .va mummm ewuumHmm .mCmCmu ozu mo Cmmusw .muumfifioo wo quEquCmn .w.D "mmuunom .momHa CH use mCHm> .NHCO mom:on.3mZ\M .ooH we eoHHaHuHCE eCm xmeCH muHmanoo ooCmEEoo mo quaquamo he emeH>He mCmHHoe quHuCu.l \H A.e.uaooe mCe mHan 188 NHN N.ww w.Nm o.ooH eemH NHN «.mN N.ww o.HoH memH HoH N.ww o.Hw H.wm qemH HmH N.em w.mN m.mw memH wHH m.mN w.HN e.ow NemH om m.mN e.we e.NN HemH «OH N.wH w.ee e.oN oemH HNH N.mH N.ee w.HN mmaH HoH o.eH e.Nw e.NN wmmH HHH o.mH N.He o.me NmmH oHH o.m w.ee w.ee emmH cHH o.w N.wm N.ee mmmH IwCHCCuommCCmE ImquECHLm Imuumum wCHmCOL HHm IwCHCCuomeCmE \H mmusume \m oEo: \wo owmuCooCmC m mm \H HHm eCm muCuHCHCm mHHeoz muHCC NHHammHuHCZ wo xoeCH mo xoeCH Com» .mNmHImmmH .mCOHuvam eCmEme e003hHa mCHC Cumsusom eCm eumoemHoHuumC CH om: How common eCm com: mume Cmsuo .eHC mHemH 189 NNN eNN NNN eqN mHN mNm Nem eNm Noe Hoe qu me OQN .ooH >3 emHHCHuHCE eCm muumum wCHHHmBe HmuOu he eoeH>He muHCC +m .muumum wCHmCom N.wN o.mN m.wN e.oq N.ww w.ee w.em N.ww e.me w.eeH e.me m.owH w.eHH «.mNH w.0NH a.wHH N.on q.eoH o.HHH q.eoH o.ooH w.HeH w.mmH o.meH N.NmH N.wHH e.NmH N.qu e.omH o.eHH H.on «.NoH «.moH o.ooH muHCC mo meCmmConu CH .ooH n NemH \m \N D. meH wNmH NNmH eNmH mNaH «NaH mNmH NNoH HNmH onH weaH wemH NemH A.e.uCooV eH< mHan 190 .ome .moHEOCoom HmHuumCeCH wo meusw .monEEoo wo uCoEquaoC .m.D ”mummN emuoonm .mHmNHmC< UHSOCoom mo Cmmunw .moCmEEoo Ho quEuumCmo .w.D mmNmH .moH>me ummuom .mCCuHCoHCw¢ mo quEquCoQ .w.D H¢H< mHan "mmousow A.e.uc00e ea< mHamH 191 mm.m wNm.N Noq.¢H e.Nw wNmH Ne.N NaH.N NNw.wH e.ww NNmH Hm.m wHH.N eoo.NH H.Nw eNmH wo.m oeo.N eoN.mH m.qm mNmH mo.w er.H eow.mH N.H¢ quH wm.m Nme.H me.NH N.Nq mNmH Ne.N eme.H mew.NH H.qq NNmH HH.m woe.H wmm.eH N.Nm HNmH Nw.N meq.H oeH.eH e.NN onH Nw.N Nem.H www.mH H.om memH I mNm.H mwm.qH m.mN wemH I mwN.H oqw.NH m.mH NemH aw”... fimfim. gamma ....flmma eooBNHm ustma emouHHmm eoo3umow Com uCCuCO Com» .mNmH INemH .mCOHuvao NHQCCm eoo3zHC mCHn Cuocunom eCm eumoemHoHquC CH mm: Cow emummu mume Conuo .NH< oHemH 192 .ome .moHEOCoum HmHHumCeCH mo Cmmunw .muumfifioo wo quEquamQ .w.D moH< eCm m< .H< mmHemH "mmoHCom .ooH u memH .mMHuHeoEEoo HHm COM xmeCH onua mHmmmHoz3 mcu Ne emumHeme meHCCmm kHCCo: mumHuoB COHuoCeoum eooBhHC eCm CmmCm> eoo3umow .mCmHHoe CH\.m .mquo CH.I \m .mHmme fioCHIw\m .ummw dpmsum mo mCoHHHHa CH\M .mquHa mo CmeECC we emeH>He COHuoCeoua UHummaom\fl oq.m mNm.N wwm.wH w.mq mNmH H.e.ucoov Nae mHan 193 m.mmm quH N.me qu NNm oNoH e.omm wqu H.Nq me qoe wNoH m.on quH o.oq weH wwm NNmH N.mow NeHH o.om ew qeq eNoH w.mwN Now o.He oe eem mNoH o.HwN www w.Hq qu New quH H.NNN NmHH N.Nq qwm qu wNoH e.me oomH w.ee Hoe NHe NNoH o.wMN HmHH N.ee mom eNm HNoH o.oNN me w.ee NON NNm oNoH o.oNN HHw N.Nq HNN New meoH Ifiusom Imupmum Imuumum wCHmCo: \msu CH \wCHmso: uHCmNmCo .w.D Hmuou mo .Isusow onu CH lzuoow mEooCH uHCC oco quoCoa m mm nuCow moo CH m wmum wCHHHmse m£u\mH muumuw HmComCmm .w.D Hmuoe muumum wCHmCon uHCC mCO uHCC +m wCHm50£ uHCC mCo Com» .oNoHIoeoH .cusow mnu Ou onHuoCm mumo eCmEmQ .wH< oHemH 194 .mummh emuomHom .mHmNHmCm oHEOCoom wo CmmCCw .moumaaoo mo quEuumamO .m.D “Ame whom» emuomem .mCmCmO mnu mo Cmmusm .moumano mo quEqume .m.D "mmousom .OoH u NNoH .moCCuHOCmoxm COHuaECmCoo Hmcompoa COM Coumeoe ooHCC uHUHHaEH one :oHs emumemo .mCmHHoo NNmH mo mCOHHHHe CH \m .oos an emHHmHuHCE ecm muumum mCHmCOL uHCC mCo .m.D Hmuou we emeH>He Lunom mSu CH muumum mCHmCo: uHCC mCO\M .muHGD mo mficmmDOLu CH\..IH- A.e.u:00e ms< mHamH 195 eomN m o.Nw HHOHHoz\CmE:on mmoN o o.mm Mao» Bmz eHmm HH o.Nw HamHz «EN 0 men «2325:: quN em o.CoO Hnmm .uw\mHHoammCCHz mHuumom eCmemo\oumHquum Cow mmHme< mOH eCmHuuom xoom oHuuHH wusemumumm .uw\maEmH eCmHm>mHO 0HC0uC< me mCoEHuHmm\C0uwCH£mm3 anwCHECHw mHHH>Comxomo mHsoa .om A.e.ocoov mae mHan 197 .ONoH .CompmeC< "mousom eNON N.ww mmstmsHHz H.e.uaoue ma< mHemH 198 .mquEoH:m eooBNHC wNoH HmuoH 1m \C uumum wCHmCom emo.eHm Nem.mH mazes .um moN.mNN mom.HH xHoHCoz\Cmaum£on moN.oHN ooO.N xuow 3oz qu.on NwH.oH HamHz Cmm.maq NmH.mH «HeaHmemHHem HNm.Nmm eoq.N mCmmHCo 3oz omq.mem ooe.q mHnaEmz on.mem qu.N~ oewHoH\oHoCumm wmw.NmN ONm.Hm owmoHCO MNw.Hem mNN.NN euuoz .uC\meHmo ooH.MNm OHm.wm Coumso: mHN.Nem mwo.ms mwooamuumeo\mocmHue NNm.Noe qu.m wouOHCmso mHmmeI:w\m .uooe oCvam mo mCOHHHHE CH wNoH CH ACmo Hsmm .um\mHHommoCCHz mHuummw eCmemo\oomHoCmCm me mmHowC< mOH eCmHuCom xoom mHuuHH wusemumuom .um\mCEmH eCmHm>mHO OHCOuC< me mCoEHuHmw\C0uwCH£mm3 EmcwCHEHCm mHHH>Comxomo H.e.ua00v o~< mHan 200 . m mummx emuum mm .mCmCmo m u o Cmmunm .moCmEEoo o quEuumamo .m.D m .Cowum C< “mmousow V H n w w oNoH e .muumum wCHmCos Cow Nxouo m mm com: mum3 meHHHmSO uHCC q eCm m.N.H Com eosmmH muHECma wCHeHHCm.I \H A.e.ocoov ome mHan 201 Table A21. Projected high, medium, and low levels of U.S. disposable personal income, 1990 and 2000 (in billions of 1972 dol— lars). Year U.S. disposable personal income High Medium Low 1990 1540 1450 1360 2000 2110 1880 1690 Source: U.S. Department of Agriculture, Forest Service, 1979. REFERENCES CITED REFERENCES CITED Adames, Darius M. and Richard W. Haynes. 1980. The 1980 softwood timber assessment market model: structure, projections, and pol- icy simulations. Forest Science Monograph 22. Washington, D.C.: Society of American Foresters. Adams, Thomas. 1978. Production opportunities in the Pacific North- west. in Structural Flakeboard from Forest Residues, USDA Forest Service General Technical Report WO-5. Washington, D.C.: U.S. Government Printing Office. Akers, L. E. 1966. Particleboard and hardboard. Oxford, New York: Pergamon Press. Ambler, John. 1980. Personal communication regarding Current Industri- al Reports, Series MA-24H, U.S. Department of Commerce, Bureau of the Census. Anderson, Robert G. 1978. Softwood plywood use in non-residential construction. American Plywood Association Market Research Report R41. Tacoma, Washington: American Plywood Association. . 1979. Regional production and distribution pat- terns of the softwood plywood industry. American Plywood Associ- ation Market Research Report E27. Tacoma, Washington: American Plywood Association. . 1980. Plywood end-use marketing profiles 1979- 1981. American Plywood Association Market Research Report E28. Tacoma, Washington: American Plywood Association. Anonymous. 1957a. Fiber-clad building board plant. The Lumberman 84(11): 58-62, 71-72. Anonymous. 1957b. Hardboard, particleboard, flakeboard plant direc- tory. The Lumberman 84(5): 159-165. Anonymous. 1957c. Two men operate entire production line in plant. The Lumberman 84(5): 57-59. Anonymous. 1962. Wizewood makes wafers from green poplar to form a new board. The Lumberman and Wood Industries 89(6): 40-42. 202 203 Anonymous. 1966a. Los Angeles story spotlights problem of noncon- formance. Forest Industries 93(1): 58-59. Anonymous. 1966b. New plywood standard simplifies specification. Forest Products Journal 16(12): 14. Anonymous. 1971a. Board line uses rotary press. Forest Industries 98(7): 47. Anonymous. 1971b. Associations backstop makers with promotion, tech- nical help. Forest Industries 98(8): 40-41. Anonymous. 1975. Oriented wood strand panels open new era in utili- zation. Forest Industries 102(3): 46-48. Anonymous. 1978. Softwood plywood, the south—-1967—1977. Forest Industries 105(3): 30. Anonymous. 1979a. PET, HDPE to lead plastics parage. Purchasing World 23(12): 46-47. Anonymous. 1979b. Upturn in production coming for composite, OSB panels. Forest Industries 106(9): 11-12. Anonymous. 1980a. Capital spending jumps in solid wood industry. Forest Industries 107(1): 30-31. Anonymous. 1980b. MDF plants, location, etc. (advertisement). Forest Industries 107(4): 28. Anonymous. 1980c. Particleboard: the versatile panel product of the '80's. Wood and Wood Products 85(2): 12-13. Anonymous. 1980d. Unrest at Georgia-Pacific. Business Week 2664: 147-148. Anonymous. 1980e. Using particleboard and MDF II. Wood and Wood Products 85(2): 8-95 (Part 2). Anonymous. 1981a. Directory of panel plants--U.S. Forest Industries 108(4): 63-67. Anonymous. 1981b. South's 'third forest' falls short of goals. For- est Industries 108(1): 42. Anonymous. l981c. Study, APA disagree on plywood industry's fate. Forest Industries 108(1): W6. Arola, R. A. 1978. Fiber for structural flakeboard--mechanized thinning and topwood recovery. iE_Structural Flakeboard from Forest Residues, USDA Forest Service General Technical Report WO-S. Washington, D.C.: U.S. Government Printing Office. 204 Arrow, Kenneth J. 1962. The economic implications of learning by doing. The Review of Economic Studies 29(3): 155-173. Baldwin, Richard F. 1977. Southern pine plywood--from 0 in '64 to 33% of U.S. output in '76. Forest Industries 104(3): 40-43. Ball, G. W., R. P. Redman, and A. D. Adams. 1979. Particleboard binder promises performance without emissions. Forest Indus- tries 106(4): 76-79. Bentley, William R. 1970. Technological change in the forest indus- tries--a problem analysis. The University of Wisconsin Forestry Research Notes, No. 151. Blackman, Ted. 1978. Electron beam system cures coated particleboard panels. Forest Industries 105(3): 70-71. 1980. Commercial production starts on new composite Elcoboard. Forest Industries 107(4): 54-55. Brown, Daniel H. 1979. How building codes affect wood markets. For- est Products Journal 29(1): 14—15. Bryan, Richard W. 1978. Continuous rotary process makes thin particle- board for wall paneling. Forest Industries 105(3): 52. . 1981. Composite pine panel plant is the last word in versatility. Forest Industries 108(4): 34-37. Buongiorno, Joseph and R.A. Oliveira. 1977. Growth of the particle- board share of production of wood-based panels in industrialized countries. Canadian Journal of Forestry Research 7(2): 383- 391 O Buzzell, Robert D. 1966. Competitive behavior and product life cycles. ig_New Ideas for Successful Marketing, edited by John S. Wright and Jack L. Goldstucker. Chicago, Illinois: American Marketing Association. Carney, Michael. 1977. Softwood plywood used in new residential con- struction--1976. American Plywood Association Market Research Report R38. Tacoma, Washington: American Plywood Association. Coppens, H. A. E., M. A. Santana, and F. J. Pastorze. 1980. Tannin formaldehyde adhesive for exterior grade plywood and particleboard manufacture. Forest Products Journal 30(4): 38-42. Cour, Robert M. 1955. The plywood age, a history of the fir plywood industry's first fifty years. Portland, Oregon: The Douglas Fir Plywood Association. Cox, Lionel A. 1974. Transfer of science and technology in success- ful innovation. Forest Products Journal 24(9): 44-48. 205 DeLess, Bruce C. 1975. Automated layup line saves wood, cuts costs. ig_Modern Plywood Techniques, Volume 2: Proceedings of the second Plywood Clinic, New Orleans, Louisiana, November 1974. San Fran- cisco, California: Miller Freeman Publications, Inc. Dickerhoof, H. Edward. 1976. Particleboard production, markets, and raw materials in the United States. Forest Products Journal 26(10): 16-20. 1977. Particleboard production, markets, and raw material in the South. USDA Forest Service Research Paper SE-157. Asheville, N.C.: Southeastern Forest Experiment Station. 1978. Use of wood in mobile homes is increas- ing. U.S. Department of Agriculture Resource Bulletin FPL 4. Madison, Wisconsin: Forest Products Laboratory. Dickerhoof, H. Edward and Thomas C. Marcin. 1978. Factors influenc- ing market potential for structural flakeboard. ig_Structural Flakeboard from Forest Residues, USDA Forest Service General Technical Report WO—5. Washington, D.C.: U.S. Government Printing Office. Dickerhoof, H. Edward and David B. McKeever. 1979. Particleboard, medium-density fiberboard, and mende process board plants in the United States--capacity, production, and raw material trends, 1956-1976. USDA Forest Service Resource Bulletin FPL-6. Madison, Wisconsin: Forest Products Laboratory. Durbin, J. and G.S. Watson. 1951. Testing for serial correlation in least squares regression, II. Biometrika 38: 159-178. Egan, Mary Lou. 1980. Wood products companies near energy indepen- dence. Forest Industries 107(9): 79. Elmendorf, Armin. 1956. New pressed board developments are reviewed. The Lumberman 83(3): 178. Farris, Mary R. 1978. The veneer and plywood industry: above aver- age productivity gains. Monthly Labor Review 101(9): 26-30. Fassnacht, Donald L. 1064. The story behind the new southern pine plywood industry. Forest Products Journal 14(1): 21-22. 1965. Southern pine plywood. Southern Lumber- man 210(2616): 15—16. Food and Agriculture Organization of the United Nations. 1976. Pro- ceedings of the world consultation on wood-based panels. Brus- sels: Miller Freeman Publications. Frashour, Ronald. 1961. An approach to particleboard standards. The Lumberman 88(7): 64-66. 206 Geimer, Robert L. and Eddie W. Price. 1978. Construction variables considered in the fabrication of a structural flakeboard. In Structural Flakeboard from Forest Residues, USDA Forest Service General Technical Report WO-S. Washington, D.C.: U.S. Govern- ment Printing Office. Greenwald, Douglas and Associates. 1973. The McGraw-Hill dictionary of modern economics. Second edition. New York, New York: McGraw-Hill Book Company. Griliches, Zvi. 1957. Hybrid corn: an exploration in the economics of technological change. Economietrica 25(4): 501-522. Hair, Dwight. 1967. Use of regression equations for projecting trends in demand for paper and board. USDA Forest Resource Report No. 18. Washington, D.C.: U.S. Government Printing Office. 1980. Timber situation in the United States—-1952-2030. Journal of Forestry 78(11): 683-686. Hall, Bronwyn and Robert Hall. 1978. Time series processor (TSP), version 2.8, user's manual. Available at Michigan State Univer- sity Computer Center. Haynes, Richard W. and Darius M. Adams. 1979. Possible changes in regional forest product output and consumption during the next 50 years. Forest Products Journal 29(10): 75-80. Hirshleifer, Jack. 1976. Price theory and applications. Englewood Cliffs, New Jersey: Prentice-Hall, Inc. Holley, D. L. 1969. Plywood's ability to compete for southern pine stumpage. Southern Lumberman 219(2728): 177—179. Irland, Lloyd c. 1972. Labor trends in southern forest industries, 1950—1969. USDA Forest Service Research Paper SO-81. New Orleans, La.: Southern Forest Experiment Station. Johnston, John. 1972. Econometric methods. Second edition. New York, New York: McGraw—Hill Book Company. Jorgensen, Richard N. 1978. Requirements, codes, and Forest Service goals, 13 Structural Flakeboard from Forest Residures, USDA For- est Service General Technical Report WO-5. Washington, D.C.: U.S. Government Printing Office. Kmenta, Jan. 1971. Elements of econometics. New York, New York: MacMillan Publishing Co., Inc. Koch, Peter. 1972. Utilization of the southern pines. USDA Agricul- tural Handbook No. 420. Washington, D.C.: U.S. Government Print- ing Office. 207 Koch, Peter. 1978. Two methods of acquiring residual wood for south- ern flakeboard plants--the shaping-lathe headrig and the mobile chipper. in Structural Flakeboard from Forest Residues, USDA Forest Service General Technical Report WO-5. Washington, D.C.: U. 8. Government Printing Office. Kotler, Phillip. 1976. Marketing management. Third edition. Engle— wood Cliffs, N. J.: Prentice-Hall, Inc. Lambert, Herb. 1980. Late '79 downturn in housing had modest effect on panels. Forest Industries 107(4): 30-33. 1981. Understatement: panels did not score high in 1980. Forest Industries 108(4): 30—33. Lancaster, Kelvin. 1971. Consumer demand: a new approach. New York: Columbia University Press. Leefers, Larry A. 1979. Forces affecting technological change, inno- vation, and diffusion: a literature survey. Survey prepared for U. S. Forest Products Laboratory under Extramural Research Agreement 12-146. Lehmann, W. F. and H. E. Walhgren. 1978. Status and prospects of residue utilization in board manufacture in the United States and Canada. Forest Products Journal 28(7): 24-29. Lekvall, Per and Clas Wahlbin. 1973. A study of some assumptions underlying innovation diffusion functions. Swedish Journal of Economics 75(4): 362-377. Lewis, Wayne C. 1965. Wood-residue utilization. Forest Products Journal 15(8): 303-307. Maddala, G. S. 1977. Econometrics. New York: McGraw-Hill, Inc. Maloney, Thomas M. 1977. Modern particleboard and dry-process fiber- board and manufacturing. San Francisco, California: Miller Freeman Publications, Inc. Mansfield, Edwin. 1961. Technical change and the rate of imitation. Econometrica 29(4): 741-766. . 1963. The speed of response of firms to new tech- niques. Quarterly Journal of Economics 77(2): 290-311. Manthy, Robert S. 1974. Opportunities for technological advance in the timber industry: an assessment. Lansing, Michigan: Green- tree Associates. Unnumbered monolith. Marra, G.G. 1969. A vision of the future for particleboard. in Proceedings of the Washington State University Particleboard Symposium, No. 3. Pullman, Washington: Washington State Univer- sity. 208 Mast, Fred R. and David A. Pease. 1969. Suddenly they're everywhere—- automatic panel layup systems abound as wraps are taken off. For- est Industries 96(1): 50-52. McCarthy, E. Jerome. 1975. Basic marketing: a managerial approach. Fifth edition. Homewood, Illinois: Richard D. Irwin, Inc. McKeever, David B. 1979. Hardboard and insulation board plants in the United States-—capacity, production, and raw material require- ments, 1955-1978. USDA Forest Service Resource Bulletin FPL- 7. Madison, Wisconsin: Forest Products Laboratory. McKillop, W. L. M. 1967. Supply and demand for forest products-- an econometric study. Hilgardia 38: 1-132. McKillop, William, Thomas W. Stuart and Peter J. Geissler. 1980. Competition between wood products and substitute structural pro— ducts: an econometric analysis. Forest Science 26(1): 134-148. McSwain, George A. 1977. Technical developments in the wood-based panel products industry. Paper presented by George A. McSwain, Assistant Director, Forest Products Laboratory, Forest Service, U. S. Department of Agriculture to the Food and Agriculture Organ- ization of the United Nation's Committee on Wood-based Panel Pro— ducts, Fifth Session in Rome, Italy in November 1977. Mills, Thomas J. and R. S. Manthy. 1974. An econometric analysis of market factors determining supply and demand for softwood lum- ber. Michigan State University Natural Resources Research Report 238. Moslemi, A. A. 1974. Particleboard. Volume 1. Carbondale, Illinois: Southern Illinois University Press. Mottet, Arthur. 1962. Particleboard now, tomorrow. Forest Industries 89(7): 94-96. National Research Council. 1978. Technological innovation and forces for change in the mineral industry. Washington, D. C.: National Academy of Sciences. Neter, John and William Wasserman. 1974. Applied linear statistical models. Homewood, Illinois: Richard D. Irwin, Inc. Oliver, F. R. 1964. Methods of estimating the logistic growth func- tion. Applied Statistics 12(2): 57-66. Pease, David. 1977. Computers raise output of particleboard line. Forest Industries 104(7): 62-63. . 1980. Structural panels in limelight as capacity growth approaches. Forest Industries 107(7): 52—53. 209 Pease, David. 1981. Waferboard plant finds market growing despite housing drop. Forest Industries 108(4): 54-57. Phelps, Robert B. 1977. The demand and price situation for forest products, 1976-77. USDA Forest Service Miscellaneous Publication No. 1357. Washington, D.C.: U. S. Government Printing Office. Price, Eddie W. and William F. Lehmann. 1978. Flaking alternatives ig_Structural Flakeboard from Forest Residues, USDA Forest Ser- vice General Technical Report WO-5. Washington, D.C.: U.S. Gov- ernment Printing Office. Quandt, Richard E. 1964. Estimating the effectiveness of advertising: some pitfalls in econometric analysis. Journal of Marketing Re— search I(2): 51-60. Reid, William H. 1958. The particleboard industry. The Lumberman 85(7): 38-41. Risbrudt, Christopher D. 1979. Past and future technological change in the U. S. forest industries. Unpublished Ph.D. dissertation, Michigan State University. . 1980. Personal communication regarding par- ticleboard furnish, U. S. Department of Agriculture, Forest Pro— ducts Laboratory. Rogers, Everett and Floyd Shoemaker. 1971. Communication of innova- tions. New York: The Free Press. Schmookler, Jacob. 1966. Invention and economic growth. Cambridge, Mass.: Harvard University Press. Stern, Louis. 1964. Market share determination: a low cost approach. Journal of Marketing Research 1(3): 40-45. Stone, Robert N. and Thomas C. Marcin. 1978. Surveys for forest re- source demand analysis. Paper presented at the Society of Ameri- can Foresters' National Workshop on Integrated Inventories of Renewable Natural Resources in Tucson, Arizona, January 9, 1978. TenWolde, Anton. 1979. Personal communication regarding purchased energy in the wood products industries, University of Wisconsin. U.S. Department of Agriculture, Forest Products Laboratory. 1974. Wood handbook: wood as an engineering material. USDA Agricul- ture Handbook No. 72, revised. Washington, D.C.: U.S. Printing Office. U.S. Department of Agriculture, Forest Service. 1974. The outlook for timber in the United States. USDA Forest Resource Report No. 20. Washington, D.C.: U.S. Government Printing Office. 210 . Department of Agriculture, Forest Service. 1977. The nation's renewable resources--an assessment, 1975. USDA Forest Resource Report No. 21. Washington, D.C.: U.S. Government Printing Office. . Department of Agriculture, Forest Service. 1978. Structural flakeboard from forest residues. USDA Forest Service General Technical Report WO-5. Washington, D.C.: U.S. Government Print- ing Office. Department of Agriculture, Forest Service. 1979. An analysis of the timber situation in the United States, 1952—2030. Review draft. Washington, D.C.: U.S. Government Printing Office. Department of Agriculture, Forest Service. 1980. An assessment of the forest and range land situation in the United States. USDA Forest Service, Fs-345. Washington, D.C.: U.S. Government Printing Office. . Department of Commerce, Bureau of Industrial Economics. 1980. Statistical series--tables. Forest Products Review 36(1): 27-50. Department of Commerce, Bureau of the Census. Selected years (a). Particleboard. Current Industrial Reports, Series MA-24L. Washington, D.C.: U.S. Government Printing Office. Department of Commerce, Bureau of the Census. Selected years (b). Softwood plywood. Current Industrial Reports, U.S. Govern- ment Printing Office. . Department of Commerce, Bureau of the Census. Selected years (c). U.S. exports--commodity by country. FT 410. Washington, D.C.: U.S. Government Printing Office. . Department of Commerce, Bureau of the Census. Selected years (d). U.S. imports--commodity by country. FT 135. Washington, D.C.: U.S. Government Printing Office. . Department of Commerce, Bureau of the Census. Selected years (e). Housing starts. Construction Reports C20. Washington, D.C.: U.S. Government Printing Office. . Department of Commerce, Bureau of the Census. Selected years (f). Value of new construction. Construction Reports C30. Wash- ington, D.C.: U.S. Government Printing Office. Department of Commerce, Bureau of Economic Analysis. Selected years. Survey of current business. Washington, D.C.: U.S. Government Printing Office. Department of Commerce, National Bureau of Standards. 1966a. Mat-formed wood particleboard. NBS Commercial Standard 236—66. Washington, D.C.: U.S. Government Printing Office. 211 U.S. Department of Commerce, National Bureau of Standards. 1966b. Softwood Plywood, Construction and Industrial. Product Standard PS 1-66. Washington, D.C.: U.S. Government Printing Office. U.S. Department of Commerce, National Bureau of Standards. 1974. Construction and industrial plywood. Voluntary Product Standard PSI-74. Washington, D.C.: U.S. Government Printing Office. U.S. Department of Commerce, National Bureau of Standards. 1977. Tabulation of voluntary standards and certification programs for consumer products. NBS Technical Note 948. Washington, D.C.: U.S. Government Printing Office. U.S. Department of Commerce, Census of Manufactures, 1972. 1975a. Industry series: millwork, plywood, and structural wood members, n.e.c., MC72(2)-24B. Washington, D.C.: U.S. Government Printing Office. U.S. Department of Commerce, Census of Manufactures, 1972. 1975b. Industry series: wooden containers and miscellaneous wood pro- ducts, MC72(2)-24C. Washington, D.C.: U.S. Government Printing Office. U.S. Department of Commerce, Census of Manufactures, 1977. 1980a. Industry series: millwork, plywood, and structural wood members, n.e.c., MC77-I-24B. Washington, D.C.: U.S. Government Printing Office. U.S. Department of Commerce, Census of Manufactures, 1977. 1980b. Industry series: wooden containers and miscellaneous wood pro- ducts, MC77—I-24C. Washington, D.C.: U.S. Government Printing Office. U.S. Department of Labor, Bureau of Labor Statistics. Selected years. Producer prices and price indexes. Washington, D.C.: U.S. Govern- ment Printing Office. Ventre, Francis T. 1979. Innovation in residential construction. Technology Review 82(2): 50—59. Vogel, B. M. 1976. Standards referenced in selected building codes. U.S. Department of Commerce, National Bureau of Standards, NBSIR- 1140. Washington, D.C.: U.S. Government Printing Office. Warner, K. E. 1974. The need for some innovative concepts of innova— tion: an examination of research on the diffusion of innovations. Policy Sciences 5: 433—451. Weiss, Doyle L. 1968. Determinants of market share. Journal of Marketing Research V(3): 290—295. Wilson, J.B. 1975. Outlook for particleboard to 1980. Forest Pro- ducts Journal 25(11): 10-16. 212 Wisdom, Harold W. 1977. The impact of the environmental and energy crisis on the U.S. timber supply. Paper presented at the Third World Pallet Congress held at Virginia Polytechnic Institute and State University, Blacksburg, Virginia, October 3-5, 1977. Wright, Maurice G. and Robert B. Phelps. 1967. Trends in particle- board, insulation board, and hardboard industries, 1956-66. USDA Forest Service Research Paper WO-S. Washington, D.C.: U.S. Government Printing Office. Zivnuska, John A. 1963. The future of wood in a competitive market. Forest Industries 90(10): 40-41, 101, 104—106.