)V153I.} RETURNING MATERIALS: P1ace in book drop to uaamuss remove this checkout from “ your record. FINES will be charged if book is returned after the date stamped be1ow. INTENSIVE AND EXTENSIVE GROWTH IN THE FINNISH AND U.S. FOREST INDUSTRIES BY Marko Tuomas Katila A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Forestry 1988 507‘6’95/3 ABSTRACT INTENSIVE AND EXTENSIVE GROWTH IN THE FINNISH AND U.S. FOREST INDUSTRIES BY Marko Katila This study examines the sources of growth in the Finnish . and 0.8. mechanical forest and pulp and paper industries during the period 1958-82. The main objective is to separate growth into extensive and intensive components, the former meaning increasing use of resources at the same technological level, the latter meaning more effiCient use of existing resources. These components are analyzed to learn about the role of technological change, labor, capital, material inputs and energy in the growth process as a guide for future Finnish and 0.8. forest policies. Two approaches containing four basic models for the pulp and paper industry and three for the mechanical forest industry were applied. The production functions applied were two variations of a Cobb-Douglas function and a factor augmenting CES production function. Translog production function was applied implicitly in a total factor productivity index. The growth processes in the Finnish and 0.8. forest industries were shown to differ in nature, the differences being more apparent after the mid-70's. Growth in Finland has become more intensive over time, emphasizing the role of technological change or total factor productivity. In the 0.8. , growth has been more of the extensive type, emphasizing the role of capital deepening. Total factor productivity analysis in a gross output framework showed that capital and increased use of wood, chemicals and other material inputs have become more central to the growth process in the U.S. P & P industry, while in Finland their relative importance decreased towards the end of the study period. Comparison of the gross output and value added productivity measures revealed that the value added framework overestimated the role of productivity both in the Finnish and U.S. P 8 P industries. Although the results emphasize the importance of technological change to output growth, capital investment is suggested to be central to the growth process through a complementary relationship between the two factors. Capital intensity of production has increased significantly faster in the Finnish forest industries than in the U.S. , which could explain the differences in the role of total productivity. Increased capital intensity of production with limited substitution possibilities has meant greater vulnerability to changes in demand for forest products. As a result, forest industries in both countries have experienced short-term fluctuations in total productivity, especially during the mid-70 ' s. ACKNOWLEDGEMENTS The impetus for this study developed during my Master's research on technological change in the Finnish pulp and paper industry. I thank Dr. Markku Simula for whetting my interest in this area of research. The work was supervised by Dr. Markku Simula, who also provided some of the data used in the study. Dr. Robert S. Manthy - the chairman of the doctoral committee - gave valuable advice, and helped me shape the final format of this dissertation. During my academic studies I have received guidance and support from Professor Paivio Riihinen. The amenable research environment at the Department of Social Economics of Forestry, University of Helsinki, greatly contributed to the completion of this study. My studies in Michigan State University were made possible by grants from the ASLA-Fulbright Foundation, the Finnish Academy, and the Yrjo Jahnsson Foundation. I am grateful to all the above. I also want to express my gratitude to the other committee members at Michigan State University and everybody else who has contributed to this study. iv TABLE OF CONTENTS EASE LIST OF TABIJESOOOQQQQQOOoooooooooooooo00000000000000... Viii LIST OF FIGURES......................... ....... ........ x I INTRODUCTION..................................... '1 The Problem................................... 1 Objectives and Scope.......................... 4 Method........................................ 6 II ANALYTICAL FRAMEWORK ............................ 8 Aggregate Production Functions................ 8 Representation of Technological Change........ 13 Total Productivity, Technological Change and the Theory of Index Numbers................. 16 III ALTERNATIVE APPROACHES IN MEASURING CONTRIBUTIONS To GROWTHOCOOOOOOOOOOOOOOOOO......OOOOOOOOOOOO 23 Production Function Approach.................. 23 Index Theoretical Approach.................... 29 "No-Quality-Change" Models...... ...... ...... 30 "Explain-Everything" Models................. 32 Value Added vs. Gross Output Measurement...... 35 IV MODELS AND VARIABLES............................. 40 Models........................................ 40 Cobb-Douglas Production Function............ 40 Modified Cobb-Douglas Function.... ...... ... 41 Estimation of Models.......................... Measurement of Variables and Sources of Data.. Factor Augmenting CES-Function............. Translog Total Factor Productivity Index... V TRENDS IN THE USE OF FACTORS OF PRODUCTION AND THEIR PARTIAL PRODUCTIVITIES.............. Mechanical Forest Industry.................... Pulp and Paper Industry....................... Discussion.................................... VI COMPARISON OF SOURCES OF GROWTH IN THE FINNISH AND U.S. FOREST INDUSTRIES ....-...... Mechanical Forest Industry.................... Pulp and Paper Industry....................... Production Function Models.................. Total Factor Productivity Index............. VII DISCUSSION AND CONCLUSIONS....................... APPENDIX APPENDIX APPENDIX APPENDIX APPENDIX APPENDIX Differences and Similarities Between Countries and Industries.................... Conclusions and Implications.................. Related Research........................... Derivation of Aberg's Model................ Data for the Finnish Forest Industries..... Data for the U.S. Forest Industries........ Derivation of the Real Material Input Series Output and Partial Productivity Changes for the Finnish and U.S. vi Forest Industries..... Page 43 45 46 47 57 57 62 67 72 72 80 80 85 97 97 99 108 111 113 115 117 122 APPENDIX J. Rates of Change in Factor Efficiencies, and Indices of Technological Change for the Mechanical Forest Industries........... APPENDIX K. Estimated CD Production Functions for the Finnish and U.S. Pulp and Paper Industries. APPENDIX L. Estimated CES Functions for the Finnish and U.S. Pulp and Paper Industries......... APPENDIX M. Rates of Change in Factor Efficiencies, and Indices of Technological Change for the P & P Industries....................... APPENDIX N. Factor Input Shares in the Gross Value of Production in the P & P Industries......... APPENDIX 0. Annual Changes in Total Factor Productivity in the Finnish and U.S. P E P Industries... APPENDIX P. Development of Alternative Real Value Added Measures in the U.S. P & P Industry........ APPENDIX Q. Contributions of Capital, Material and Energy Deepening and TFPI to Labor Productivity Growth........................ BIBLIOGMPHY ...0.0............OOOOOOOOOOOOOOOOO0...... 127 129 130 131 133 134 135 136 137 LIST OF TABLES IQQIQ Page 1. Sources of Labor Productivity Growth in the Finnish and U.S. Mechanical Forest Industries (percent)...... 75 2. Contributions to Output Growth in the Finnish and U.S. Mechanical Forest Industries, 1959-82....... 78 3. Sources of Labor Productivity Growth in the Finnish and U.S. P & P Industries (percent).................. 83 4. Contributions to Output Growth in the Finnish and U.S. P & P Industries, 1959-82................... 84 5. Alternative Measures of Annual TFP Change in the Finnish and U.S. P & P Industries (percent).......... 87 6. Contributions to Output Growth in the Finnish P & P Industry (percent)............................. 91 7. Contributions to Output Growth in the U.S. P & P Industry (percent)00.000.000.000...00......00......O. 91 C1. Data for the Finnish Mechanical Forest Industry...... 113 C2. Data for the Finnish P & P Industry.................. 114 D1. Data for the U.S. Mechanical Forest Industry......... 115 D2. Data for the U.S. Mechanical P & P Industry.......... 116 F1. Output and Partial Productivity Changes in the Mechanical Forest Industries......................... 122 F2. Output and Partial Productivity Changes in the P&PInduStrieSOOOO....00............OOOOOOOOO0.0.0. 122 Jl. Rates of Change in Factor Efficiencies, and Indices of Technological Change for the Finnish Mechanical Forest Industry......OOOOOOOOOOOOOO......OOOOOOOOOOO. 127 viii 111111.: J2. Rates of Change in Factor Efficiencies, and Indices Q1. oz. of Technological Change for the U.S. Mechanical Forest IndustrYOOOOOOOOOOOOOOO......OOOOOOO0.0... Rates of Change in Factor Efficiencies, and Indices of Technological Change for the P & P Industries..... Factor Input Shares in the Gross Value of Production in the P & P Industries.............................. Annual Changes in Total Factor Productivity in the Finnish and U.S. P & P Industries.................... Contributions to Labor Productivity Growth in the Finnish P & P Industry (percent)..................... Contributions to Labor Productivity Growth in the U.S. P 8 P Industry (percent).................... ix 128 129 133 134 136 136 LIST OF FIGURES E19223 Page 1. Isoquants of Production Functions with Alternative Elasticities of Substitution........... 12 2. Output and Input Development in the Finnish Mechanical Forest Industry......................... 58 3. Labor and Capital Productivity in the Finnish Mechanical Forest Industry......................... 59 4. Output and Input Development in the U.S. Mechanical Forest Industry......................... 60 5. Labor and Capital Productivity in the U.S. Mechanical Forest Industry......................... 61 6. Labor Productivity in the Finnish and U.S. Mechanical Forest Industries....................... 62 7. Output and Input Development in the Finnish P8PInduStry.OO.......OOOOOOOOOOOOOO....0.......O 63 8. Output and Input Development in the U.S. P & P IndustrYOOOO0.00.00.00.00.........OOOOOOOOOOOOO.... 64 9. Partial Productivities in the Finnish P & P Industw..........O............OOOOOOOOOOOO.......O 65 10. Partial Productivities in the U.S. P 8 P Industry.. 66 11. Labor Productivity Development in the Finnish and U.S. P&PIndustrieSOOOOOOOOOO......OOOOOOOOOOOOOO 67 12. Capital-Labor Ratios in the Finnish and U.S. Mechanical Forest Industries....................... 69 13. Capital-Labor Ratios in the Finnish and U.S. P&PIndustrieSOOO.........OOOOOOOOOOOOOOO0.00.... 7o 14. CES Indices of Technological Change in the Finnish and U.S. Mechanical Forest Industries...... 79 Figure 15. 16. 17. 18. 19. CES Indices of Technological Change in the Finnish andUOSO P8PIndustrieSOOOOOOOOOOOO00...... ...... 85 Total Factor Productivity (TFPI) in the Finnish andUOS. PaPIndustrieSOOOOO......OOOOO0.0......O 89 Total Factor Productivity (TFPI) in the Finnish and U.S. P & P Industries, Moving Average..... ..... 90 Total and Value Added Productivities in the FinniShP&PIndustry.....OOOOOO0.0.0.0...... ..... 93 Total and Value Added Productivities in the U.S. P5PIndustry..........OOOOOOOOOOOOOOOOOO0.0.0.... 94 Development of Alternative Real Value Added Measures in the U08. P&PIndustrYOOOO....OOOOOOOOOOOOOIOOO 135 xi I INTRODUCTION Ths_£r9hlem Economic growth is a sustained increase in the output of goods and services used to satisfy human wants. Widely accepted as a national goal, economic growth implies greater real income per capita, and usually makes other economic and social objectives easier to achieve. Measures of growth are widely used as indicators of overall economic performance; the presence or absence of growth is looked upon as an indicator of the success or failure of economic policy. Formation of economic policy is conditioned by views about the causes of growth. In quantitative studies of 'growth, growth has usually been separated into extensive and intensive components, the former meaning increasing use of resources at the same technological level, the latter meaning more efficient use of already existing resources by use of a different technology and/or improved quality of inputs. In the more precise terminology used in production theory, extensive growth refers to capital and material deepening, intensive growth corresponds to the concepts of technological progress or total factor productivity (TFP) growth.1 The essential quantitative characteristic of intensive growth or ;In this study the terms technological and total factor productivity change are used interchangeably. 1. 2 technological change is a shift of a production function, enabling greater output to be produced with the same quantity of inputs, while extensive growth or capital deepening corresponds to a movement along a production function whose shape is determined by given technology (Solow 1957; Usher 1980a, p. 259-60). In Finland there have been only a few studies on the sources of economic growth. This is true especially at the industry’ level - also in ‘the forest industries (related research is reviewed in Appendix A). This study is carried out to learn more about the process of growth in the forest industries both in Finland and the U.S. Since the 1950's the output of the Finnish forest industries has been growing rapidly; the average annual growth rate in output volume for the period 1955-80 was 5 percent. High growth rates have meant increased use of resources needed to produce the desired output. A raw material and production oriented "growth strategy" has been possible because of advantageous preconditions for growth. Demand for forest products has risen steadily, providing new market opportunities for Finnish forest products. The domestic timber base has been growing even faster than needed to satisfy the actual demand for industrial roundwood. In addition, low real interest rates and institutional factors (e.g., tax laws) have created a positive economic environment for new forest industry investments. 3 However, the Finnish forest industries now face a situation where growth based on the increased use of inputs ‘will not be possible to the same extent as before. Significant expansion of production capacity is not possible because present capacity is already nearly equivalent to the allowable cut, and imports of roundwood cannot be increased in a significant amount. Even if the plans for increasing the allowable out were succesful, the changed selling behavior of nonindustrial private forest owners may reduce the supply of roundwood permanently below the allowable cut. Future expansions may also be reduced by increased real interest rates and by uncertainty about future energy supply (prices). Thus, it appears that future growth of the Finnish forest industries must be based on the more efficient use of existing resources. The assessment of opportunities for growth and formulating of new growth strategies requires knowledge of the factors underlying past growth. Information about the role of total factor productivity (that is, technology) in growth is especially needed. International competitiveness depends not only on price and cost development of forest products and factor inputs, but also on productivity growth. Increases in total factor productivity improve profitability, and can secure and even create jobs in the long run and provide higher wages. High unemployment levels since the mid-19708, and increasing competition for resources, especially for capital, have also increased the 4 need to know more about the sources of growth in the Finnish forest industries to improve the allocation of resources. W This study estimates the contributions of various factors to the growth of output in the Finnish and U.S. forest industries during the period 1958-82, with emphasis on growth in the Finnish forest products industries. The study aims at decomposing the growth into extensive and intensive components, and analyzing these components in more detail as a guide for future forest policy. The role of capital, labor, material inputs, energy, and especially, the role and nature of technological change responsible for this growth are examined. Trends in the use of factors of production and their partial productivities are also studied. Before discussing empirical applications, the theoretical aspects of growth analysis and the_ measurement of variables are examined. Comparative analysis of the sources of growth in the Finnish and U.S. forest industries is carried out for two main reasons. First, it helps in appraising a country's past performance and provides information about possible differential technological opportunities for increasing the role of productivity in growth. Second, international comparison facilitates validation of applied models. Production functions are estimated separately for the mechanical wood and pulp and paper industries. Total factor 5 productivity indices are calculated for pulp and paper industries. The study period, the level of aggregation and the inclusion of forest industry subsectors for study were determined by the availability of comparable data. The following industry sectors are included in the study under the headings "Mechanical Wood" and "Pulp and Paper" (note: the Finnish coding system changed in 1970): Finland USA ISIC code SIC code W Sawmills and planing mills 331111 2420 Other manuf. of wooden stuct. 331122 - 2430 Veneer and plywood mills 331191 W Pulp mills 34111 2611 Paper and paperboard mills 34112 2621,2631 Building paper and board mills 2661 The "mechanical wood" industry in the United States is defined to consist of SIC (Standard Industrial Code) industry 2420, which includes Sawmills and Planing Mills (SIC 2421), Hardwood Dimension and Flooring (SIC 2426) and Special Product Sawmills (SIC 2429), and of the industry SIC 2430, which includes Millwork (SIC 2431) , Wood Kitchen Cabinets (SIC 2434), Hardwood Veneer and Plywood (SIC 2435), Softwood Veneer and Plywood (SIC 2436) and Structural Wood Members (SIC 2439). In 1972 small revisions took place in the classification of products in SIC industries 2420 and 2430, but those changes partly cancel each other out, and the remaining source of error is insignificant. The Finnish and U.S. data are not fully consistent in coverage and 6 classification, but it is assumed that these differences will not have a significant effect on the results. fleshed In this study both a production function and total factor productivity (TFP) index approach is applied to quantify the sources of growth in the Finnish and U.S. forest industries. Functional forms for the production functions are chosen on the basis of their suitability for growth analysis. The validity of underlying assumptions and the robustness of the results are studied with the use of alternative models. The production functions used here are the Cobb-Douglas and the CES (constant elasticity of substitution) functions. A third type of function - the translog production function - is applied implicitly in a TFP productivity index. Application of index number theory permits circumvention of some of the problems generally associated with explicit production function analysis. Given certain assumptions, TFP indices allow the estimation of the contributions of various factors to the growth of output without explicitly specifying the form of the production function. They are also easier to apply than production functions in growth analysis when more than two factors of production are considered. This is important in this study because an attempt is made to quantify the role of four factors in the growth process: labor, capital, material inputs and energy. On the other 7 hand, TFP indices cannot be used to test the associated hypotheses (e.g. , the type of the production function or technological change), or to obtain parameter estimates describing production relationships. The choice of applying two alternative approaches to study the sources of economic growth of the defined forest industry sectors in Finland and the U.S. is based on the recognitions that the choice of the study approach may have a great effect on the results, and that the two approaches complement each other to some extent. When alternative models are used the validation of the results also becomes. easier. It is also valuable for future research to find out to what extent the two approaches produce comparable results. II ANALYTICAL FRAMEWORK Aggregate_£reduetien_runstigne In an attempt to identify the sources of economic growth, or to examine the policies affecting growth, the explicit or implicit use of an aggregate production function is almost indispensable.2 A production function defines the relationship (based on physical or engineering considerations) whereby alternative combinations of factors of production are transformed into outputs. In the economic theory of production, a production function represents the locus of minimum input requirements needed to produce given level of output. Mathematically production function can be written as Q = f(x) 6f/6xi = £120, 62f/6xi2 =f11$0 (1) where Q is the quantity of output, x represents the vector of inputs, 6 is the derivative, and f is a concave function. Often output is expressed as a function of only two homogeneous inputs - labor (L) and capital (K): Q = f(L,K) . (2) A production function is not an unambiguous concept, but it can represent a number of diverse concepts. One can make a distinction between ex ante and ex post, micro and macro, frontier’ and average, and short-' and. long-run. production '2The (neoclassical) concept of a production function is not accepted by all economists (see discussion in e.g., Jones 1975) . 9 functions. An ex ante or blueprint production function represents the set of most advanced techniques currently available (Sato 1975, p. xxiv). In empirical work we are not studying theoretical relationships, but we try, ex post, to link the actual recorded output with the actual factors of production. Ex post functions are of a short-run nature because techniques are more or less fixed. Macro production functions are employed to describe industry or economy-wide production relationships. They are traditionally built from micro units assuming that all firms share an identical production function. This kind of traditional macro production function can produce results that are difficult to interpret or operationalize, because it represents average technology that does not necessarily exist as such. Firms do, however, differ in productive efficiency, and due to this, they are not so much interested in average as in best-practice technology. (Johansen 1972; Sato 1975, xxi) Macro functions based on the efficiency distribution of firms, and frontier production functions, which refer to the best-practice technologies existing at a given point of time, have been developed to provide a more realistic description of the industrial structure and productive efficiency of a firm or an industry, and to produce results that would be easier to operationalize at a firm level or in developing instruments for industrial policy (see e.g., Johansen 1972: 10 Sato 1975: Forsund et a1. 1980). The lack of suitable data does not permit to use this approach in this investigation. The production function has four characteristics which are central in the analysis of changes in the production technology: the efficiency of production, degree of economies of scale (or returns to scale), factor intensity, and the elasticity of substitution. An increase in the efficiency of production refers to a reduction in the quantities of factors used in producing the unit output. Alternatively, it can be seen as an equal reduction in the unit cost of all factors of production by applying better techniques. The degree of- economies of scale expresses the change in output that results from an equiproportional change in all inputs. A production function is said to be homogeneous of degree n, if Q=f(ax)=anf(x) for all a>0. Factor intensity refers generally to capital intensity; i.e., the capital-labor ratio for given relative prices. Capital intensity may change as a result of changes in relative prices, or the change may have a purely technological origin. The fourth central characteristic of a technology is the ease with which one input can be substituted for another. It can be measured locally by the elasticity of substitution (0) a - dln(K/L)/dln(fL/fx) (3) Assuming perfect competition and profit maximization, a can be written a , glam/1.1 = W (4) n(w/r) d(w/r)/ (w/r) 11 where w and r represent factor prices. The elasticity of substitution is therefore a measure of how rapidly factor proportions can change for a change in relative factor prices (Brown 1966, p. 12-20; Intriligator 1978, p. 264-5). Production functions can be classified according to the size of a (see e.g., Ollonqvist 1974). When a = 1, we have one of the most widely used production functions for empirical applications - the W. In a two-factor case it is written Q = ALaxfi A20, azo, 320 (5) where a and B are elasticities of output with respect to inputs, and A is a parameter embodying changes in the efficiency of technology. The sum of the parameters a and 8 indicates the degree of economies of scale. In the W W, the elasticity of substitution is not given a priori, but it is assumed to be constant (Arrow et al. 1961): Q - u[(6K‘“ + (1-6)L'a]’V/° “>0, v20, 0<6<1, 62-1 (6) where n is the efficiency parameter, v shows the degree of economies of scale, 6 is a distribution parameter representing capital intensity, and a is a substitution parameter. The elasticity of substitution (0) is derived from the substitution parameter as a - 1/(1+a). The CD-function is a special case of the CES production function, because when a --> 0, a approaches one. ngntigfiLg 12 input-output function (Q=min(aL,bK)) is also a special case of the CES-function, for as a --> on, a approaches zero (Intriligator 1978, p. 273-5). The isoquants of these three types of production functions are shown in Figure 1. a=0 \ 0:1 can L Figure 1. Isoquants of Production Functions with Alternative Elasticities of Substitution. All the above described production functions share one common characteristic - the elasticity of substitution is constant. In the v ' s ’t s t' o c o 'o , a is allowed to vary with changes in factor proportions (see Sato 8 Hoffman 1968; Revankar 1971). In the recent years, one of the most widely used production functions is the gnaneeengen§e1 logeritnmic (tganslog) pgedueeien functien of the form l3 an a lnao+ Eailnxi +(1/2)Ezfiij1nxilnxj (7) where anutput, ao=efficiency parameter, and Xj=input j. 0:1 and 31:) are unknown parameters such that fo( x) is concave, nondecreasing, and linearly homogeneous over the relevant range of x's (Christensen et al. 1973). The translog production function can be viewed in three alternative ways: as an exact production function, as a second-order Taylor-series approximation to a general but unknown production function, or as a second-order approximation to a CES production function (Boisvert 1982, p. 5-6) . Because of its generality and flexibility, translog production functions have most often been used to approximate any unknown linearly homogeneous production function, either explicitly, or implicitly in TFP indices (see p. 20). WW Technological change has been claimed to play a crucial role in the process of economic growth, and consequently its analysis occupies a central place in growth studies. Technological change (progress) can be characterized in several ways, but its essential quantitative characteristic is to shift the production function enabling greater output to be produced with the same gneneiey of inputs, or the same output with lesser inputs (Kennedy & Thirwall 1972, p. 12). More formally, technological change (TC) can be defined by: TC = dln(Q)/dln(t) (3) 14 where time (t) represents technology. Technological change can also be defined from the dual side of production. According' to the ‘theory' of’ duality, for’ each. production function there exists a dual cost function reflecting the production technology (e.g., Smith 1978). On the cost side, the rate of technological change can be measured as TC 2 dln(C)/dln(t) (9) where C is the cost of producing output Q, and t represents technology. Another important aspect of technological change is its neutrality or bias, the latter meaning change that leads to a greater saving in one input relative to another. Production ‘theory recognizes several definitions of technological change, of which definitions by Hicks, Harrod and Solow have been widely used (see e.g., Beckman & Sato 1969; Fun & Wibe 1980, p. 3-21). Technological change is Hicks-neutral if it does not change the marginal rate of substitution between the inputs at given factor proportions: that is, the ratio of marginal products stays constant. Harrod-neutral and Solow- neutral technological change, on the other hand, refer respectively to cases in which the ratio of marginal products remains constant when measured with respect to identical capital-output or labor-input ratios (e.g., Nadiri 1970, p. 1142-3). Definitions of labor- and capital-saving technological bias follow from the definitions of neutrality. We may define, for example, Hicks bias as 15 a $11 311 BR Kdt LLL K/L constant (1°) >0 labor-using (capital-saving) =0 neutral <0 capital-using (labor-saving) A distinction is made conventionally between embodied and disembodied technological change. In the first case, technological change is built into new capital goods, while in the second case, it is a function of time and independent of any changes in factor inputs. Disembodied technological change is generally attributed to improvements in the organization and management, and in techniques that enhance_ the productivity of old equipment along with new (Kennedy & Thirwall 1972, p. 31-39). A neoclassical production function exhibiting disembodied (neutral) technological change (progress) may be written in a general form Q - f(x,t), ft>o (11) where x is the vector of factors of production, and t denotes time or level of technology at point t (see Solow 1957). The change in output over time is defined by taking the total derivative of the above equation dQ/dt - 2(6f/6x1)(dxi/dt) + 6f/6t. (12) The first term on the right indicates the change in output due to increased inputs, and the second term on the right indicates the change in output due to disembodied technological change. Equation (11) is often expressed in factor augmenting form that may be written in a two-factor case as 16 Q = f(a(t)L.b(t)K) (13) where a(t)K and b(t)L can be identified as "effective" labor and capital. Technological change is said to be purely capital-saving if a(t)>0 and b(t)=0, whereas, it is purely labor-saving if b(t)>0 and a(t)=0. When a(t)=b(t)>0, technological change is neutral. Another distinction is that between exogeneous and endogeneous (or induced) technological change. In the case of the former, technological change as such is not explained; it is just exogeneous to the economic system. In the case of endogeneous technological change, the expansion of technological possibilities is explained explicitly by economic factors such as long term changes in the relative prices of inputs, investment in education and research plus in inputs required for changing over to improved industrial methods (Kennedy & Thirwall 1972, p. 13: Heertje 1977, p. 173-4). In recent years, the empirical analysis of production has changed its focus from explicit production function analysis to the application of ‘total factor'jproductivity (TFP) indices. Total factor productivity is often defined as a ratio of output to the aggregate of all factor inputs. If an aggregate index of inputs is defined as Ft = g(x), where x 17 is the input vector, total factor productivity (At) can be measured as the ratio At = Qt/Ft = Qt/g(x)- (14) From this equation two important conclusions can be derived. First, total factor productivity and technological change are similar concepts, because both measure that part of the output growth that cannot be accounted for by the factors of production. Second, a productivity index always implies the existence of a production function and vice versa (Diewert 1976). The implicit production function underlying the measure At can be expressed as Qt - PtFt - Ptg(x). (15) One of the main problems in the index approach is, how should the inputs be arranged, or in other words, what form should the (implicit) production function take. In order to be consistent, factors of production should be aggregated in such a way that would correspond to the "factual" production function Qt = f(x). (15) TFP index approach does not therefore free us from the choice of the production function, even if the precise form of the function does not necessarily have to be estimated directly. For example, application of Kendrick's arithmetic index (1961) implies the assumption that the "factual" production function is of CES-type (Kendrick & Sato 1963). It is evident that even if the inputs were measured properly and 18 the weighting (index number) problem were solved, the validity of TFP measures depends greatly on how well the underlying economic theory describes the actual production relationships. Difficulties involved in the productivity measurement have been significantly reduced by the recent developments in the index number theory and by the introduction of flexible functional forms. The modern theory of total productivity measurement is based on the explicit recognition of the connection between the theory of production and the application of the Divisia index principle (Diewert 1976, 1980, p. 443-52). Given certain conditions3 production function can take the special form Q(t) = A(t)f(X(t)) (17) where the multiplicative factor A(t) measures cumulatively the shifts in the production function over time (Solow 1957: Jorgenson S: Griliches 1967: Diewert 1980, p. 443). If one takes the total derivative of Equation (17), and divides it by Q(t), one obtains the following equation fungus-.1 2113mm Q(t) A(t) * A‘t’ 6x1 Q(t) ‘18) where dots indicate time derivatives. Then, if inputs are being paid the value of their marginal products (Sf/6x1)=pi), TFP growth can be expressed as the difference between the 3The production function should be linearly homogeneous, concave and nondecreasing, and it should exhibit neutral technological change and constant returns to scale. Cost minimization is also assumed. 19 rate ‘of change of output and weighted (average) rates of changes in inputs: 11:1 3 élri _ £1111 A(t) o Ewix1(t) (19) where the weights Vi are relative cost shares (wi=pixi/Q) at given points of time. Note that this Divisia index defines the geneinnene rate of total factor productivity change, but because data are available only in discrete form, Equation (19) must be approximated. using' a discrete index number formula.4 Diewert (1976) has shown that if a particular form is assumed for the production function, it is possible to construct an exee; discrete index of productivity change. A quantity index Q is defined to be. exact for ‘the given functional fomm f, if the ratios of the outputs between any two periods are equal to the index of inputs: fx)'°1'1/° (28) Before statistical estimation is attempted, the type of factor augmentation needs to be specified to avoid the Diamond and McFadden "impossibility" theorem, which states that variations in individual efficiencies of inputs and 0 cannot be separately identified (Nerlove 1967, p. 92-98). A common specification is to assume that factor augmentation occurs at a constant exponential rate (a(t) = aoegt, b(t) = boegt). This type of models have been estimated, e.g., by David 8 van der Klundert (1965) , Ferguson & Moroney (1969) and Kalt (1978). In this study a factor augmenting CES-type function allowing nonneutral technological change will be applied implicitly. The main advantage of the CES-production function is its generality and flexibility compared to the CD-function. On the other hand, it has also several limitations. The 28 estimation of the CES-function is relatively difficult, and the empirical evidence seems to indicate that its parameters are highly sensitive to slight changes in the data, measurement of the variables, and the methods of estimation (Nadiri 1970, p. 1151). This was proved to be true also in the case of ‘the Finnish forest industries (Simula 1979, 1983) . Another limitation of the CES-function is that it is difficult to apply in the case of’ multiple factors of production (n>2). Also, the analysis of technological change is not unambiguous, because technological change can be manifested in one or more of the coefficients (Brown 1966, p. 59-61). Some of the limitations cf the CD and CES production functions can be handled. with. alternative functional forms. The variable elasticity of substitution function which is a generalization of the CES-function, and the translog production function, which is a generalization of the CD- function, both allow for different a at different input-input and input-output ranges (Intriligator 1978, p. 280). The latter function can also readily handle the problem of pairwise differing elasticities of substitution for a set of several inputs. Empirical applications of the variable elasticity of substitution. production functions seem to, however, often produce results that are statistically insignificant and sometimes even economically meaningless. 29 Application of the widely used translog production function (or its dual) is also fraught with some difficulties. First, it is theoretically somewhat ambiguous because of its several possible interpretations (see p. 12- 13) . Second, the problems encountered in obtaining reliable estimates of the parameters of the function are difficult to solve. Third, in order to take advantage of the function's flexibility, one encounters numerous computational difficulties and additional data requirements (Boisvert 1982, p. 31-35). In this study the translog production function will be applied - not explicitly but implicitly in a total factor productivity index as a second-order approximation to an unknown function. In this type of application most of the criticisms raised against the translog production function do not apply because the function is not estimated statistically. In general, the gain of realism from alternative production forms like, e.g., translog, vintage or frontier production. functions, comes at the cost of additional - sometimes excessive - data requirements and computational difficulties. These complications often make the use of simpler and more manageable functions advantageous compared to more sophisticated and complex models. WW Total factor productivity indices have been widely used in assessing the contributions of individual factors of 30 production to the growth of output. All the models based on this approach share the same logic: the change in output can be attributed to increasing factor inputs by weighting different inputs in some "appropriate" way, while the change in output due to increase in total factor productivity can be obtained by subtracting this aggregate input component from the actual increase in output(s) . Direct estimation of a production. function is unnecessary, even if every productivity index implies a particular production function. All measures of TFP are more or less based on the neoclassical framework: weighting is based on the equality of factor prices and marginal productivities, production technology exhibits constant returns to scale, etc. In the measurement of total factor productivity, two major approaches can be distinguished, which could be called the "no-quality-change" approach and the "explain-everything" approach using Tolley's (1961) terminology (see Denison 1961). In the "no-quality-change” approach only conventional factors of production are measured; i.e. , inputs are not adjusted for changes in quality. The remaining residual output that cannot be explained by the increase in conventional inputs is called total factor productivity or technological change. For example, the familiar arithmetic (Abramovitz 1956, Kendrick1961) and geometric (Solow 1957) 31 index applications belong to this category. Abramovitz's "residual index" is expressed as dA/A = dQ/Q - aodL/L - bodK/K (29) where a0 and b0 are fixed weights. Kendrick measures total factor productivity using a distribution equation derived from a homogeneous production function and the Euler condition: .. ____K LQ —_ dA/A - (wL1+rK1)1/ (8L0+1‘K0) - 1 (30) where the weights w (wage rate) and r (the rate of return on capital) are allowed to change from period to period. The geometric index applied by Solow is based on a linear homogeneous production ‘function with constant returns to scale and disembodied technological change: dA/A = dQ/Q - MdL/L) - 3(dK/K) (31) where a and H are the shares of labor and capital. The validity of these indices depends greatly on the way of measuring inputs and outputs and on the weighting scheme used for their aggregation. The derivation and application of these models require also several simplifying assumptions which don't necessarily apply in reality. For example, Abramovitz's index is based on the unrealistic assumption that marginal productivities are not affected by a change in capital-labor ratio (Lave 1966, p. 8). Solow's index, on the other hand, is based on the assumption that the underlying production function is of Cobb-Douglas type, while Kendrick's 32 index implies a CES-function (Kendrick & Sato 1963, Lave 1966, p. 7-13). Common to all these indices is also the implicit assumption that only capital and labor are important factors of production; material inputs and energy are excluded. In recent years, it has become more common to measure TFP change in gross-output framework applying, e.g., the T6rnqvist index presented earlier in this study. The development in the economic index number theory has also led to the use of more flexible indices. concerning the underlying production function which have replaced the "traditional" indices in empirical work (see Chapter II). Also these indices could be applied without making any quality adjustments and measuring output with value added, but most often they are applied in a more "ambitious" way. II ' _. II All the indices where inputs are measured conventionally share the same :major’ limitation: they' cannot be used. to "explain" economic growth, because the resulting measure of technological or total factor productivity change includes all the possible factors that could shift the production function, e.g., economies of scale, improved quality of labor and capital, more efficient management, measurement errors, etc. Therefore, even if the contribution of TFP to the growth can be estimated, the sources of that growth - which are of 33 most interest to policy makers and firm managers - remain unknown. Dissatisfaction with models that left much of the growth unexplained9 led to the new approach that tried to narrow the residual, and thus reduce our ignorance concerning sources of economic growth. The works of Denison (1962, 1967, 1974) and Jorgenson with several colloborators - Griliches (1967) , Christensen (1969), Gollop (1977), and Fraumeni (1980) - are the most important studies that employ this approach. Denison uses the production function (similar to CD) as an organizing device or accounting format to decompose the residual into economies of scale, improvements in resource allocation, and finally advances of knowledge (Nadiri 1970, p. 1165-6). Denison sought to narrow the residual by adjusting labor for changes in quality due mainly to more education, changes in the composition by sex and age and shorter working hours. After quantifying the contributions to growth of all major factors, he was left with a "final residual", which he called advances in knowledge. The underlying relation between growth of output and various explanatory factors in Denison's model is: dQ = u(Eaidxi + Eyj + J) (32) where dQ is the growth rate of output, )1 is a measure of economies of scale, :11 represents the income share of the factor represented by dxi, Yij refers to the growth rates of 9For example, in Solow's (1957) study the residual amounted to almost 90 percent. 34 various disequilibrium factors, and J is the final residual (Nadiri 1970, p. 1166). Because Denison adjusted only labor for changes in quality and left capital unadjusted, his approach could be called a "partial-quality-change" approach. Jorgenson and Griliches (1967) , on the other hand, tried to explain away the very existence of total factor productivity by adjusting total output and input data for errors of aggregation, errors in investment goods prices, and errors of utilization of cap- ital and labor. The rate of TFP growth 1 is calculated using Tornqvist's discrete approximation of the Divisia index and assuming producer's equilibrium, constant returns to scale and perfect competition: A/A =- 6/9 - 32/): = “133 - “lg: (33) where wi is the relative share of the value of the ith output in the value of total output, and wj is the value of the jth input in the value of total input. Maybe the most significant difference between this approach and Denison's approach is in the measurement of capital: Denison uses capital stocks (excluding depreciation) while Jorgenson 8 Griliches measure capital input by flows of capital services. Attempts to reduce the magnitude of the residual or explain away its existence are necessary to obtain a deeper understanding of the growth process, and to identify the major variables for decision makers. Unfortunately, this approach can be time consuming and very expensive because of 35 its great demand for data. In fact, the lack of data often precludes the application of this type of approach. The results are also sensitive to the classification of the sources growth and the assumed causal relationships, and especially to the choice of conventions for measuring real factor inputs (Nadiri 1970, p. 1167-9). In summary, given certain assumptions TFP indices allow the estimation of the contributions of various factors to the growth of output without explicitly specifying the form of the aggregate production function. They are also easy to apply in the case of multiple factors of production. 0n the. other hand, they cannot be used for testing the associated hypotheses, e.g. , the type of the production function or technological change, and moreover, the parameters of the production function reflecting, e.g., changes in the individual efficiencies of inputs that underlie the total productivity change, are left unknown. In this study both the production function and index theoretical approach are used. Before the models to be applied in this study are presented, an important question concerning the proper way of measuring the output is briefly discussed. WW Most analyses of productivity growth and technological change have measured output in terms of real value added (real gross output less real intermediate inputs). In such 36 studies output (Qva) is a function only of labor, capital and time: . Qva I f(L,K,t) (34) On the economy level, working with a value added model of production is correct when dealing with a closed economy, or with an open economy where imports are considered as final goods or as being separable from primary factors of production (Denny 8 May 1977) . A value added technology at the economy-wide level also seems intuitively a reasonable concept, because when the production accounts of all industries are aggregated, inter-industry flows of intermediate inputs cancel out (no fear of double counting). At the subsector .or industry level the relevancy of value added as an output measure can, however, be strongly questioned. In 1970's the restrictions of the value added approach were formally presented, and attempts were made to test the existence of a real value added function (Sims 1969, Arrow 1972, Berndt 8 Wood 1975, Humprey 8 Moroney 1975, Denny 8 May 1977, Bruno 1978). . Sims (1969) and Arrow (1972) demonstrated that e;__tne Wueal) value added is an economically meaningful concept only given weak separability between the primary inputs (L,K) and intermediate inputs. A production function is defined to be weakly separable with respect to the grouping of factors of production, if the marginal rate of substitution between pairs of factors in the separated 37 group are independent of the levels of factors outside that group, or alternatively, if the Allen partial elasticities of substitution between a factor in the separable group and some factors outside the group are equal for All factors in the group (Berndt 8 Christensen 1973). Thus, only when the marginal rate of substitition between capital and labor is independent of the quantity of intermediate inputs M (including energy), can value added (Ova) be separated from the original function ng - F(L,K,M,t) (35) and it can be written as a sub-function of F: ng a F(f(L,K,t),M) (36) Qva - f(L,K,t) (37) In the value added model of production quite strict assumptions have to be made on the nature of the production process that reduce the generality of the model and often cause bias in the results. The value added approach at the subsector or industry level has also been criticized for implying irrational producer behavior because producers are not allowed to equate the prices of intermediate inputs to the respective values of marginal products as required by the profit maximization assumption. This again creates restrictions on the production technology in terms of substitution possibilities. Moreover, from Equation 37 it can be seen that technological change is allowed to affect output only through 38 function f, which leaves no room for intermediate inputs as a source of productivity growth or decline (see Hulten 1978). In the case of forest industries these restrictions mean that producers would not (necessarily) respond to changing relative stumpage or energy prices by substitution or increasing the efficiency of wood utilization (see Bengston 8 Stress 1986). It can also be asked how relevant it is to describe, e.g., lumber production with a production function where roundwood is not entered as an input. The calculation of reel value added is also problematic. In many countries (not in the U.S. or Finland), the method used to calculate real value added is the double deflation procedure. This requires price and quantity data on intermediate inputs at a disaggregated level, or reliable deflators (at purchaser's prices). Double deflation is also likely to produce biased real value added measures unless one of the following conditions is satisfied (Bruno 1978): 1) The volume ratio of intermediate inputs to outputs remains constant (regardless of the price of intermediate inputs) 2) The relative price of intermediate input (Pm/Png) remains constant 3) Production technology is separable with respect to primary inputs and intermediate inputs The Divisia quantity index of sectoral value added can then be derived from Equation 36 by logarithmic differentiation with respect to time and assuming constant 39 returns to scale, perfect competition and producer's equilibrium: dln(Qva/dt) - (dln(ng/dt) - wm(dlnM/dt))/wva, (38) where wm and wva are value shares of intermediate inputs and value added respectively (Sims 1969). As a result of the difficulties involved in the application of the value added approach, it can be concluded that if possible, it is preferable to use 1970. The year 1970 was chosen because it divides the study period into two subperiods of equal length. Compared to the CD-models, the model derived by Aberg can be considered as an improvement, if capital data are of poor quality, or if one wants to measure capital services, but relevant data on capital are missing. The derivation of this model and its estimation require however some 43 restrictive assumptions, of which the assumption of constant real return on capital is the most serious one. A detailed derivation of this model and its underlying assumptions are presented in Appendix B. W The estimation of a factor augmenting type of production function gives more detailed information about factors underlying TFP change. In the factor augmenting formulation with two factors of production Q = £4a(t)L,b(t)K), a(t) and ‘b(t) stand for the input augmenting factors or the "efficiencies" of labor and capital, respectively. The jproblem is that when technological change is factor augmenting, estimates for the rates of efficiencies of labor and capital cannot be uniquely determined unless the form of the production function is known a priori (Diamond 8 McFadden 1966, Sato 8 Beckman 1968). When differentiating the above equation with respect to time and assuming constant returns to scale we get Q - a(L + a) + pa’c + 6) (43) where ah= (6f/6L)(L/Q), fl - (l-a) = (6f/6K)(K/Q), and the dot refers to rate of change over time. If we write x = L/K then (Q/L) = as + as - 39c (44) Estimation of the efficiencies of capital and labor from Equation (44) is not possible because we have two unknowns and one equation. Sato and Beckman (1968) solved this problem 44 by deriving the following two equations using the definition of the elasticity of substitution (0) 4-4- (mm-5+4) (4S) i~=i>+ (a/a)(a-b+x) (46) where r and w stand for input prices of capital and labor. By solving Equations 45 and 46, and using Equation 44, we get the fundamental relationships for estimating a and b: _ é = (.4 - (Q/L)/(a-1). a f 1 (47) 5 = (of - (Q/K)/1! ;_ ' -‘ :- --, o_, -; ..,.. !!_‘ Analysis of sources of output growth showed that the growth processes in the Finnish and U.S. mechanical forest and P 8 P industries differ in nature, the differences being most. apparent after 'the 'mid-70's. Growth in Finland. has become more intensive over time, emphasizing the role of technological change or total factor productivity; in the U.S., it has been more of the extensive type emphasizing the role of capital deepening. Total factor productivity analysis in a gross output framework with a translog productivity index showed that capital and increased use of wood, chemicals and other material inputs have become more central to the growth process in the U.S. P 8 P industry, while in Finland their relative importance decreased towards the end of the study period. 97 98 All the models except the CD production function for the mechanical and P 8 P industries and CES function for the P 8 P industries showed that even before the mid-70's the role of total productivity in the growth process has been greater in the Finnish forest industries than in the U.S. Estimations of the factor augmenting CES production function showed no significant difference in the sources of output growth in the Finnish and U.S. P 8 P industries during 1958-75. But they do show technological change as a major contributor to output growth in the U.S. mechanical forest industriesl4. Estimated CD production functions, on the other hand, could not be used for reliable comparison because of their poor quality, caused most likely by a combination of restrictive assumptions underlying CD functions, simultaneity errors, and deficiencies in (capital) data. Inter-industry performance comparison based on the estimates of rates of technological change revealed that the mechanical forest industries are not less progressive than the P 8 P industries, opposite to conventional belief (see also Risbrudt 1979). When interpreting estimates of technological change, it should be remembered that the use of value added framework overstates the importance of technological change; this was demonstrated here for the pulp and-paper industries, but it has been shown to be true also 14Greber and White (1982) obtained the same result in their study on the U.S. lumber and wood products industry for a slightly different period. 99 for mechanical forest industries (Bengston 8 Strees 1986) . Therefore, one should interpret carefully (conservatively) the results of value added models. Comparison of total factor productivity (technological change) measures developed in this study to technological change in other industries (see e.g., Gollop 8 Jorgenson 1977, Fraumeni 8 Jorgenson 1980, Wyatt 1983, Karhu 8 Vainionmaki 1985) indicates that productivity has increased slowly both in the Finnish and U.S. forest industries. This gap can be interpreted as a signal that technological opportunities may be open to the forest industries for increasing the role of TFP in the growth process. Reliable comparison of the rates of productivity growth between different industries would, however, require the use of similar kind of methods and data for approximately the same period. Conclusions and Implications Although the results of this study emphasize the importance of productivity, the role of capital in the growth process should not be ignored: investment makes the application of new, more efficient technology possible, and the creation of this new technology through R 8 D efforts is not costless either. For instance, more efficient use of wood fibre input required large investments (and disinvestments) to change pulping methods from sulphite to sulphate. Also, 100 although nothing conclusive could be said about the nature and degree of economies of scale in this study, it is known that some of the technological change is scale-related. A growth characteristic in both countries is the rapid increase in the capital intensity of production, with increase in the K/L—ratio over three times faster in Finland than in the U.S. since the early 1970's. This suggests that capital intensity of production is one of the major factors explaining the differences in the role of productivity. Examination of numerical data and Figures 12, 13, and 17. showed that increases in TFP and capital-intensity are related. This is, naturally, only inferential, descriptive analysis. The relationship between capital intensity and labor and total factor productivity in the Finnish forest industries found in this study is supported econometrically by Simula (1979, 1983). Compared to the Finnish forest industries the United States forest industries have underinvested during the 1970's and early 1980's. It appears that if the U.S. forest industries are to increase their productivity, investment rates should be increased. Where investments are most likely to provide the greatest benefits is discussed later on. The type of complementary role of capital and productivity in contributing to growth discussed above is likely to apply to other inputs, too. For instance, application of a new, more efficient pulping process may 101 require increased energy intensity and new technical skills, which again require new investments. The problem lies in establishing these relationships in a quantitative form. It must be noted that in the late 1960's and early 1970's K/L -ratios were increased by significant investment for environmental protection. During those years TFP measures are to a small extent affected negatively by these investments. Increased capital intensity of production may have been important in raising productivity but it has also created problems. Increased amounts of fixed capital together with limited substitution possibilities make forest industries more vulnerable to fluctuations in demand for forestry products. This is a very serious problem for the Finnish P 8 P industry because of its high capital intensity and almost complete dependency on export markets, and also for the U.S. mechanical forest industries that suffer from fluctuations in housing starts. Examination of the changes in the levels of output and productivity revealed clearly their close relationship. The adverse effects of decreases in demand were strongly demonstrated in 1975 when there was almost a thirty percent drop in Finnish P 8 P production, and an almost ten percent decline in total factor productivity. The differences in output increases after 1975 in the Finnish and U.S. forest industries provide a very plausible explanation of the 102 differences in productivity development between 1975 and 1982. During those years both output and productivity in the mechanical and P 8 P industries grew significantly faster in Finland than in the U.S. The close relationship between TFP and changes in output has been shown in other studies (Simula 1979, 1983, Berndt 8 Watkins 1981, Martinello 1985). Changes in the demand and limited substitution possibilities also explain the negative efficiencies of labor and capital during several years shown by the factor augmenting CES production function. Negative labor efficiency is due to labor's quasi-fixed nature: it has not been possible to reduce labor input as much as needed when output has contracted. As a result labor productivity has decreased, and simultaneously wages have risen. Negative capital efficiency is due to capital's fixity and increased capital intensity in production. Increased capital intensity means that less labor is needed relative to capital to produce one unit of output, which has important implications for the demand for labor. The K/L-ratio can have changed either through the substitution process or through labor-saving technological change. According to this study, substitution possibilities are small, so increases in K/L ratios must be due to biased technological change that may have been induced by relatively greater changes in wages than in the price of capital. In Chapter V in connection with a factor augmenting CES 103 function, technological change was shown to be nonneutral. Additional analysis following Sato (1970) revealed technological change to be labor saving and capital using (according to nick's definition). Similar results have been reported by Stier (1980a) and Greber and White (1982). If this development continues in the future, not even significant increases in output will increase the demand for labor. Material and energy using/saving technological change was not tested econometrically in this study, and in the mechanical forest industries intermediate inputs were totally excluded from the analysis because of data problems. However, some conclusions concerning the role of material and energy input in the Finnish and U.S. P 8 P industry can be made. In the Finnish P 8 P industries, increased material productivity was an important source of output and TFP growth during the last years of the study period. If the growth of output and TFP is to continue, attention must be paid to maintaining favorable material productivity development because of the wood input's cost importance and limited possibilities for increasing its use. This is going to mean increased use of chemicals and changes in the product mix towards products where wood costs are relatively less important. It is also possible that substitution possibilities between labor and wood exist: e.g., allocation of additional labor to improving maintenance of existing 104 equipment could increase output, at least in the mechanical forest industries. The U.S. P 8 P industry hasn't been able to adapt smoothly to increases in relative timber prices. The use of material inputs, mainly pulpwood, during the study period increased faster than the use of any other input, and material productivity stayed practically constant over the study period. This development offers a partial explanation of the slowdown in TFP in the U.S. P 8 P industries after the mid-70's. Inability to respond to changes in relative factor input prices through substitution or through (material- saving) technological change means higher production costs and reduced profits and investment, which finally may result in stagnation. Slow material productivity development and large material cost share imply that to increase the rate of productivity, more should be invested on increasing material productivity. 7 Material input's large value share in the gross output suggests a need for a more detailed look at the components of material input and their productivity development. That was not possible in this study because of the difficulties in getting reliable, disaggregated annual price and quantity data. In both countries energy productivity increased only slightly over the 1958-74 period, with most of the increase taking place after 1975 as a response to sharp increases in 105 energy prices. In Finland, the high growth rates of output even with increases in energy productivity have meant increased 'use of energy. Because the forest industry is already now a major consumer of energy, the economic scarcity (rising prices) of energy may become a factor limiting new investments and growth. In the preceding discussion some explanations for changes in TFP and productivity differentials between the Finnish and U.S. forest industries were offered. These explanations can also be understood as suggestions where. research and data improvement efforts should be directed to learn more about the factors affecting growth. In this study an attempt was made to account for changes in TFP in the pulp and. paper industries. by adjusting labor and. capital for quality changes, but the amount of unexplained output growth was not reduced noticeably (Chapter VI). The "explanation" (reduction) of the residual would seem to require better data on changes in the quality of inputs (e.g. , vintage data on capital), or possibly a different analytical approach. It is also possible that a significant part of the reported TFP change has been disembodied, and/or some central variables have been left out from the analysis, e.g., economies of scale, R 8 D expenditures and management input. It is also evident that a part of TFP change is "explainable" by errors in data and by market imperfections. 106 Fluctuations in demand for forest products combined with the high capital intensity of production emphasize the importance of good management, especially in investment planning. Management is also central in effective organization planning and direction of R 8 D efforts. Although the role of management in increasing productivity is recognized, management as an input has not really been quantitatively studied, mainly because of the concept's complexity and intangibility. Since the efficiency with which an industry converts resources into outputs affects the level of‘ profits and subsequent capital investment, total factor productivity is an important determinant of an industry's performance and eventually of it's rise or decline. In Finland, the forest industry's performance has great repercussions on social well-being because of its central importance in the economy. To maintain the relatively favorable productivity development, more attention must be paid to supporting research, education and labor training, and to creating a positive environment for investments. At the firm level, attention should be paid to investment planningandto management of technology diffusion. The above recommendations are quite general. More concrete recommendations would require a separate study where factors affecting total factor productivity development would be examined in detail. Because of the nature of the concept, 107 total factor productivity studies can provide only little guidance for policy formulation unless causal, quantitative relationships between policy instruments» and. productivity development are established. The problem in explaining total factor productivity is that behind productivity lie all the dynamic forces of economic life, and it is very difficult to account for these factors. Nevertheless, the linking of quantitative knowledge of production relationships to policy instruments and objectives should be attempted. APPENDICES 108 APPENDIX A. Related Research. Empirical research on the sources of economic growth and technological change at national level is extensive. Studies on the forest industry and its subindustries are, however, less common. In Finland, Simula (1979, 1983) has studied productivity in the forest industries using production functions and total factor productivity (TFP) indices. His first study examined factors affecting productivity in individual branches using time-series data and several alternative models and ways of measuring inputs and outputs. In the latter study, cross-section data were used in analyzing the suitability of TFP indices and alternative production functions for explaining productivity differences in individual branches and at the aggregate level. The study revealed that in most cases productivity variation is best explained by differences in the capital-labor ratio and plant size. The study also demonstrated the importance of output quality and the level of output for total productivity. Katila (1983) has analyzed the role and nature of technological change in the growth of output in the Finnish pulp and paper (P 8 P) industry. The study attributed most of the growth to the increased use of capital, and the rate of technological change was found to be quite slow. Wyatt (1983) 109 APPENDIX A. (cont'd.). compared sectoral total factor productivity growth in the Finnish and Swedish industries using a TFP index. The study demonstrated that total factor productivity growth in the Finnish forest industries has been slow and below the corresponding growth rates in Sweden. The reliability of wyatt's conclusion's can, however, be questioned because of problems with the Finnish real material data. Karhu 8 vainionmaki (1985) used the same data in their work on TFP measurement in the Finnish economy. In the United States, Robinson (1975) has applied Solow's (1957) residual model to the U.S. mechanical wood industry. The study indicated that increased use of capital had a greater impact on labor productivity than technological change. Manning and Thornburn (1971) applied the same method to the Canadian P 8 P industry but with opposite results. Risbrudt (1979) has examined the rate of technological change in the U.S. forest industry using five alternative models. He attempted also to analyze qualitatively the factors underlying technological change. The validity of the assumptions underlying the alternative models was not assessed in the study. Stier (1980a, 1980b) has applied dual cost functions in studying the structure and role of technological change in the U.S. forest industry. Greber and White (1982) applied 110 APPENDIX A. (cont'd.). Sato's (1970) method. to study' the role of technological change in the U.S. lumber and wood products industry. In their study, most of the growth was attributed to technological change. Martinello (1985) has estimated the rate of technological change and returns to scale in the Canadian forest industries, using cost functions. Bengston and Strees (1986) have emphasized the need to. analyze technological change in a gross output framework. They found out that the use of value added output measures overestimated considerably the rate of technological change in the U.S. lumber and wood products industry. Two Swedish studies must also be acknowledged. Wohlin (1970) applied Salter's (1960) approach. in. his study on structural change, expansion possibilities and technological change in the. Swedish forest industries. Forsund. et al. (1978) have analyzed technological and structural change in the Swedish pulp industry using plant-level cross-section data and frontier production functions. 111 APPENDIX B. Derivation of Aberg's Model. The starting point is a CD-production function with constant returns to scale: Q = ALaK3c°e9t (1) where Q is the real value added, a and 8 stand for output elasticities, c is the rate of capacity utilization, 0 is the elasticity with respect to c, and g represents the rate of technological change. It is assumed that firms have a constant real return requirement (r) on capital, and they try to equate the marginal product of capital with this rate (fx 2 r). If the "true" rate of return is assumed to be proportional to the rate of capacity utilization we can write Q = Kcr + Lw (2) where Kc represents the utilized capital stock. This equation can be rewritten R = Kcr = Q - Lw (3) where R is the real (capital income. From this, c can be reformulated as c = R/Kr. In empirical analysis the ratio R/K can thus be used as a measure of the utilization rate. Equation 1 can then be expressed by Q = B afiLaegt (4) where B is equal to Ar'fl (for simplicity's sake 8 = 8). Given constant returns to scale Equation 3 can be written in labor intensive form ln(Q/L) = lnB + filn(R/L) + gt (5) 112 APPENDIX B. (cont'd.). In order to be able to include r in the constant factor B, the assumption of the constant real return requirement on capital is crucial. However, it doesn't necessarily have to be fully constant, but it is enough if there's no clear trend in r during the period in question, so that the technological change term 9 is not affected. Another requirement is that the share of capital income (R/Q) is not constant during the period of estimation. 113 APPENDIX C. Data for the Finnish Forest Industries. Table C1. Data for the Finnish Mechanical Forest Industry. Finland Year OVA K L w r1 r2 8 8 1958 .5446 .5968 .9789 .5819 .8257 .8492 .6360 .3640 59 .6061 .6313 1.0057 .6079 .7024 .9473 .6758 .3242 60 .7618 .6457 1.1851 .6223 1.1469 1.2367 .6001 .3999 61 .7508 .6778 1.1785 .6656 .9446 1.0328 .6494 .3506 62 .7032 .7127 1.1023 .7145 .8415 .8077 .7732 .2268 63 .7422 .7379 1.0966 .7252 .5226 .8971 .7571 .2429 64 .7872 .7803 1.0869 .7397 .5372 .9763 .7435 .2565 65 .8173 .8058 1.0702 .7785 .5096 .9846 .7541 .2459 66 .7399 .8181 .9808 .8247 .3322 .7769 .8181 .1819 67 .7629 .8332 .9082 .8692 .3877 .8675 .7870 .2130 68 .7900 .8311 .9290 .8758 .8091 .9075 .6466 .3534 69 .9148 .9025 1.0030 .9348 1.0942 .9755 .5894 .4106 70 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 .6019 .3981 71 .9974 1.0320 .9661 1.1094 .6874 .8574 .6955 .3045 72 1.0047 1.0595 .9330 1.1907 .7424 .7967 .6810 .3190 73 1.0833 1.1338 .9544 1.2644 1.6913 .7908 .4875 .5125 74 .9614 1.2833 .9142 1.3830 1.4371 .3923 .5089 .4911 75 .6567 1.5583 .7612 1.3905 .0006 .0316 .9994 .0006 76 .7482 1.8083 .7800 1.4994 .2671 .0615 .7855 .2145 77 .7966 1.9577 .7790 1.5043 .4187 .1171 .6837 .3163 78 .8712 2.1519 .7987 1.5204 .4476 .1639 .6558 .3442 79 1.0393 2.2516 .8915 1.5614 .6986 .2248 .5723 .4277 80 1.1337 2.4129 .9304 1.6092 .9069 .2422 .5084 .4916 81 .9854 2.6252 .8344 1.6323 .3771 .1585 .6753 .3247 82 .8933 2.7501 .7219 1.6518 1775 .1604 .7869 .2131 Avg. .6829 .3171 Note: All monetary variables are in constant prices. Base year is 1970. OVAavalue added, Katotal capital, Latotal labor hours, w-wage rate, r1, r2 are alternative real rates of return, aslabor'e value added income share, fiscapital's value added income share. Construction of time series is explained in Chapter IV. APPENDIX C. (cont'd.). 114 Table C2. Data for the Finnish P 8 P Industry. Finland Year 06V OVA M E K L w r1 r2 8 8 1958 .4065 .4065 .4674 .4200 .4083 .8185 .5463 .9282 .9358 .4814 .5186 59 .4315 .4315 .5002 .4440 .4478 .8546 .6088 .7636 .8441 .4518 .5482 60 .5040 .5040 .5605 .5149 .4918 .9321 .5971 .8403 .9604 .4594 .5406 61 .5937. .5937 .6626 .6050 .5721 1.0139 .6345 .8385 .9854 .4676 .5324 62 .6164 .6164 .6709 .6366 .6550 1.0214 .6721 .5514 8766 .4770 .5230 63 .6729 .6729 .7072 .6796 .6865 1.0170 .6928 .6015 .9524 .4687 .5313 64 .7431 .7431 .7490 .7716 .7220 1.0340 .7213 .5650 0269 .4581 .5419 65 .7980 .7980 .8236 .8142 .7857 1.0382 .7590 .4517 1 0234 .4512 .5488 66 .8360 .8360 .8453 .8371 .8572 1.0161 .8034 .3546 .9891 ;4463 .5537 67 .8173 .8173 .8270 .8406 .9010 .9752 .8374 .3800 .9075 .4721 .5279 68 .8693 .8693 .8526 .8828 .8922 .9807 .8594 .7723 .9923 .4812 .5188 69 .9616 .9616 .9296 1.0052 .9735 1.0051 .9103 .8998 0167 .4866 .5134 70 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 0000 .5017 .4983 71 .9902 .9902 1.0827 1.0851 1.0532 1.0441 1.0802 .5941 .8614 .5343 .4657 72 1.1069 1.1069 1.1200 1.0842 1.1517 1.0121 1.1807 .5402 .9151 .5174 .4826 73 1.1802 1.1802 1.1991 1.0701 1.1800 1.0167 1.2821 .6225 .9372 .4825 .5175 74 1.2009 1.2009 1.2937 1.0491 1.2034 1.0545 1.3332 1.0662 .8954 .4031 .5969 75 .8966 .8966 .8264 .9821 1.3557 1.0237 1.4514 .5229 .3995 .4334 .5666 76 .9393 .9393 .8918 .9963 1.5657 1.0062 1.4650 .1627 .3942 .4362 .5638 77 .9287 .9287 .8536 .9979 1.6896 .9392 1.4674 .1788 3894 .4668 .5332 78 1.0572 1.0572 .9879 1.1095 1.7774 .9158 1.4701 .3775 4968 .4682 .5318 79 1.2060 1.2060 1.0338 1.1684 1.8063 .9289 1.5603 .5798 5865 .4453 .5547 80 1.2376 1.2376 1.0683 1.1709 1.8658 .9474 1.6506 .5371 5580 .4607 .5393 81 1.2554 1.2554 1.0216 1.1586 1.9653 .9310 1.6724 .4740 5463 .4534 .5466 82 1.2230 1.2230 .8877 1.0838 2.0765 .8696 1.7003 .3193 5148 .4759 .5241 Avg. .4672 5328 Note: All monetary variables are in constant prices. Base year is 1970. Gavegross value of production, OVAsvalue added, Hsmaterial input, E=ener9y, Kstotal capital, L=total labor hours, =wage rate, r1, r2 are alternative rates of return, calabor's value added income share, flacapital's value added income share. Construction of time series is explained in Chapter 1V. APPENDIX D. Table D1. 115 Data for the U.S. Forest Industries. Data for the U.S. Mechanical Forest Industry. Year 1958 OVA .7881 .8653 .8319 .8251 .8792 .9218 .9660 .9740 .9656 .9726 1.0167 1.0146 1.0000 1.0869 1.3242 1.3132 1.1668 1.1175 1.2543 1.3554 1.3796 1.3763 1.2933 1.2576 K .7994 .8221 .8517 .8619 .8607 .8755 .8962 .9107 .9182 .9379 .9385 .9589 1.0000 1.0138 1.0512 1.0789 1.0965 1.1327 1.1897 1.2229 1.2494 11.2873 1.3229 1.3369 1.3607 L 1.1222 1.1999 1.1554 1.0627 1.0750 1.0898 1.1060 1.1160 1.1067 1.0851 1.0460 1.0571 1.0000 1.0076 1.1116 1.1777 1.0956 .9695 1.0562 1.1424 1.1775 1.2319 1.1127 1.0438 .8249 w .7370 .7748 .7650 .7879 .8038 .8520 .9067 .8934 .8979 .9051 .9883 1.0023 1.0000 1.0499 1.1257 1.1095 1.0739 1.0948 1.1319 1.1755 ’1.2140 1.1579 1.0983 1.0695 1.1635 r1 .9370 1.2443 .8715 .8522 .9634 1.0955 1.1478 1.0681 1.1427 1.1224 1.5866 1.5211 1.0000 1.4129 1.9040 2.3301 1.5287 1.0777 1.5896 1.7897 2.0112 1.6948 1.0244 r2 .8913 .9004 .8583 .9297 1.0556 1.0383 .9985 1.0213 .9921 1.0171 1.0482 .9670 1.0000 1.1277 1.3944 1.2290 1.0469 1.0829 1.1503 1.1282 1.0269 .9935 1.0820 1.1457 1.0241 Note: All monetary variables are in constant prices. Base year is 1970. OVA-value added, Kstotal capital, L-total labor hours, w-wage rate, r1, r2 are alternative real rates of return, e-labor's value added income share, pucapital's value added income share. Construction of time series is explained in Chapter IV. APPENDIX D. Table D2. Year 1958 OGV .6044 .6719 .6765 .6911 .7247 .7520 .7879 .8333 .9013 .9043 .9719 1.0249 1.0000 .9951 1.0641 1.1668 1.2201 1.0057 1.1302 1.1579 1.1980 1.2521 1.2451 1.2681 1.2201 OVA .5650 .6300 .6440 .6640 .7000 .7370 .7840 .8300 .8940 .8920 .9590 1.0090 1.0000 1.0110 1.0900 1.1410 1.1360 .9650 1.1230 1.1750 1.2030 1.2460 1.2370 1.2030 1.2120 I .6400 .7031 .6994 .7060 .7312 .7494 .7847 .8296 .8968 .8994 .9684 1.0231 1.0000 .9969 1.0679 1.1660 1.2139 .9962 1.1146 1.1369 1.1836 1.2446 1.2453 1.2759 1.2351 (cont'd.). E .6290 .6925 .6904 .6984 .7250 .7448 .7722 .8082 .8650 .8586 .9392 1.0077 1.0000 1.0119 1.1298 1.1978 1.2098 .9618 1.0081 1.0180 1.0153 1.0102 .9900 .9800 .8970 K .6132 .6464 .6687 .6960 .7175 .7349 .7566 .7964 .8427 .9021 .9448 .9713 1.0000 1.0242 1.0450 1.0563 1.0603 1.0898 1.1894 1.2497 1.2904 1.3375 1.4076 1.4712 1.5847 116 L .9430 .9902 .9811 .9512 .9618 .9591 .9679 .9775 1.0036 1.0174 1.0286 1.0451 1.0000 .9636 .9564 .9596 .9574 .8570 .9050 .9104 .9032 .9150 .9098 .8927 .8428 H .7542 .7695 .7882 .8095 .8393 .8654 .8859 .8995 .9246 .9449 .9618 .9764 1.0000 1.0349 1.0953 1.1137 1.1097 1.1467 1.2056 1.2616 1.2834 1.2720 1.2444 1.2367 1.2707 r1 1.2584 1.4406 1.3705 1.2685 1.2422 1.2892 1.3503 1.3520 1.3756 1.2001 1.1368 1.1160 1.0000 .8548 .9417 1.0930 1.4944 1.1872 1.1939 1.0572 1.0275 1.0914 .9484. Data for the U.S. P 8 P Industry. r2 .6813 .7691 .7684 .8008 .8251 .8754 .9385 .9799 1.0203 .9114 .9828 1.0270 1.0000 1.0006 1.0839 1.1491 1.1413 .8691 .9713 .9615 .9665 .9935 .9538 .8855 Note: All monetary variables are in constant prices. Base year is 1970. OGngross value of production, OVAsvalue added, Ismaterial input, Esenergy, Katotal capital, Lstotal labor hours, wswage rate, r1, r2 are alternative rates of return, ealabor's value added income share, plcapital's value added income share. Construction of time series is explained in Chapter 1V. 117 APPENDIX E. Derivation of the Real Material Input Series For the 11.5: 2 8 2 industm detailed annual price and quantity data on materials consumed were not available, so data from census years were used in constructing the real material input index. Different items comprising the aggregate cost of materials, supplies, etc. reported in the Census of Manufactures were deflated respectively by their wholesale/producer price indices. These real costs were used in computing material input-output ratios for the census years. The real material input for the intermediate years was obtained by estimating input-output ratios for these years by linear interpolation, and then multiplying the corresponding real outputs by these ratios. To improve the interpolation procedure an attempt was made to relate changes in input- output ratios to changes in capacity utilization rates, but no clear relationship was found using the data from the census years. In the following the cost items and their deflators are presented. The correctness of some of the deflators can be questioned, but the best existing deflator was always chosen. In some cases a deflator was constructed combining several price series with "consumption" or cost shares as weights. . W: Deflated by a product's price index. (Resales should have been excluded from both the gross output 118 APPENDIX E. (cont'd.). and material input, but it was not possible, because of lack of data). ggn;;agt__flg;k: Deflated by an index with following weights: 0.5*(wage index in metal industries) + 0.5*(PPI for machinery and equipment). Materials_and_§unnlies: - Different pulpwood species and chips were deflated using price indices from U.S. Timber Production, Trade.... 1984. Southern pine and mixed hardwoods were deflated by an average of Midsouth, Southeastern 8 Louisiana pine and hardwood pulpwood prices, respectively. Item "Other hardwoods" was deflated by the average of these two price indices. Cost of stumpage cut was deflated by an index with following weights: 0.7*(Southern pine) + 0.1*(Southern hardwoods) + 0.1*(Wisconsin spruce) + 0.1*(Wisconsin hemlock). - Chemicals: Deflated by PPI for industrial chemicals. - Woodpulp: Deflated by PPI for woodpulp. — Wastepaper: Deflated by a price index from Bureau of Labor Statistics (BLS). - Paperboard boxes: Deflated by a price index from BLS. - All other materials and supplies were deflated by PPI for intermediate materials and supplies. 119 APPENDIX E. (cont'd.). For the W the real material input was constructed. using annual data reported. in. Industrial Statistics Parts I and II. Nominal material costs were deflated with various indices and procedures to make these costs represent physical material input. W: This was the largest cost item reported in Industrial Statistics Part I. To deflate it, it was necessary to divide this aggregate cost. group into three sub-groups, pulpwood, chemicals and others, using detailed value and quantity data reported in Industrial Statistics Part II. The prices of different pulpwood species, chips, etc. were calculated using the reported value and quantity data, but in the construction of the real wood input quantity data reported in Yearbook of Forest Statistics were used. The total real cost of chemicals was obtained by summing up all chemicals product by product, and deflating this aggregate by WPI for chemicals. The value of "others" was computed as a residual that remained after the value of pulpwood, chemicals and own pulp was deducted frmm the total cost. The resulting residual was deflated by WPI. The deduction of own pulp was necessary to avoid (reduce) double counting of“ material costs. These three subaggregates were then added together to get the real raw 120 APPENDIX E. (cont'd.). material and semifinished products. This procedure was carried out for years 1958-74: from 1975 to 1982 price index for :gw_m§;§11§1§ provided by Indufor Ky, Helsinki was used for deflating. This index is based on detailed data and computations, and can be considered to be a reliable deflator for raw materials. The use of two different deflating procedures causes error in the material input index, but the difference between these two ways of deflating is not great because wood raw material was the major cost item in this cost aggregate, and wholesale prices for all products and chemicals developed similarly enough. with this "correct" index between 1975 and 1982. The effect on TFP measures is probably small except during 1975 when the change in the deflator takes place. The material input and TFP measures for this year are also of suspect because of the change in the correction procedure for double counting. Eggkgging__mgt;: Deflated by WPI for paper and paperboard. A Lubricants: Deflated by WPI for mineral fuels and lubricants. WW: Deflated by WPI for machinery and equipment. 121 APPENDIX E. (cont'd.). antrggt__wgzk: Deflated. by an index ‘with following weights: 0.4*(wage index in P 8 P industry) + 0.6*(WPI for machinery and equipment). Repairs: Deflated by an index with following weights: O.5*(wage index in metal industries) + 0.5*(WPI for machinery and equipment). 1 22 APPENDIX F. Output and Partial Productivity Changes for the Finnish and U.S. Forest Industries Table F1. Output and Partial Productivity Changes in the Mechanical Forest Industries. Mechanical Forest Industries Finland Year Q Q/L 1959-82 0.0206 0.0333 1959-70 0.0506 0.0489 1971-82 -0.0094 0.0178 1976-82 0.0440 0.0515 (percent) Q/K -o.04 30 0.0076 -0.09 -0.03 37 72 - Q 0.0142 0.0198 0.0085 0.0013 Q/L 0.0270 0.0295 0.0245 0.0218 Q/K -0.0080 0.0012 -0.0172 -0.0275 Table F2. Output and Partial Productivity Changes in the P 8 P Industries. Pulp and Paper Industries Year Q 1959-82 0.0459 1959-70 0.0750 1971-82 0.0168 1976-82 0.0444 Year Q 1959-82 0.0293 1959-70 0.0420 1971-82 0.0166 1976-82 0.0277 Finland (percent) Q/L Q/K 0.0434 -0.0219 0.0583 0.0004 0.0284 -0.0441 0.0677 -0.0166 USA (percent) Q/L Q/K 0.0339 -0.0103 0.0371 0.0012 0.0308 -0.0218 0.0300 -0.0259 Q/M 0.0192 0.0116 0.0267 0.0341 Q/M 0.0019 0.0048 0.0010 0.0031 Q/E 0.0064 0.0027 0.0101 0.0303 Q/E 0.0145 0.0033 0.0256 0.0376 123 APPENDIX G. Estimated C-D Production Functions in the Finnish and U.S. Mechanical Forest Industries Finnish Mechanical Forest Industry, 1959-82 dln(Q) = 0.0679 + 1.4307 dln(L) - 0.4573 dln(K)1 (3.6331) (9.7611) (-1.8495) fi2 = 0.8315 F = 57.7500 D-W = 2.1444 dln(Q/L) - 0.0668 - 0.4386 dln(K/L) (5.0462) (-4.0753) 52 a 0.4043 F = 16.6078 D-W = 2.1304 dln(Q/K) - 0.0668 + 1.4386 dln(L/K) (5.0480) (13.3674) i? = 0.8854 F = 178.6880 D-w = 2.1296 U.S. Mechanical Forest Industry, 1959-82 dln(Q) = 0.0264 + 0.7638 dln(L) - 0.1088 dln(K) (1.3193) (6.1112) (-0.1389) fi2 = 0.6186 F = 19.6527 o-w = 1.8043 dln(Q/L) = 0.0186 + 0.2412 dln(K/L) (1.8817) (1.9729) i? = 0.1117 F = 3.8922 D-W = 1.7869 dln(Q/K) = 0.0186 + 0.7590 dln(L/K) (1.8831) (6.2092) 52 = 0.6202 F = 38.5544 o-w 1.7864 1T-values are in the parentheses. 124 APPENDIX H. Estimated Modified C-D Functions in the Finnish and U.S. Forest Industries Finnish Mechanical Forest Industry, 1959-82 dln(Q/L) = 0.0325 + 0.1121 dln(s)1 (5.6787) (10.2324) P? = 0.8185 F - 104.703 o-w = 2.0212 dln(Q/L) = 0.0428 - 0.02050 + 0.1103 dln(s) (5.5661) (-1.8840) (10.5907) NZ 3 0.8373 F 8 60.1929 D-W = 2.2857 U.S. Mechanical Forest Industry, 1959-82 dln(Q/L) = 0.0120 + 0.3686 dln(s) (2.0202) (6.7643) i2 = 0.6605 F - 45.7558 o-w = 1.6261 dln(Q/L) = 0.0155 - 0.00710 + 0.3696 dln(s) (1.8975) (-0.6343) (6.6874) §2 = 0.6511 F = 22.4576 D-W = 1.6924 Finnish Pulp and Paper Industry, 1959-82 dln(Q/L) = 0.0254 + 0.4441 dln(s) (7.4905) (23.1761) i2 - 0.9589 F - 537.130 o-w - 1.8967 dln(Q/L) = 0.0304 - 0.00958 + 0.4404 dln(s) (6.4407) (-1.4663) (23.3630) i2 = 0.9609 F = 238.680 D-W = 1.9902 1T-values are in the parentheses. 125 APPENDIX H. (cont'd.). U.S. Pulp and Paper Industry, 1959-82 dln(Q/L)* = 0.0099 + 0.5175 dln(s) (3.0077) (19.5133) E? = 0.9452 F =380.771 0-w = 1.5824 dln(Q/L)* = 0.0088 + 0.00130 + 0.5218 dln(s) (2.1216) (0.2509) (18.3244) E? = 0.9385 F = 168.800 0-w = 1.5298 »* Hildreth-Lu correction for autocorrelation 126 APPENDIX I. Estimated CES Functions for the Finnish and U.S. Mechanical Forest Industries. Finnish Mechanical Forest Industry, 1959-82 9/w = 0.0441 + 0.0082 (x/x)1 (5.6169) (0.1278) -0.0447 0.0163 1.6441 0 = 0.387/0.0082 = 47.1951, not sign., wrong sign f/r* = 0.6130 + 8.7107 (x/x) (1.8130) (2.9088) 0 = 0.6829/8.7107 = 0.0780, significant f/r = 0.2625 + 4.3470 (x/x) (2.3467) (4.7802) 0 = 0.6829/4.3470 = 0.1571, significant U.S. Mechanical Forest Industry, 1959-82 0/w = 0.0206 + 0.0447 (i/x) (2.2550) (0.3952) 0 = 0.3794/0.0447 = 8.4877, wrong sign i/r = 0.0783 + 2.9538 (x/x) (1.9473) (5.9321) 0 = 0.6206/2.9538 = 0.2101, significant *Hildreth-Lu correction for autocorrelation 1T-values are in the parentheses. §2 F D-W §2 F D-W 0.2533 8.4612 2.2421 0.4872 22.8503 1.8010 -0.0381 0.1562 1.6523 0.5978 35.1898 2.1849 127 APPENDIX J. Rates of Change in Factor Efficiencies and Indices of Technological Change for the Mechanical Forest Industries. Table J1. Rates of Change in Factor Efficiencies and Indices of Technological Change for the Finnish Mechanical Forest Industries. fiNLAID Year b1 b2 b3 81 a2 a3 VAPi VAPZ VAP3 1958 1.0000 1.0000 1.0000 59 .0913 .1703 .3694 .0868 1002 .1341 1.0883 1.1243 1.2150 60 .1520 .0462 -.2202 .0723 .0876 .1262 1.1894 1.1970 1.2161 61 -.0381 .0106 .1334 -.0235 -.0519 -.1233 1.1604 1.1685 1.1890 62 -.0129 .1882 .6948 -.0118 -.0377 -.1029 1.1483 1.1954 1.3143 63 .0107 -.0060 -.0481 .0676 .0841 .1257 1.2026 1.2584 1.3992 64 -.0017 -.0109 -.0341 .0768 .0946 .1396 1.2598 1.3267 1.4954 65 .0165 .0382 .0929 .0534 .0541 .0559 1.3039 1.3768 1.5605 66 -.0551 .0614 .3549 -.0255 -.0515 -.1171 1.2721 1.3493 1.5439 67 -.o148 -.0677 -.2010 .1180 .1384 .1900 1.3639 1.4471 1.6568 68 -.0955 -.3552 -1.0098 .0131 .0148 .0191 1.3466 1.3584 1.3882 69 .0191 -.0693 -.2919 .0711 .0729 .0776 1.3978 1.3770 1.3247 70 .0010 .0294 .1010 .0967 .1058 .1289 1.4558 1.4519 1.4423 71 .0308 .1575 .4770 .0181 -.0087 -.0762 1.4783 1.5014 1.5599 72 -.0373 -.0729 -.1629 .0368 .0261 -.0007 1.4920 1.4966 1.5087 73 -.1478 -.4513 -1.2161 .0512 .0484 .0415 1.4609 1.3384 1.0302 74 -.2585 -.2884 -.3637 -.1o79 -.1696 -.3252 1.2774 1.1092 .6857 75 .7833 3.4370 10.1241 -.2368 -.3125 -.5033 1.1339 1.1395 1.1540 76 -1.1695 3.4178 -9.0837 .1119 .1234 .1522 1.1954 1.1235 .9427 77 -.1055 -.2790 -.7162 .0756 .0981 .1551 1.2233 1.1223 .8684 78 -.0187 -.0454 -.1127 .0749 .0950 .1456 1.2673 1.1709 .9287 79 .0713 -.0455 -.3399 .0740 .0888 1261 1.3402 1.2079 .8752 80 -.0285 -.1190 -.3470 .0470 .0522 .0655 1.3525 1.1814 .7511 . 81 -.1001 .1429 .7552 -.0401 -.os71 -.1000 1.2880 1.2056 .9986 82 -.0287 .1977 .7683 .0534 .0664 .0992 1.3196 1.3070 1.2760 Avg. -.0390 -.0312 -.0115 .0314 .0276 .0181 Note: Numbers 1, 2, and 3 refer to alternative elasticities of substitution (0.16, 0.36, 0.60, respectively). Symbols 8 and b represent labor and capital efficiencies. VAP's represent value added productivity indices. 128 APPENDIX J. (cont'd.). Table J2. Rates of Change in Factor Efficiencies and Indices of Technological Change for the U.S. Mechanical Forest Industries. 034 Year b1 b2 b3 61 .2 a3 VAP1 VAPZ VAP3 1958 1.0000 1.0000 1.0000 59 .0075 -.os60 -.2613 .0202 .0101 -.0089 1.0159 .9775 .9053 60 .0000 .1207 .3471 .0015 .0063 .0153 1.0169 1.0217 1.0306 61 -.0195 - 0186 -.0168 .0876 .1073 1442 1.0716 1.0904 1.1255 62 .0497 .0249 -.0215 0606 .0744 .1003 1.1287 1.1489 1.1866 63 .0041 -.0380 -.1170 .0272 .0167 -.0031 1.1481 1.1470 1.1449 64 .0174 .0076 -.0108 .0241 .0112 -.0131 1.1699 1.1569 1.1326 65 .0096 .0373 .0893 0033 .0094 .0208 1.1753 1.1758 1.1767 66 -.0398 -.0763 -.1449 -.0021 -.0045 -.0090 1.1602 1.1466 1.1210 67 -.0129 -.o112 -.0079 .0319 .0400 .0553 1.1762 1.1685 1.1539 68 -.0366 -.1663 -.4097 .0793 .0764 .0709 1.2102 1.1501 1.0371 69 -.0188 - 0108 .0041 -.0198 -.0313 -.0529 1.1908 1.1274 1.0081 70 .0401 1959 .4881 .0526 .0713 .1062 1.2387 1.2456 1.2583 71 -.0037 -.1221 -.3443 .0830 .0946 .1165 1.2892 1.2589 1.2019 72 .1247 .0659 -.0446 .1072 .1199 .1437 1.4039 1.3555 1.2644 73 -.0971 - 1984 -.3885 -.0798 -.1020 -.1435 1.3159 1.2078 1.0046 74 -.0581 .0650 .2959 -.0494 -.0551 -.0659 1.2625 1.2080 1.1054 75 -.0028 .1148 .3354 .0951 .1208 .1691 1.3187 1.3264 1.3406 76 -.0193 -.1576 -.4170 .0289 .0274 .0245 1.3278 1.2779 1.1840 77 .0317 0023 -.0531 - 0113 -.0280 -.0592 1.3360 1.2636 1.1276 78 -.0357 -.0873 -.1842 -.0244 -.0436 -.0796 1.3064 1.1997 .9993 79 .0046 .0642 .1761 - 0477 -.0478 -.0481 1.2827 1.2032 1.0539 80 .0204 .1979 .5309 .0642 .1038 .1782 1.3293 1.3448 1.3738 81 .0186 .1108 .2839 .0525 .0793 .1297 1.3701 1.4350 1.5569 82 -.0667 0594 .2961 .1144 .1246 .1438 1.4301 1.5400 1.7465 Avg. -.0034 0039 .0177 .0291 .0326 .0390 Note: Numbers 1, 2, and 3 refer to alternative elasticities of substitution (0.21, 0.41, 0.60, respectively). Symbols a and b represent efficiencies of labor and capital. VAP's represent value added productivity indices. 129 APPENDIX K. Estimated CD Functions for the Finnish ' and U.S. Pulp and Paper Industries. Finnish Pulp and Paper Industry, 1959-82 dln(Q) = 0.0621 + 1.0912 dln(L) - 0.2798 dln(K)1 (1.9060) (2.4786) (-0.6775) 82 = 0.1570 F = 3.1420 0-w = 1.8421 dln(Q/L) = 0.0559 - 0.1925 dln(K/L) (2.1298) (-0.6156) P? = -0.0277 F = 0.3790 0-w = 1.8276 dln(Q/K) = 0.0559 + 1.1923 dln(L/K) (2.1301) (3.8142) i2 = 0.3707 F = 14.5487 0-w = 1.8274 U.S. Pulp and Paper Industry, 1959-82 dln(Q) = 0.0315 + 1.4276 dln(L) + 0.1768 dln(K) (2.1032) (7.4031) (0.5246) §2 a 0.7239 F - 31.1485 0-w = 1.6900 dln(Q/L) = 0.0510 - 0.3289 dln(K/L) (4.7401) (-1.6972) 82 = 0.0756 F = 2.8806 0-w = 1.9635 dln(Q/K) = 0.0510 + 1.3286 dln(L/K) (4.7358) (6.8549) £2 = 0.6666 F = 46.9896 0-w = 1.9622 1T-values are in the parentheses. 130 APPENDIX L. Estimated CES Functions for the Finnish and U.S. Pulp and Paper Industries. Finnish Pulp and Paper Industry, 1959-82 . . R2 = -0.0374 w/w = 0.0508 + 0.0535 (x/x)1 F = 0.1705 (4.6694) (0.4129) 0—w = 1.7955 0 = 0.4832/0.0535 = 9.032, not sign., wrong sign . . §2 = 0.3561 r/r = 0.2766 + 4.9208 (x/x) F = 13.7174 (2.4789) (3.7037) 0-w = 1.7187 0 = 0.5168/4.9208 = 0.1050, significant U.S. Pulp and Paper Industry, 1959-82 . . fi2 = 0.0079 w/w* = 0.0170 - 0.0993 (x/x) F = 1.1747 (2.0595) (-1.0838) 0-w = 1.3662 0 = 0.5328/0.0933 = 5.7106, not sign. . . §2 = 0.6189 r/r = 0.1031 + 2.8191 (x/x) F = 38.3506 (4.0765) (6.1928) 0-w = 1.4050 0 = 0.4672/2.8191 = 0.1657, significant * Hildreth-Lu correction for autocorrelation 1T-values are in the parentheses. 131 APPENDIX M. Rates of Change in Factor Efficiencies, and Indices of Technological Change for the P 8 P Industries. Table M1. Rates of Change in Factor Efficiencies, and Indices of Technological Change for the Finnish P 8 P Industries. FINLAND Year b1 b2 b3 81 a2 a3 VAP1 VAPZ VAP3 1958 1.0000 1.0000 1.0000 59 -.0129 .0102 .0751 .0051 -.0079 -.0448 .9952 1.0216 1.2150 60 .0575 .0527 .0392 .0793 .0917 .1269 1.0625 1.1031 1.2163 61 .0143 .0164 .0223 .0820 .0847 .0923 1.1099 1.1623 1.1834 62 -.0580 -.0123 .1167 .0267 .0228 .0118 1.0915 1.2381 1.3317 63 .0349 .0283 .0096 .0997 .1085 .1333 1.1664 1.3296 1.4447 64 .0627 .0786 .1235 .0878 .0938 .1108 1.2547 1.4852 1.5837 65 .0126 .0425 .1266 .0693 .0717 .0782 1.3096 1.6339 1.6868 66 -.0156 .0131 .0940 .0694 .0710 .0754 1.3554 1.7693 1.6588 67 -.0900 -.1102 -.1671 .0157 .0124 .0032 1.3199 1.6556 1.8461 68 -.0073 - 0980 -.3536 .0598 .0641 .0762 1.3551 1.4309 1.3502 69 -.0035 -.0233 -.0791 .0786 .0813 .0888 1.3955 1.4150 1.2645 70 .0008 -.o124 -.0496 .0380 .0310 .0110 1.4164 1.3775 1.4132 71 -.0049 .0604 .2445 - 0691 -.0876 -.1398 1.3688 1.4774 1.5794 72 .0364 .0530 .0997 .1491 .1567 .1782 1.5003 1.6860 1.4985 73 .0272 .0128 -.0280 0568 .0535 .0444 1.5645 1.7030 .7815 74 -.0690 -.1459 -.3627 -.0263 -.0346 -.0579 1.4870 1.3224 .5123 75 -.3742 -.3313 -.2106 -.3056 ~.3550 -.4943 .9794 .8649 .7521 76 .0350 .1874 .6169 .0704 .0782 .1000 1.0370 1.0971 .5932 77 -.1105 -.1368 -.2111 .0645 .0725 .0949 1.0595 1.1191 .5491 78 -.0034 -.0983 -.3656 .1737 .1955 .2568 1.1767 1.1583 .5822 79 .0768 .0322 -.0934 .1245 .1328 .1559 1.2952 1.1948 .5511 80 .0022 .0122 .0405 - 0001 -.oo72 -.0274 1.2966 1.2050 .4827 81 -.0270 -.0146 .0202 .0341 .0368 .0443 1.3007 1.2437 .6022 82 -.0422 0025 .1285 .0451 .0488 .0589 1.3068 1.3570 .7692 Avg. -.0191 - 0159 -.0068 .0429 .0423 .0407 .-..--...--..--.-.-.-----...--.-.....----.-..........-...-.-...---..--..-.---- Note: Numbers 1, 2, and 3 refer to alternative elasticities of substitution (0.11, 0.21, 0.40, respectively). Symbols a and b represent efficiencies of labor and capital. VAP's represent value added productivity indices. 132 APPENDIX M. (cont'd.). Table M2. Rates of Change in .Factor Efficiencies, and Indices of Technological Change for the U.S. P 8 P Industries. USA Year b1 b2 b3 81 82 a3 VAPi VAPZ VAP3 1958 1.0000 1.0000 1.0000 59 .0400 .0270 .0035 .0682 .0748 .0866 1.0532 1.0423 .9053 60 -.0043 .0019 .0131 .0328 .0340 .0362 1.0665 1.0669 1.0187 61 .0045 .0157 .0359 .0686 .0744 .0848 1.1029 1.1293 1.1154 62 .0313 .0385 .0513 0428 .0437 .0454 1.1434 1.1841 1.1836 63 .0254 .0237 .0208 .0592 .0631 .0701 1.1906 1.2363 1.1342 64 .0301 .0278 0239 0587 .0635 .0722 1.2422 1.2935 1.1203 65 .0066 .0073 .0086 0538 0590 .0685 1.2770 1.3400 1.1697 66 .0179 .0180 .0182 .0521 .0554 .0614 1.3194 1.3904 1.1045 67 -.0567 -.0458 -.0262 -.0236 -.0298 -.0410 1.2646 1.3445 1.1408 68 0427 .0559 .0798 .0705 .0777 .0907 1.3353 1.4588 1.0076 69 .0316 0384 0507 .0391 .0424 .0483 1.3823 1.5310 .9784 70 -.0234 -.0116 .0096 .0374 .0392 .0426 1.3914 1.5706 1.2232 71 .0166 .0404 .0832 .0508 .0531 .0571 1.4392 1.6801 1.1542 72 .0466 .0397 .0273 .0881 .0925 .1002 1.5376 1.7903 1.2263 73 .0117 -.0071 -.0409 .0477 .0519 .0595 1.5833 1.8070 .9077 74 -.0739 -.1269 -.2221 -.0018 -.0015 -.0011 1.5168 1.5824 .9992 75 -.1825 -.1760 -.1643 -.0697 -.0838 -.1090 1.3114 1.3589 1.2342 76 .0762 .0858 .1032 .1068 .1146 .1286 1.4288 1.5141 1.0410 77 .0198 .0392 .0741 .0379 .0369 .0351 1.4688 1.5997 .9822 78 -.0044 -.0011 .0048 .0345 .0369 .0412 1.4891 1.6345 .8562 79 -.o132 -.0233 -.0414 .0286 .0337 .0430 1.4978 1.6298 .9030 80 -.0414 -.0277 -.0032 .0026 .0059 .0119 1.4656 1.6357 1.1918 81 -.0787 -.0841 -.0937 -.0095 -.0099 -.0107 1.3966 1.5444 1.4101 82 -.0406 -.0194 .0187 .0726 .0789 .0901 1.4134 1.6245 1.6774 Avg. -.0049 -.0027 .0015 .0395 .0419 .0463 Note: Nunbers 1, 2, and 3 refer to alternative elasticities of substitution (0.17, 0.27, 0.60, respectively). Symbols 8 and b represent efficiencies of labor and capital. VAP's represent value added productivity indices. 133 APPENDIX N. Factor Input Shares in Gross Value of Production. N. Factor Input Shares in Gross Value of Production. Year 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 wL .1369 .1557 .1457 .1443 .1589 .1571 .1576 .1619 .1690 .1738 .1612 .1488 .1455 .1630 .1671 .1684 .1393 .1987 .2179 .2208 .1937 .1804 .1875 .1837 .1917 Finland WK WM .1927 .5305 .1700 .5292 .1796 .5307 .1787 .5512 .1387 .5732 .1528 .5657 .1430 .5847 .1211 .6071 .1044 .6162 .1209 .5994 .2189 .5287 .2365 .5193 .2415 .5227 .1501 .5769 .1445 .5749 .1575 .5667 .2111 .5154 .1574 .4971 .0624 .5549 .0805 .5367 .1604 .4961 .2165 .4621 .1994 .4423 .1825 .4411 .1427 .4536 wE .1399 .1451 .1439 .1257 .1292 .1245 .1147 .1099 .1104 .1059 .0912 .0954 .0993 .1100 .1135 .1074 .1342 .1469 .1648 .1620 .1499 .1411 .1709 .1926 .2120 wL .2316 .2239 .2285 .2301 .2325 .2326 .2285 .2244 .2214 .2335 .2333 .2332 .2380 .2432 .2418 .2232 .1808 .1901 .1899 .2008 .2039 .1939 .1881 .1831 .1950 USA VK .2495 .2717 .2689 .2620 .2549 .2637 .2704 .2730 .2746 .2611 .2515 .2460 .2363 .2120 .2256 .2394 .2676 .2485 .2454 .2293 .3216 .2416 .2201 .2208 .2147 "M .4623 .4514 .4486 .4517 .4564 .4478 .4466 .4484 .4520 .4511 .4616 .4682 .4680 .4769 .4637 .4693 .4632 .4567 .4609 .4557 .4515 .4481 .4658 .4614 .4523 wE .0566 .0530 .0541 .0561 .0562 .0559 .0546 .0542 .0520 .0542 .0535 .0527 .0577 .0679 .0689 .0681 .0884 .1047 .1037 .1142 .1131 .1164 .1260 .1347 .1379 APPENDIX 0. Annual Changes in Total Factor.Productivity in Table 0. Annual Changes in Total Factor Productivity in the 134 the Finnish and U.S. P 8 P Industries. Finnish and U.S. P 8 P Industries. FIN USA Year TFP1 TFP2 TFP3 TFP1 TFP2 TFP3 59 -.0073 -.0076 -.0078 .0327 .0319 .0323 60 .0441 .0448 .0448 .0022 .0011 .0012 61 .0122 .0120 .0120 .0131 .0124 .0124 62 .0014 .0003 .0004. .0189 .0189 .0189 63 .0433 .0421 .0423 .0188 .0186 .0186 64 .0410 .0394 .0401 .0141 .0143 .0143 65 -.0031 -.0041 -.0041 .0126 .0117 .0117 66 .0213 .0203 .0202 .0184 .0182 .0181 67 -.0084 -.0099 -.0100 -.0189 -.0194 -.0195 68 .0404 .0401 .0401 .0191 .0195 .0196 69 .0198 .0193 .0189 .0132 .0129 .0129 70 -.0041 -.0046 .-.0047 -.0101 -.0110 -.0111 71 -.0785 -.0796 -.0797 -.0006 -.0012 -.0013 72 .0840 .0839 .0841 .0245 .0247 .0248 73 .0222 .0209 .0210 .0439 .0438 .0438 74 -.0305 -.0302 -.0304 .0246 .0242 .0237 75 -.0730 -.0743 -.0727 -.0668 -.0680 -.0687 76 -.0080 -.0090 -.0092 .0283 .0287 .0301 77 .0219 .0216 .0215 .0012 .0006 .0013 78 .0368 .0370 .0376 .0103 .0097 .0097 79 .0967 .0965 .0964 .0110 .0106 .0107 80 .0003 -.0005' -.0003 -.0141 -.0143 -.0145 81 .0293 .0297 .0297 .0020 .0021 .0019 82 .0540 .0539 .0539 -.0169 -.0175 -.0173 Avg. .014823 .014241 .014343 .007571 .007185 .007236 Note TFP1 - no quality adjustments TFP2 - quality adjustments, capital gains excl. TFP3 - quality adjustments, capital gains incl. 135 APPENDIX P. Development of Alternative Real Value Added Measures in the U.S. P 8 P Industry. USA -JJ -015- '92 1 li‘iilill 11 Year 11r1111111 111 5859606162636 65666768697071”73747576777879308182 D 0W! + DOUflR Figure P. Development of Alternative Real Value Added Measures in the U.S. P 8 P Industry. Note: QVAR is based on the output indices calculated by the Bureau of Labor Statistics. DQVAR is the double deflated real value added. 136 APPENDIX Q. Contributions of Capital, Material and Energy Deepening and TFP1 to Labor Productivity Growth. Table Ql. Contributions to Labor Productivity Growth in the Finnish P 8 P Industry (percent). Year Q/L wk(K-L) wm(M-L) we(E-L) TFP1 1959-82 4.34 0.95 1.43 0.47 1.48 (21.9) (32.9) (10.8) (34.1) 1959-62 4.87 1.06 1.88 0.66 1.26 (21.8) (38.6) (13.5) (25.9) 1963-67 6.57 0.91 3.02 0.76 1.88 (13.8) (46.0) (11.57) (28.6) 1968-72 5.32 0.77 2.89 0.34 1.23 (14.5) (54.3) (8.1) (23.1) 1973-77 -2.01 1.11 -l.82 0.05 -1.35 (55.2) (-90.5) (2.5) (-67.2) 1978-82 7.04 0.91 1.30 0.49 4.34 (12.9) (18.5) (7.0) (61.6) Table Q2. Contributions to Labor Productivity Growth in the wm(fi-i) 1959-82 1959-62 1963-67 1968-72 1973-77 1978-82 U.S. P 8 P Industry (percent). wk(X-L) 1.08 (31.8) 0.91 (22.5) 0.93 (28.2) 0.99 (22.0) 1.15 (43.1) 1.41 (54.4) 1.47 (43.4) 1.28 (31.7) 1.35 (40.9) 2.18 (48.5) 1.04 (38.9) 1.46 (56.4) we(E-L) 0.08 (2.3) 0.17 (4.2) 0.12 (3.6) 0.40 (8.9) -0.14 (-5.2) -0.12 (-4.6) TFP1 0.76 (22.4) 1.67 (41.3) 0.90 27.3) 0.92 (20.5) 0.62 (23.2) -0e15 (-5.8) BI BLIOGRAPHY BIBLIOGRAPHY Abramovitz, M. 1956. Resource and Output Trends in the U.S. Since 1870. American Economic Review 46(2): 5-23. American Paper Institute. 1984. Statistics of Paper, Paperboard 8 Woodpulp. Data through 1983. American Paper Institute, New York. Arrow, K.J. 1972. The Measurement of Real Value Added. Tech. Report no. 60, Econ. Series, Institute for Mathematical Studies in the Social Sciences, Stanford Univ. Arrow, K.J., Chenery, H.B., Minhas, 8.8. 8 Solow, R.M. 1961. Capital-Labor Substitution and Economic Efficiency. Review of Economics and Statistics 43(3): 225-250. Beckman, M. and Sato, R. 1969. Aggregate Production Functions and Types of Technical Progress. A Statistical Analysis. American Economic Review 59: 88-101. Bengston, D.N. and Strees, A. 1986. Intermediate Inputs and the Estimation of Technical Change: The Lumber and Wood Products Industry. Forest Science 32(4): 1078-1085. Berndt, E.R. and Christensen, L.R. 1973. The Internal Structure of Functional Relationships: Separability, Substitution, and Aggregation. Review of Economic Studies XL(3): 403-441. Berndt, E.R. and Watkins, G.C. 1981. Energy Prices and Productivity Trends in the Canadian Manufacturing Sector, 1975-76. Economic Council of Canada, Ottawa, Ontario. Berndt, E. R. and Wood, D. O. 1975. Technology, Prices, and the Derived Demand for Energy. The Review of Economics and Statistics, August: 259- 268. Boisvert, R. N. 1982. The Translog Production Function: Its Properties, Its Several Interpretations and Estimation Problems. Dept. of Agricultural Economics, Cornell University Agricultural Experiment Station, Ithaca, New York. Brown, M. 1966. On the Theory and Measurement of Technological Change. Cambridge University Press. 214 p. 137 138 --------- (ed.) 1967. The Theory and Empirical Analysis of Production. Studies in Income and Wealth, vol. 31, The National Bureau of Economic Research. Columbia University Press. New York. Bruno, M. 1978. Duality, Intermediate Inputs and Value-Added. In Fuss, M. and McFarren, D. (Eds.) Production Economics: A Dual Approach to Theory and Application, vol. 2: 3-16. Central Statistical Office of Finland. Industrial Statistics. Helsinki, various years. -------- Statistical Year Book of Finland. Helsinki, various Christensen, L. and Jorgenson, D. 1969. The Measurement of U.S. Real Capital Input, 1929-1967. Review of Income and Wealth, 15(4): 293-320. Christensen, L., Jorgenson, D. and Lau, L. 1973. - Transcendental Logarithmic Production Frontiers. Review of Economics and Statistics 55(1): 28-45. ' David, P. A. and van der Klundert, T. 1965. Biased Efficiency Growth and Capital-Labor Substitution in the U.S. 1899- 1960. American Economic Review 55: 375-95. Denison, E. F. 1961. Measurement of Labor Input: Some Questions of Definitions and the Adequacy of Data. Discussion by G. Tolley. In "Output, Input and Productivity Measurement”, vol. 25. The National Bureau of Economic Research. Princeton University Press. --------- 1962. The Sources of Economic Growth in the United States and Alternatives Before Us. Committee for Economic Development. New York. --------- 1967. Why Growth Rates Differ: Postwar Experience in Nine Western Countries. The Brookings Institution. Washington, D.C. --------- 1974. Accounting for United States Economic Growth 1929-1969. The Brookings Institution. Washington. D.C. Denny, M. and May, D. 1977. The Existence of a Real Value- Added Function in Canadian Manufacturing. Journal of Econometrics 5: 55-69. Diamond, P. A. and McFadden, D. 1966. Identification of the Elasticity of Substitution and the Bias of Technological: an Impossiblity Theorem, Working Paper no. 62. Institute 139 of Business and Economic Research, University of California, Berkeley. Diewert, W. E. 1976. Exact and Superlative Index Numbers. Journal of Econometrics 4: 115-45. --------- 1978. Superlative Index Numbers and Consistency in Aggregation. Econometrica 46: 883-900. --------- 1980. Aggregation Problems in the Measurement of Capital. In Usher (ed.) 1980b. Ennakkotietoja ETLA:n tyévoimakustannustiedusteluista. 1987. Unpublished Report, The Research Institute of the Finnish Economy. Helsinki. Farrar, D. E. and Glauber, R. R. 1967. Multicollinearity in Regression Analysis: The Problem Re-visited. Review of Economics and Statistics 49: 92-1077. Ferguson, C. and Moroney, J. R. 1969. The Sources of Change in Labor's Relative Share: A Neoclassical Analysis. Southern Economic Journal 35: 308-22. Forsund, F., Gaunitz, S., Hjalmarsson, L. and Wibe, S. 1978. Technical Progress and Structural Change in the Swedish Pulp Industry 1920-74. In Puu and Wibe. 1980. Forsund, F., Lovell, K. and Schmidt, P. 1980. A Survey of Frontier Production Functions and of Their Relationship to Efficiency Measurement. Journal of Econometrics 13(1): 5-25. Fraumeni, B. M. and Jorgenson, D. W. 1980. The Role of Capital in U.S. Economic Growth, 1948-1976. In von Furstenberg (ed.) 1980. von Furstenberg, G. M. (ed.) 1980. Capital, Efficiency and Growth. Ballinger Pub. Co., Cambridge, Massachusetts. Gollop, F. M. and Jorgenson, D. W. 1977. U.S. Productivity Growth by Industry, 1947-1973. Harvard Institute of Economic Papers. Discussion Paper 570. Cambridge, Massachusetts. Greber, B. J. and White, D. E. 1982. Technical Change and Productivity Growth in the Lumber and Wood Products Industry. Forest Science 28(1): 135-47. Heertje, A. 1977. Economics and Technical Change. Weidenfeld and Nicholson. London. 140 Hulten, C.R. 1978. Growth Accounting with Intermediate Inputs. Review of Economic Studies 78: 511-518. Humprey, D.B. and Moroney, J.R. 1975. Substitution among Capital, Labor, and Natural Resource Products in American Manufacturing. Journal of Political Economy, 83(1): 57-82. Intriligator, M. D. 1965. Embodied Technical Change and Productivity in the United States in 1929-1958. Review of Economics and Statistics 47: 65-70. --------- 1978. Econometric MOdels, Techniques and Applications. Englewood Cliffs. Prentice-Hall. New Jersey. 250 p. Johansen, L. 1972. Production Functions: An Integration of Micro and Macro, Short Run and Long Run Aspects. North- Holland. Amsterdam. Jones, H. G. 1975. An Introduction to Modern Theories of Economic Growth. McGraw-Hill, New York. 250 p. Jorgenson, D. W. and Griliches, z. 1967. The Explanation of Productivity Change. Review of Economic Studies 34(3): 249-83. Kalt, J.P. 1978. Technological Change and Factor Substitution in the United States 1929-67. International Economic Review 19: 761-76. Karhu, V. and Vainionmiki, J. 1985.Tutkimus kokonaistuottavuuden mittaamisen teoreettisista perusteista ja kokonaistuottavuuden muutoksista Suomen teollisuudessa 1960-80. Department of Economics, University of Tampere, Series B 60/1985. Tampere. 243 p. Katila, M. 1983. Teknisen kehityksen luonne ja tuotannon kasvu Suomen massa- ja paperiteollisuudessa 1954-1980. Unpublished Master's Thesis. Dept. of Social Economics of Forestry, University of Helsinki. Helsinki. 64 p. Kendrick, J. 1961. Productivity Trends in the United States. National Bureau of Economic Research. Princeton University Press. New York. 630 p. Kendrick, J. and Sato, R. 1963. Factor Prices, Productivity and Economic Growth. American Economic Review 53: 974-1003. Kennedy, C. and Thirwall, A. D. 1972. Surveys in Applied Economics: Technical Progress. Economic Journal 82: 11-72. 141 Xoskenkyli, H. 1979. The Definition and Measurement of Capital and Its Role in the Investment Function. Bank of Finland, Research Papers 4/79. Helsinki. 59 p. Lau, L. J. 1979. On Exact Index Numbers. Review of Economics and Statistics 61: 73-82. Lave, L. B. 1966. Technological Change: Its Conception and Measurement. Englewood Cliffs. Prentice-Hall. New Jersey. Manning, G. H. and Thornburn, G. 1971. Capital Deepening and Technological Change. The Canadian Pulp and Paper Industry 1940-1960. Canadian Journal of Forest Research 1(3): 159-66. Martinello, F. 1985. Factor Substitution, Technical Change and Returns to Scale in Canadian Forest Industries. Canadian Journal of Forest Research 15(6): 1116-1124. May, J.D. and Denny, M. 1979a. Factor-Augmenting Technical Progress and Productivity in U.S. Manufacturing. International Economic Review 20(3): 759-774. --------- 1979b. Post-War Productivity in Canadian Manufacturing. Canadian Journal Of Economics XII(1): 29- 41. Monthly Labor Review. 1984. Survey of Hours Worked. June 1984. Nadiri, M. I. 1970. Some Approaches to the Theory and Measurement of Total Factor Productivity. A Survey. Journal of Economic Literature 8: 1137-77. Nelson, R. R. 1964. Aggregate Production Functions and Medium-Range Growth Projections. American Economic Review 54: 575-606. Nerlove, M. 1967. Recent Empirical Studies of the CES and Related Production Functions. In Brown (ed.) 1967. Niitamo, O. 1958. Tuottavuuden kehitys Suomen teollisuudessa vuosina 1925-1952. Kansantaloudellisia tutkimuksia X. Helsinki. Ollonqvist, P. 1974. Teknologian kehityksen ja kapasiteetin kiytén mittaus tuotanto- ja kustannusfunktiossa. Unpublished Licenciate's Thesis. School of Economics. Helsinki. 142 Parkkinen, Matti. 1982. Teollisuustuotannon volyymi-indeksin laskenta. Central Statistical Office of Finland. Studies, No 76. Helsinki. 80 p. Pindyck, R. and Rubinfeld, D. 1981. Econometric Models 8 Econometric Forecasts. McGraw-Hill. Tokyo. 630 p. Productivity Analysis in Manufacturing Plants. BLS Staff Paper 3. U.S. Department of Labor. Bureau of Labor Statistics. Washington, D.C. 96 p. Profitability, Productivity, and Relative Prices in Forest Industries in the ECE Region. 1987. Report prepared by Indufor, Ky for the ECE/FAG Agriculture and Timber Division. Helsinki. Puu, T. and Wibe, S. (eds.) 1980. The Economics of Technological Change. St. Martin's Press. New York. 336p. Revankar, N. S. 1971. Capital-Labor Substitution, Technological Change and Economic Growth: The U.S. Experience 1929-1953. Metroeconomica XXIII: 154-176. Risbrudt, C. D. 1979. Past and Future Technological Change in the U.S. Forest Industries. Unpublished Ph.D. Dissertation. Michigan State University, East Lansing. Robinson, V. L. 1975. An Estimate of Technological Progress in the Lumber and Wood Products Industry. Forest Science 22(2): 149-54. Salonen, I. 1981. Teknisen kehityksen mittaamisesta tuotantofunktion avulla ja sovellutus Suomen kansantalouteen. Bank of Finland's Publications 0:51: 1-93. Salter, W. E. G. 1960. Productivity and Technological Change. Cambridge University Press. Cambridge. Sato, K. 1975. Production Functions and Aggregation. North.- Holland. Amsterdam. 313 p. --------- 1976. The Ideal Log-Change Index Number. Review of Economics and Statistics 58: 223-8. Sato, R. 1970. The Estimation of Biased Technical Progress and the Production Function. International Economic Review 11(2): 179-208. --------- 1980. The Impact of Technical Change on the Holothecity of Production Functions. Review of Economic Studies 47: 767-76. 143 Sato, R. and Beckman, M.J. 1968. Neutral Inventions and Production Functions. Review of Economic Studies: 57-65. Sato, R. and Hoffman, R. 1968. Review of Economics and Statistics L: 453-60. Sims, C. 1969. Theoretical Basis for a_Double Deflated Index of Real Value Added. Review of Economics and Statistics, LI(4): 470-471. Simula, M. 1979. Tuottavuus Suomen metséteollisuudessa. Unpublished Licenciate's Thesis. Dept. of Social Economics of Forestry, University of Helsinki. Helsinki. 193 p. --------- 1983. Productivity Differentials in the Finnish Forest Industries. AFF 180. Helsinki. 67 p. Smith, R. 1978. Duality Principles and Measuring the Production Technology: A Heuristic Introduction. Socio- Economic Planning Sciences 12: 161-6. Solow, R. 1957. Technical Change and the Production Function. Review of Economics and Statistics 39(3): 182-8. --------- 1960. Investment and Technical Progress. Mathematical Methods in the Social Sciences, 89-104. Stanford. --------- 1962. Substitution and Fixed Proportions in the Theory of Capital. Review of Economic Studies 30: 201-226. Stier, J.C. 1980a. Estimating the Production Technology in the U.S. Forest Products Industries. Forest Science 26(3): 471-82. --------- 1980b. Technological Adaptation to Resource Scarcity in the U.S. Lumber Industry. Western Journal of Agricultural Economics, Dec. 1980: 165-75. Thursby, J. G. 1980. Alternative CES Estimation Techniques Review of Economics and Statistics 72: 295-99. Tinbergen, J. 1942. Zum Theorie der langfristigen Wirtschaftsentwicklung. Weltwirtschaftlisches Archiv. 55: 511-49. Térnqvist, L. 1936. The Bank of Finland's Consumption Price Index. Bank of Finland, Monthly Bulletin no. 10: 1-8. U.S. Department of Agriculture. Forest Service. 1984. U.S. 144 U.S. Department of Commerce, Bureau of the Census. Annual Survey of Manufactures. U. S. Government Printing Office, Washington, D. C., various years. U.S. Department of Commerce, Bureau of the Census. Annual Survey of Manufactures. U. S. Government Printing Office, Washington, D. C., various years. --------- Census of Manufactures. U.S. Government Printing Office, Washington, D.C., various years. -------- 1981. Bureau of Economic Analysis. The National Income and Product Accounts of the United States, 1929-76, Statistical Tables. U.S. Government Printing Office, Washington, D.C., 429 p. U.S. Department of Labor, Bureau of Labor Statistics. 1985. Handbook of Labor Statistics. Bulletin 2217. U.S. Government Printing Office, Washington, DC; 449 p. -------- 1986. Productivity Measures for Selected Industries, 1958-84. Bulletin 2256. U. S. Government Printing Office, Washington, DC. 295 p. Usher, D. 1980a. The Measurement of Economic Growth. Basil Blackwell. Oxford. 306 p. - -------- (ed.) 1980b. The Measurement of Capital. Studies in Income and Wealth vol. 45. National Bureau of Economic Research. University of Chicago Press. New York. Vartia, Y. 1974. Ideal Log-Change Index Numbers. Mimeograph, Dec. Helsinki. --------- 1976. Relative Changes and Index Numbers. The Research Institute of Finnish Economy, Series A4. Helsinki. Wickens, M. R. 1970. Estimation of the Vintage Cobb-Douglas Production Function for the United States, 1900-1960. Review of Economics and Statistics 52: 187-93. Wohlin, L. 1970. Skogsindustrins strukturomvandling och expansions méjligheter. Stockholm. Wyatt, G. 1983. Multifactor Productivity Change in Finnish and Swedish Industries, 1960 to 1980. ETLA B 38. Helsinki. 119 p. Yearbook of Forest Statistics. Official Statistics of Finland. Series XV11 A: various years. Helsinki.. 145 You, J. K. 1976. Embodied and Disembodied Technical Progress in the United States, 1929-1968. Review of Economics and Statistics 58: 123-7. Aberg, Y. 1969. Produktion och produktivitetet i Sverige 1861-1965. IUI. Stockholm. --------- 1981. Produktiviteten i svensk industri éren 1953-1976. En studie i berékning av produktivitets- samband. Stockholm. -------- 1984. Produktivitetsutvecklingen i industrin olika OECD-lénder 1953-1980. IU, Forskningsrapport nr 25. Stockholm. 2‘