”$223.?" “1mg:\WWN This is to certify that the dissertation entitled USE OF IMPLAN TO ESTIMATE ECONOMIC IMPACTS STEMMING FROM OUTDOOR RECREATION EXPENDITURES IN THE UPPER LAKE STATES presented by Lawrence D. Pedersen has been accepted towards fulfillment Of the requirements for DOCTOR OF PHILOSOPHY dcgreein ESOURCE DEVELOPMENT Date May 18, 1990 MS U is an Affirmative Action/Equal Opportunity Institution 0- 12771 PLACE IN RETURN BOX to remove thlo checkout from your record. TO AVOID FINES Mum on or baton dale dun. DATE DUE DATE DUE DATE DUE " .1 1997' ll __ —_____l —_ 1:31. 1 a; 1:393 ' .4» d? HELL/.4995 ’- MSU Is An Atflmmlve ActionlEqunl Opportunity lnotltmlonm . .1.— USE OF IMPLAN TO ESTIMATE ECONOMIC IMPACTS STEMMING FROM OUTDOOR RECREATION EXPENDITURES IN THE UPPER LAKE STATES by Lawrence D. Pedersen A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Resource Development 1990 ABSTRACT USE OF IMPLAN TO ESTIMATE ECONOMIC IMPACTS STEMMING FROM OUTDOOR RECREATION EXPENDITURES IN THE UPPER LAKE STATES By Lawrence D. Pedersen The USDA Forest Service's IMPLAN input-output (I-O) has been used to generate estimates of outdoor recreation economic impacts, but the reliability of such estimates is largely unknown. Problems with IMPLAN's regional purchase coefficient (RPC) trade estimates were identified. Alternative RPCs were constructed from a reconciled 1977 Multi-Regional Input-Output (MRIO) database. Comparisons with IMPLAN's current RPCs reveal the alternative RPCs to be more internally consistent and in line with regional economic theory. The 1985 National Survey of Fishing, Hunting, and Wildlife-Associated Recreation (FEW) data were used with alternative IMPLAN models to generate outdoor recreation impact estimates for the upper Lake States region. Variables examined include recreation activity participation levels and spending patterns, sampling errors associated with the FHW activity and spending data, I-O model sectorization schemes, sector spending allocations, and trade estimates. All variables affected the magnitudes of total economic impact estimates. The range of spending estimates constructed to account for FHW sampling error had a large influence on impact magnitudes, indicating that reports of deterministic impact estimates may be misleading. Evidence presented concerning IMPLAN's RPC estimates indicates that improvements in. impact estimates could result from estimation of IMPLAN's RPCs from the reconciled MRIO database. 11 ACKNOWLEDGEMENTS Several persons made immense contributions to this study. The research was facilitated through cooperation from the IMPLAN staff. Eric Siverts deserves special mention for patiently helping me and literally hundreds of other IMPLAN users sort through I-O issues and problems. Professor Daniel Chappelle served as my major professor and mentor throughout my years in graduate school. He provided inspiration, financial support, technical guidance on research issues, and editorial assistance. He also imparted much wisdom on a variety of subjects, from research methods through regional economics. Many other MSU faculty and staff also had roles in enabling this research. Professor Larry Leefer's generous giftsof COIPPBEQLME and insights were especially important as were frequent contacts with Professor Dennis Propst. Julie Peeler's word processing and guidance on many otpejmmatters was much appreciated. There have also been various ,cOmpadres‘, including George Erickcek and Michael Oden, who have made ‘Ehe prSEess more pleasureable. My family and relatives have been continual sources of support. Finally, the person to whom I owe a great debt for making it all possible is my wife, Helen. To all, thanks. iii Table of Contents List of Tables II. III. INTRODUCTION Background Need Use of Outdoor Recreation Data in Impact Analysis Input- Output Model Influences on Outdoor Recreation Impact Estimates The Concept of Accuracy as It Applies to I- O Analysis Application of Valuation and Impact Concepts to Recreation Impact Analysis . . Terminology . . Residents Versus Nonresidents Comprehensive Impact Analysis Study Premises and Assumptions Regarding Influences of Variables on Impact Estimates Study Objectives INPUT-OUTPUT ANALYSIS AND IMPLAN Introduction Input- Output Analysis . Nonsurvey I- O: Adaptation of National Coefficients for Regional Models Location Quotients Supply- -Demand Pooling Regional Purchase Coefficients IMPLAN . IMPLAN Version 2. O' s Adaptation of RFC Method Derivation of IMPLAN' s RPCs IMPLAN Differences from the Stevens ApprOach Evidence of Problems with IMPLAN' 5 RFC Estimates Nonsurvey Input-Output Model Validation Procedures for Evaluating IMPLAN's RPCs Measurement of Outdoor Recreation Economic Impacts Secondary Sources of Data for Outdoor Recreation I-O METHODS Introduction . Examination of IMPLAN' 5 Regional Purchase COefficients Initial Identification of RFC Problems . Development and Comparison of Alternative RPCs Lake State Outdoor Recreation Economic Impacts Variables Examined . Activity and Spending Data . Temporal and Spatial Considerations Different I- O Sectorization Schemes Allocation of Survey Spending Data to IMPLAN Sectors Influence of RPC- Based Trade Estimates iv 1’ .W #3 <1» 22 22 26 28 34 36 4O 4O 4O 43 44 45 45 47 50 50 51 52 55 62 63 68 7O 7O 7O 7O 71 83 86 88 89 91 93 IV. VI. Table of Contents (cont'd.) PREPARATION OF INPUT DATA AND IMPLAN MODELS Introduction Converting Fishing, Hunting, and Wildlife Associated Data for Use with IMPLAN . Stage 1: Compile 1985 Recreation Activity Levels for More Heavily Forested Areas of the Lake States Stage 2: Estimate Study Area Recreation Spending Totals . . . Stage 3: Disaggregate Spending Totals for Bridging with IMPLAN Stage 4: Allocate Spending to IMPLAN Sectors Construction of Lake State IMPLAN Models Development of Alternative RPC Trade Estimates Different I- O Sectorization Schemes RESULTS Introduction Examination and Documentation of Problems with IMPLAN' s RPCs . Sector Allocation of OutdOor Recreation spending . Lake State Outdoor Recreation Economic Impacts . Differences Between Lake State IMPLAN Model RPCs Differences Between Lake State IMPLAN Model Multipliers Estimates of Lake State Outdoor Recreation Economic Impacts . Range of FHW Impact Estimates by Major Spending Category . Relationships Among Categories of Lake State FHW Recreation Spending . . . Influence of Sectorization Scheme on Lake State FHW Impacts . . Influence of RPCs on Lake State FHW Impacts . Comparison to Outdoor Recreation Impact Estimates Prepared for the 1987 Lake State Governors’ Conference on Forestry . Comparison of Recreation Multipliers to "Average" Multipliers SUMMARY, IMPLICATIONS, AND CONCLUSION Introduction Use of Nonsurvey I- O Models IMPLAN . . Outdoor Recreation Economic Impacts Influences on the Size of Lake State Outdoor Recreation Impacts Further Research Needs Conclusion Page NoI . 9 9 4 l 94 96 99 106 108 114 114 116 119 119 119 130 138 138 142 147 147 150 152 154 156 158 161 161 161 164 168 170 175 177 Page No; VII. APPENDICES A. Industry Classification of the Micro-IMPLAN 528 Sector Input/Output Tables . . . 179 B. Comparison of IMPLAN, REMI, and Corrected MRIO RPCs . 188 C. Alternative Lake State RPCs and RPCs Used as Guides for the Estimation of the Lake State RPCs . 191 D. IMPLAN Aggregation and Sectorization Schemes . 202 E. Lake State Outdoor Recreation Economic Impacts . 212 F. Bridge of National Survey of Fishing, Hunting, and Wildlife-Associated Recreation (FHW) Spending to IMPLAN 221 G. FHW Used Equipment Allocation Table . 226 H. Lake State Multiplier Analysis . 229 1. Michigan RPC and Multiplier Analysis . 234 VIII.REFERENCES . 240 Table of Contents (cont'd.) vi 10. 11. 12. 13. 14. 15. 16. 17. 18. LIST OF TABLES Page No, Michigan RPC and SDP Trade Estimates . . 17 1983 and 1985 REMI Motor Vehicles and Equipment RPCs 77 FHW Lake State Study Area % of 3 State Recreation Activity 99 FHW Lake State Study Area Spending Totals . 104 Comparison of IMPLAN and REMI RPCs for Major Sector Groupings . 122 IMPLAN and Revised State Average Service Sector RPCs and Rankings . . . . . . . . . . . . . . . . . 125 Ten Lowest and Highest Revised MRIO Average Service Sector RPCs 126 Comparison of Recreation-Related Service Sector RPCs for Michigan, New Jersey, New York, and North Dakota . 127 Michigan Output Multipliers for Recreation-Related Sectors 129 Pce I¢O Category 9400: Wheel Goods, Durable Toys, Sport Equipment, Boats & Pleasure Aircraft . . . . . . 132 Contrast of Sector Allocations Based on I-0 Category #9400 Versus Small Arms, Optical Goods, and Boat Building & Repair PCE Items within I-O Category #9400 . . . . . . . . Contrast of Lake State Output Impacts Based on 1-0 Category #9400 Versus Small Arms, Optical Goods, and Boat Building and Repair PCE Items Within I-O Category #9400 . . . . Measurement of Lake State Model RPC Differences Mean Values of Lake State Model Multiliers Lake State Multiplier Um (Theil Index Bias) Measurements . Lake State Economic Impacts from Upper Lake State Fishing, Hunting, & Wildlife-Associated Recreation, Based on Disaggregated, Adjusted RPCs (BSTLK) IMPLAN Model Lake State Subcategories of Recreation Spending as a Percent of Total Upper Lake State Fishing, Hunting, & Wildlife- Associated Recreation, Based on Disaggregated, Adjusted RPCs (BSTLK) IMPLAN Model . . . . . . . . . Impacts of Lake State FHW Recreation Spending for Different Sectorization Schemes as a Percent of the Disaggregated, Adjusted RPCs (BSTLK) IMPLAN Model . . . vii 134 135 141 144 146 . 148 . 151 153 19. 20. 21. 22. LIST OF TABLES (cont'd.) Page Impacts of Lake State FHW Recreation Spending for Different Sets of RPCs as a Percent of the Disaggregated, Unadjusted (UNCHLK) IMPLAN Model . . . . . . . . . . . . . . Comparison of Estimated Type III Impacts Based on 1987 Recreation Spending Profiles . . . . . . . . . . BSTLK % of 1987 Lake State Outdoor Recreation Impact Estimates . . . . . Contrast of Impacts from Using Average Vs. Sector-Specific Multipliers for Recreation Spending, Based on a 31-Sector Model and Revised (BSTLK) IMPLAN RPCs viii No, . 155 . 157 . 158 160 CHAPTER I INTRODUCTION Mermaid State governments have become involved in sponsoring a variety of initiatives to foster economic development in recent years. Fosler (1988) documents this trend in 1113 m W 33]; of Mexican States. Nothdurft (1984) reviews recent state economic initiatives that focus on natural resources. Tourism and recreation have been major areas of emphasis for state economic development efforts. Reliable estimates of the relative economic impacts from alternative development programs are useful in planning economic development. Knowledge of economic impacts can improve governmental programs aimed at promoting economic growth. An industry's relative economic impacts vis-a-vis other industries are an appropriate consideration in the case of "targeting" a particular industry for special governmental assistance. Other factors to consider include the industry's growth prospects, the comparative advantages for the industry that exist in the state versus elsewhere, and anticipated social benefits. People recognize the need for better economic data and information on outdoor recreation. This is illustrated by the summary of outdoor recreation key issues and recommendations in Appendix I of Americans Outdoors; The Legacy, the Challenge, (President's Commission on Americans Outdoors, 1987). According to the study, states are reported to have expressed needs for "improved information collection and analysis to provide a better base for decisions" and "better 1 2 identification of values to help justify actions recommended" (p. 281- 282). The report goes on to state: "Most state assessments reference the values of recreation and the outdoors. These concerns about values are often closely related to overall planning and research needs." Two of the three primary areas encompassed by the concerns expressed were ”better recognition by and communication to the public about recreation values" and. the "need. for' more research and documentation. of quantitative benefits of recreation to the economy..." (p. 282). Two recent examples of efforts explicitly aimed at improving the quality of outdoor recreation impact analysis are the Public Area Recreation.‘Visitor Survey (PARVS) and the estimation of ‘upper Lake States outdoor recreation. economic impacts (Pedersen and Chappelle, 1987). The PARVS effort has involved Federal and State agencies in coordinating surveys of park recreationists in several regions of the United States. The primary objective of PARVS was to "generate spending _fi'ffflgfli data needed to determine the economic impacts (jobs, income, etc.) of public agency “expenditures for recreation facilities and services" (Propst, 1988, p.4). The estimation of outdoor recreation economic impacts in the upper Lake States (Michigan, Minnesota, and Wisconsin) was part of an effort to measure the contribution of the forest resources to the region. The funding for these and other similar recreation studies indicates there/is a demand forireliable economic impact measurements of .. .- outdoor recreation. At the same time, however, reliable recreation economic impact measurement is constrained by many -factors. These factors may be viewed as generally falling within three categories: the definition of outdoor recreation, the input data used with the impact 3 model, and the impact model itself. This thesis focuses on issues relating to outdoor recreation input data and impact models. Brief attention is also given to outdoor recreation definition issues. The need for this study is further discussed in the following section. Other preliminary concepts and an introduction to the subject are then presented to establish the context of the dissertation research. The chapter concludes with a statement of study objectives. Neeg A plethora of data on outdoor recreation and related statistics on travel and tourism exists, but there has been little effort devoted to comparing and relating such data, especially within the context of recreation economic impact analysis. In the absence of consistency checks across studies and statistical information, regional outdoor recreation participation and spending profiles largely remain untested and unreliable. This, in turn, inhibits the development of credible outdoor recreation economic impact estimates. There are several reasons why comparisons of recreation data and economic impacts are difficult to undertake, including a lack of consensus concerning outdoor recreation and tourism definitions, the multi-purpose nature of many outdoor recreation trips (recreationists often participate in a mix of activities), and different objectives in conducting outdoor recreation and tourism studies. As stated in the Methodological Notes accompanying the paper presented by Pedersen and Chappelle at the 1987 Lake States Governors' Conference on Forestry (p. 4): . At the time of this study, no complete, consistent, and reliable outdoor recreation data base exists that can be used in an economic impact analysis of a sub-region of a multi-state 4 area (such as the more heavily forested areas of our three state region). The lack of standardization in data reflects different ' objectives under which the data were originally collected. For example, outdoor recreation-related data may be gathered to measure tourism, state park attendance, or the multiple demands on lakes or streams. Units of measurement range from simple head counts through trips, "occasions," recreational visitor and activity days, and hours spent in the activity. Studies also include and exclude different types of recreationists, again, depending on their objectives. There are many other recreation analysis pitfalls. The sheer diversity of activities and recreationists makes impact estimation difficult. Double-counting is a risk when using multiple sources of data as in this study. Fishermen camp and campers fish; separating out what multiple data sources have included may be impossible. Typical spending categories mentioned in studies frequently do not fit the SIG codes used in impact analysis. Examples include: ”transportation" and "vehicle-related," which could refer to a myriad of sectors besides gasoline, and "food," which may or may not include restaurants and beverages. Spending will also differ depending on the recreationist's origin and destination, lodging, activities engaged in, and the recreation season. In addition to the differences between studies noted above, outdoor irecreation economic impact studies often differ in their treatment of two categories of economic activity affected by recreation activity: recreation-related equipment purchases and fiscal impacts on various levels of government. Many studies focus only on trip expenditures and do not address either durable equipment spending or the public costs of providing the recreation experience. Comprehensive profiles of outdoor recreation economic impacts might also give consideration to expenditures in the region that occur in preparation for recreation outside the region. There are also more esoteric impact issues such as analyzing changes in personal consumption expenditures for food and other items that stem from successful hunting and fishing. However, budgetary, data, and time constraints often prohibit the‘ development of comprehensive recreation impact analyses. \‘/ 5 This thesis examines several specific means to refine recreation economic impact estimates generated by the USDA Forest Service's IMPLAN (IMpact analysis for PLANning) model (U.S. Department of Agriculture, 1983). Important issues faced in estimating recreation economic impacts will be illustrated through generating estimates of upper Lake States outdoor recreation economic impacts using Micro IMPLAN (Version 2.0) (Alward et a1. , 1989). MQEQL—ltdoo wmmmam Accurate recreation participation and spending profiles need to be constructed before reliable recreation economic impacts may be estimated. Spending profiles may then be converted into final demand vectors which, in turn, "drive" input-output (LO) models. Thus, the levels of direct, indirect, and induced sales, value-added, income, or employment impacts generated by a recreation I-O analysis critically depend on participation and spending estimates. In order to expand their usefulness and address the myriad of objectives facing state planners, spending profiles and associated final demand estimates would ideally be developed (and, subsequently, be capable of being delineated) according to a number of variables. These variables include recreation activities, types of accomodation used, recreationists' residency status, and, preferably, substate regions and season. Developing distinct spending profiles according to these variables would permit: - the flexibility to examine issues for different objectives and multiple interests. Among other reasons, such flexibility is desirable with multiple definitions of tourism and recreation existing and private interests often centered around specific activities (e.g., hunting or fishing); - improved estimates as new primary and secondary data become available, be they' estimates for lodging, recreation activity levels, or other recreation-related variables; and - more crosschecking of estimates across studies in order to examine their consistency and reliability. Regarding this last point, outdoor recreation data come from national, state, and local sources. There have been few efforts to contrast participation level and spending pattern estimates across and within different levels of the spatial hierarchy. Conclusions stemming from comparisons of estimates across recreation studies are inhibited by differences in time when studies were conducted, what they measured and the measurement units they used, and low precision caused by small sample sizes in some cases. However, such comparisons may at least give some qualitative impressions of consistency and provide a measure of the reliability of outdoor recreation economic impacts. The reliability of outdoor recreation economic impacts should be questioned if gross inconsistencies between outdoor recreation data are found and not resolved before generating the impact estimates. In light of the importance of the final demand estimates for input- output analysis, the reliability of recreation participation levels and spending profiles are probably at least as important and in need of review as is the input-output model used to generate the economic impact estimates. However, problems ‘with. an I-O :model may sometimes be ixientified and rectified, leading to generic changes in the I-0 modeling system. Such changes could then help all future users of the LO modeling system. In addition to possible generic I-O improvements, 7 refinements in the analysis at the stage of using the I-0 model may also sometimes be more easily and quickly achieved. Attention to the LO model in such cases may be justified on the basis of efficiency. This study will describe where generic improvements may be made in the IMPLAN modeling system and associated IMPLAN user materials, especially as they apply to outdoor recreation economic impact estimation. Issues and variables encountered throughout the impact estimation process are. described, along ‘with their influences on. the ultimate impact estimates. This examination provides some indications of where the greatest amount of future research time and effort should be spent. In order to improve the reliability of recreation impact measurements, it may be wise to extrapolate values from other studies for certain variables which have consistent values across studies and to devote more time to those variables with wide ranging values. A savings of time resulting from extrapolating values from other studies could also be spent in developing more comprehensive assessments of impacts through pursuing the often-neglected aspects of outdoor recreation impact analysis mentioned earlier (durable equipment purchases and public fiscal impacts). Despite interest in outdoor recreation impacts, funding for impact studies is limited. Efficient allocation of research resources requires a marginal return approach. This approach would devote the greatest amount of attention to refining measurement of those variables which most affect the precision and reliability of impact estimates. ut- ut ut Model Influences ee Outdoor Recreatiog Impact Eseimaees Recreation data and 1-0 models are two major sources of influences run the magnitude of outdoor recreation impact estimates. The quality and consistency of recreation participation and spending data are, at best, untested. In contrast, there have been recent input-output (I-O) improvements in terms of structural I-O techniques used, model accessibility, and user aids for measuring outdoor recreation I-O impacts. This is specifically true with regards to the USDA Forest Service's IMPLAN model. IMPLAN version 2.0 is available in a personal computer (pc) version which allows for faster turnaround and greater user input in model development. It has adopted a new trade estimation technique in place of one which is known to overestimate regional impacts. It also has separate retail and wholesale trade sectors which permit greater precision in measuring impacts from recreation-related expenditures. Additionally, IMPLAN training materials and computer spreadsheet aids have been developed for recreation impact analysis. The range in size of multipliers provide a further indication that research on recreation participation and spending data may provide greater refinements in generating reliable recreation economic impact estimates. Sectoral multipliers generated by IMPLAN for any given region and type of economic variable tend not to vary from each other by more than a factor of one. If estimated properly, they seldom are outside of a range of one to three. For any given region, most multipliers of the same type (e.g., sales, income, jobs, Type I, or Type III) are within 50%, plus or minus, of the average multiplier for that type. In contrast to most multipliers, recreation participation and spending estimates are "all over the map." In other words, on a percentage basis, differences betwen multipliers appear to be less than differences between recreation participation and spending estimates. fITnerefore, successful efforts at improving the accuracy of participation zitmd spending data are likely to improve the reliability of estimates of t:c>tal outdoor recreation. economic impacts more than refinements in sectoral allocations of spending (the distribution of recreation eeacpenditures across input-output accounts) or improvements in multiplier accuracy. This assumes the objective is to develop reliable recreation economic impact estimates. However, objectives may sometimes be tainted by political motivations to inflate the importance of a particular iJndustry. Also, the objective may be to measure multipliers or assess tflne distribution of spending across economic sectors, or both, rather than estimate total economic impacts. This second point is further addressed below under the heading, "The Concept of Accuracy as It Applies to I-O Analysis." Ideally, consistency checks and sensitivity analyses are conducted tlrroughout an entire impact estimation process. Several dimensions at time stage of utilizing IMPLAN may have a significant bearing on final inmpact results. Addressing some of these issues may be warranted on the basis that they may be more cost effective than improving the quality of reacreation participation and spending data. Five specific issues relating to use of the input-output model could be investigated in the Process of conducting sensitivity analysis of economic impacts: 1) the sectorization scheme: minimizing aggregation error and testing for sector spending allocation error (this is essentially an extension of checking the consistency of spending profiles across outdoor recreation studies); 2) alternative deflators: Appendix D of the IMPLAN Version 1.1 Analysis Guide ‘bridges BLS deflators for 110+ sectors to IMPLAN's (version 1.1) 464. (Updated BLS deflators have also been bridged to version 2.0’s 528 sectors and are available to IMPLAN model users.) There are alternative deflators, ranging from gross national or local consumer price indices's (CPIs) 10 through very narrowly defined, industry-specific producer price indices (PPI). (Regardless of whether deflators form some part of a sensitivity analysis, they should be used to convert data for any year other than 1982 to IMPLAN version 2.0's 1982 base year. Accuracy of the deflators is related to the issue of sectorization and aggregation error.); 3) allocation of spending to 1-0 sectors: "bridging" spending from survey responses to I-O sectors is compounded by ambiquities in survey responses and survey spending categories, and by aggregation or classification differences between survey spending categories and LO model sectors. The process of converting purchaser prices to producer prices ("margining") must be conducted for most recreation expenditure surveys to correctly use them with LO tables which are based on producer prices. Appendix E of the IMPLAN Version 1.1 Analysis Guide provides margins for most of the 100+ personal consumption (PCE) categories associated with the 1977 U.S. input-output accounts. Detailed Bureau of Economic Analysis (BEA) worksheets are available which further differentiate these PCE categories into over 1700 PCE items. Use of the margins from the IMPLAN manual may involve aggregation or spending allocation error, while the detailed BEA worksheets tend to be cumbersome and time consuming; 4) employment estimates: County Business Patterns (CBP) is a major data source used to formulate IMPLAN's employment estimates. The CBP does not incorporate estimates of self- employed persons and is based on March surveys. Another complication for deriving reasonably accurate estimates of recreation employment is the seasonality and transitory nature of recreation employees. As noted by Propst et al. in Assessing the Secondary Economic Impacts 9_f Recreatiog m Toerism; Work Team Recommendations (in Propst, compiler, 1985, p. 59), the induced portion of the impact may be overstated if the summer recreation employees do not match average employee spending patterns; 5) consistency checks with other secondary economic statistics: the literature on validation of LG models and estimates has tended to emphasize comparisons with primary models, but there have been several suggestions and some studies made (e.g., Siverts, 1988) which have focused on looking at additional secondary economic data. A concern related to this fifth issue is the reliability of IMPLAN's trade estimates. Version 1.1 of IMPLAN used a technique called supply-demand pooling to generate trade estimates between a region's industries and the outside world. This approach to trade estimation originated with Isard (1953). Supply-demand pooling is based on a net 11 trade concept. The difference ‘between regional demand and regional supply (output) is assumed to be imported if there is greater demand than supply; if supply exceeds demand, the excess is assumed to be exported. In other words, local supplies are assumed to be exhausted before imports are turned to or, conversely, local demand is filled before exports occur. The net trade concept ignores crosshauling which is frequently observed in the real world. Imports tend to occur for most goods and services even if local supply is adequate to meet local demand and, similarly, exports tend to occur even if local supply can not meet local demand. This is a general phenomenon across all regions and sectors, although it holds true more for small regions than. large, complex (diversified) regions and more so for manufactured goods than services. Also, the degree of crosshauling observed will be affected by the degree of sectorization detail. The consequence for adopting the supply-demand pooling approach is that economic impacts may easily be overestimated” As stated in "Regional Non-Survey Input-Output Analysis with IMPLAN" (Alward, et al., 1985, page 8), "In general, IMPLAN multipliers tend to be larger... probably due in large part to the maximum trade assumption.” This makes intuitive sense. If every time something is purchased locally it generates round after round of purchases of products that are assumed to be produced locally rather than imported, then the estimated impact will be greater. Less "leakage" as it is called, creates larger impacts. This is reflected in inflated multipliers. Version 2.0 of IMPLAN uses an alternative approach that indirectly .accounts for crosshauling and is based on gross trade flows estimated 12 through the use of regional purchase coefficients (RPCs). The approach has been developed largely through the work of Stevens and his colleagues (Stevens, et al., 1983) and is also used by Regional Economic Models, Inc. (REMI). Basically, the amount of local output purchased to meet local demand is determined econometrically. Independent variables used to estimate the RPCs include such factors as the physical size of the region, transportation and other factor costs such as wages, and the relative share of total regional employment an industry comprises versus the industry's share of total employment on the national level. After estimating the proportion of demand supplied locally, the remainder of demand is assumed to be met through imports, and the difference between total output and the amount of output consumed locally is assumed to be exported. The estimation of trade between regions has been found to be a critical factor in determining the size of impacts. Richardson (1972, p. 175) has stated, "It is widely known that the effects of changes in trade coefficients, especially in an expanding region, can have a bigger impact on the structure of the regional economy than changes in technological coefficients due to technological change or product mix." While acknowleding some dissent, Stevens et a1. (1986, IL. 2) contend there appears to be "general agreement" that the accuracy of regional purchase coefficients is "most crucial to the accuracy of any regional I-O model." 111 his 1985 review article on input-output and economic base inultipliers, Richardson calls the Stevens et a1. RPC approach a "welcome change from endlessly repetitive and mechanical location quotient methods..." of trade estimation, but cautions the approach may face 13 difficulties due to reductions in the Census Bureau's Transportation Survey data collection and dissemination. The RPC approach has relied quite heavily upon an aging 1977 Commodity Transport Survey database. The 1982 Commodity Transport Survey was not released by the Census Bureau because of substantial discrepancies found in subsampling after the Survey was completed. Theoretically, the RPC approach may generate more realistic trade estimates and subsequent economic impact estimates than the supply- demand pooling technique. This is because it does not automatically assume maximum local trade. However, there is little empirical evidence that the RPC technique is consistently more accurate. Most literature on the performance of alternative nonsurvey trade estimation techniques predates. the full. development of the econometric RPC approach. Two papers coauthored by Stevens, the originator of the technique, (1983 and 1986) comprise the major evidence on the performance of the RPC technique versus other techniques. Stevens et a1. (1983) reported mixed results when comparing RPC- based I-O models to survey-based models for the states of Washington and West Virginia. The causes of the mixed results were partly attributed to the lack of documentation for the West Virginia survey-based model. The authors also noted that the regression-derived RPCs were, at that time, underestimating true RPCs because they were based on the 1972 Census of Transportation which did not report shipments under 25 miles. Most unreported shipments under 25 miles would be intrastate shipments. As a result of not incorporating shipments under 25 miles, intrastate shipments as a percent of total shipments were underestimated. This would tend to make RPCs based on the 1972 Census underestimates of 14 actual (real world) regional purchase coefficient values. This problem was eliminated for later versions of RFCs based on the 1977 Census of Transportation which reports shipments under 25 miles. Another reference to an empirical comparison of trade estimation techniques including the RPC approach is in an unpublished paper by the Regional Science Research Institute (RSRI) (Stevens, et al., 1986). The authors argue that survey-based tables or multipliers may be flawed due to missing data, small sample sizes, and their higher level of aggregation relative to secondary I-O models. Thus, they contend it is dubious to assess the accuracy of nonsurvey I-O models based solely on comparisons with survey-based tables or multipliers. In contrast, the authors adopt a different approach that compares the RPC technique against the supply-demand pooling technique and two other nonsurvey trade estimating techniques (location quotients based on employment and output) in estimating what are termed "known" or "observed" RPCs. These "known" RPCs are "constructed by the Regional Science Research Institute" from a multitude of secondary data sources, instead of being derived from primary surveys. As would be expected on a theoretical basis, the RFC technique outperforms the other techniques. The statistical, comparisons indicate the. RPC technique is most accurate for those cases where the actual RPC is small (less than 0.3). Perhaps more interestingly, because of the inconsistency with some other authors' nonsurvey comparisons, the performance of the other three techniques are virtually identical. For example, the RSRI RPC's root mean square error (RMSE) from the "known" RPCs is reported to be 0.223 overall versus between 0.60 and 0.601 for the other three techniques. For small "known" RPCs, the RSRI RPC's RMSE is 0.122 while the other 15 three techniques are between 0.694 and 0.706. For large “known" RPCs, the RSRI RPC's RMSE is 0.319 while the other three techniques are between 0.423 and 0.432. These results imply both that the RSRI RPC trade estimation technique is superior to the other three techniques and that there is virtually no performance difference between the three alternative techniques. It should be noted that the RMSE, will accentuate large differences more than. some alternative statistical measurements such as the mean absolute difference. Thus, the RMSE makes the performance difference between the RSRI RPCs and the other nonsurvey trade estimation techniques appear larger than alternative measurements. However, reported results from Theil's inequality index and regression results also lend evidence to the better performance of the RSRI RPCs and similarity of performance between the other estimation techniques. In any case, the empirical evidence on the superior performance of the RPC approach to estimating trade is meager, probably due largely to its recent vintage. There have been a number of articles (e.g., Garhart (1985), Ralston et a1. (1985), and Garhart and Giarrattani (1987)) that address the error generation created by using a single vector of RPCs to estimate trade (such as is done by REMI, RSRI, and, now IMPLAN) rather than a matrix of RPCs. The articles describe simulation experiments with survey-based models to demonstrate that use of a nmtrix of RPCs could improve the accuracy of RPC-based trade estimates. This issue is briefly .addressed in chapter 2. It has relevance for the comparison of different trade estimation techniques. However, the simulation results reported thus far are not very dramatic. For example, Garhart' and Giarratani C1987) report multiplier differences of less than fifteen percent. Such differences are not very significant when contrasted against differences 16 several magnitudes greater which were found stemming from other variables in this study. Furthermore, the issue is largely beyond the scope of this research study, which is to examine trade estimate used by IMPLAN (i.e., applied across rows). It is known for other reasons that many of the RPC values currently being generated by version 2.0 of IMPLAN are highly questionable. These RPCs may contribute to significant distortions in economic impact estimates, especially for particular industries, including many sectors affected by recreation. This conclusion was originally reached in the process of preparing economic impact estimates for a 1987 Lake States Governors' Conference on Forestry (Pedersen and Chappelle, 1988). It was based initially on comparisons of RPC estimates generated by IMPLAN to IMPLAN's estimates of output and demand for the same sectors, and on comparisons with RPCs derived from REMI models leased by state governments in the Lake States region. Table 1 presents some of the questionable IMPLAN RPC values observed for the State of Michigan. These RPC values were generated by IMPLAN' and. would influence impact estimates unless the model user changed them. IMPLAN estimates of the ratio of local production to local demand appear in the SDP column. The SDP value indicates the maximum potential value the RPC can attain, given IMPLAN's estimates of regional demand and output. The actual RPC may be well below the SDP ratio due to imports and exports. IMPLAN and SDP trade estimates are for 1982. 1985 REMI RPCs for the state of Michigan are also shown. Table 1 includes only' a 'portion. of’ the Michigan sectors with dubious RPC values. A pattern can be discerned of negligible value RPCs appearing in clusters of sectors. The negligible RPC values for the pulp Table 17 1. Michigan RPC and SDP Trade Estimates Intraregional Michigan Trade Estimates Paper Mills,exc Bldng Paper Paperboard Mills Envelopes Sanitary Paper Products Building Paper & Bldg Board Paper Coating and Glazing Bags,except Textile Bags Die-cut Paper and Paperboard Pressed & Molded Pulp Goods Stationery,Tablets & Related Cnvrted Paper & Paperbrd,nec Fabricated Rubber Prdcts,nec Misc Plastics Prdcts Rubber & Plastics Hose & Belting Pumps & Compressors Ball & Roller Bearings Blowers & Fans Industrial Patterns Power Transmission Equip Industrial Furnaces & Ovens General Industrial Machinery,nec Carburetors,Pistons,Rings,Va1ves Machinery,exc Electrical,nec Other Wholesale Trade Misc Repair Shops Svcs to Buildings Personal Supply Svcs Computer & Data Processing Svcs Management & Consulting Svcs Detective & Protective Svcs Equip Rental & Leasing Photofinishing,Commercial Photography Other Business Svcs Advertising Legal Svcs Engineering,Architectural Svcs Accounting,Auditing,& Bookkeeping,nec Auto Rental & Leasing Auto Repair & Svcs Auto Parking & Car Wash 0000000000000000000000000000000000000000 0000t-400t-‘t-‘00H00000000001-‘00000000000t-‘H00l-‘H 0000000000000000000000000000000000000000 18 and paper sectors (including sectors 188 through 198 shown in Table l) were of greatest concern for the 1987 Lake States forestry economic impact study. These industries account for the majority of forest product industry sales in the Lake States region. Their combined sales were in excess of ten billion dollars in 1982. Their negligible IMPLAN RPC values imply that virtually no Michigan demand for pulp and paper products is met by regional production, which is contrary to firsthand knowledge of the industry. Negligible RPCs for the service sectors listed (beginning with sector 461 through the end of the Table 1 list) are perhaps even more at odds with what is known about these sectors. Service industries tend to supply local markets and, overall, are likely to have higher RPCs than manufacturing industries. The magnitude of their RPCs should be expected to be closer to 1.0 (as the SDP and REMI values are for the sectors listed) rather than 0 (as IMPLAN's unchanged RPCs are for the sectors listed). An extensive examination of IMPLAN-generated RPC values was a major focus of this research in light of these and other observations which raised. concerns about IMPLAN's RPC trade estimates. The objectives related to this phase of the research are to identify and measure the extent of problems with IMPLAN's RPCs and to propose means of ameliorating these for IMPLAN users. Review and use of alternative RPCs is timely in terms of widespread use of the pc IMPLAN release. Although it: has implications for all IMPLAN applications, the RFC analysis relates directly' to the reliability of outdoor recreation economic impacts. It will be shown that many of IMPLAN's RPCs affecting recreation impacts are at odds with regional economic theory. 19 The 929229.12 2: Am as It Applies. t2 L9 Anguilla The degree of accuracy required for input-output estimates will depend, in part, on the purposes of the LO analysis. Input-output analysis is used by public agencies for at least two distinct purposes. The first of these is to convey a measure of the total or absolute impacts associated with some type of economic activity. The true purpose of such a use of LC analysis may be to justify the importance or budget of the agency associated with the activity. In such cases, the purpose is to use I-O as a descriptive public relations tool to convey an impression of the importance of a particular economic sector or activity. However, measuring total impacts associated with some types of activity may be appropriate and even required in cases involving major public expenditures of funds or uses of public lands. In these and other cases, the LO analysis may serve to better illuminate which sectors are affected by particular public or private actions. The second purpose is to use input-output as an analysis tool for economic development or industrial targeting. Here, the emphasis is likely to be more on establishing the relative merits of different sectors or public assistance strategies. Computing total impacts may not be as important as relative impacts; a comparison of select multipliers or industrial and institutional linkages may be what is needed by decisionmakers. Accuracy in input—output trade estimates and other I-O parameters is critical for reliable evaluations of differences between individual sectors. It is common to see authors borrow multipliers from other studies or to simply select a number (often 2.0) and multiply it by their estimates of direct sales (or income or jobs) to arrive at estimates of 20 "total" impacts. Generally, the authors are not claiming to be accurate in such cases and may even admit that their estimation procedures leave much to be desired. However, whether due to funding constraints or other priorities, unique multipliers for the time, space, and activities under consideration could not be calculated. Also, the authors may have believed it inappropriate not to mention that impacts extend beyond the direct impacts measured. Exaggeration. of impact size is a danger when. multipliers are borrowed, and such exaggeration could, in turn, contribute to widespread discounting of impact estimates and input-output analyses more generally. However, much recreation planning and many recreation analyses are not critically dependent upon the level of accuracy in the measurement of impacts. Borrowed multipliers may be used simply to indicate that secondary effects from recreation activity occur, rather than to illustrate the exact magnitude of expected impacts associated with the recreation activity being studied. In contrast, accurate impact measures are much more critical in industrial targeting and economic development studies which must assess the relative contributions or potential contributions of economic sectors. Improved accuracy of input—output estimates also may be addressed in the context of reporting results. Reporting of a range of estimates (even though the range does not constitute a true statistical confidence interval) relates more information and may more accurately convey the level of ‘knowledge regarding likely economic impacts than. a single number. Thus, despite what might appear to be a loss of precision, the reporting of impacts in ranges -- based on familiarity with the 21 variability in data and model parameters -- may be less misleading than a single number which falsely connotes a high level of precision. There has been only limited theoretical and applied work on stochastic and probabilistic I-O models. Jackson (1986), for example, has described the basis of what amounts to a probabilistic specification of technical and trade coefficients which would generate interval multipliers or impact estimates. Aggregations of industries are treated like samples of firms within an industry. Unlike the usual I-O aggregation, the information on the disaggregated industries' technical coefficients and trade are aggregated together into probability functions, weighted on the basis of output. The author notes that, in distinction from a model that is generated at least in part from random influences, his model takes into account "expected systematic variation" observed at the disaggregated data level. The accuracy of the disaggregated data is a critical constraint on the accuracy of such a probabilistic model. Also, estimation of final demands used to drive the model remain critical to the model's results. Computing hardware continues to advance with each passing year. Further development and applications of stochastic or probabilistic I-O models can be expected to accompany increased computing capabilities. Admittedly, decisionmakers may 'well prefer singular (point) values, which do not reflect any uncertainty, to interval or range estimates with probabilities attached to them which are more difficult to interpret. However, it remains the analyst's task to avoid oversimplifying or deceptive estimates which do not relate the level of uncertainty encountered. 22 Applieetion pf Valuation apd Impact Concepts pp Recreatien Impac; Apalysis a. Terminology The use of‘certain terms pertaining to impact analysis is not always consistent in the recreation literature. The dominant, conventional usage is to refer to impacts as "secondary" or "indirect" impacts. On the other hand, the experience felt by recreationists, or their utility from the experience, is generally denoted as the "direct" or "primary" effect. This orientation may stem from the perspective of benefit-cost analysis, which conventionally does not allow for a counting of indirect benefits (defined as market transactions associated with the recreation experience), to be counted as benefits. Thus, Walsh (1986) states, "Economists distinguish between the primary benefits and secondary' impacts of recreation. economic decisions...The net benefits of individual consumers represent the social benefits of public recreation programs. ‘The consumer surplus of individual users may not be spent in the region of the recreation site or spent at all, but this does not make it any less real to individual consumers. ...the secondary effects of the actual expenditures by individual consumers and managers of private and public recreation resources...are the regional economic impacts on business output or sales, employment, net income, tax. revenues, government spending, and environmental quality. The essential idea is that primary costs to individual consumers and managers become secondary gains, in part, to the regional economy supplying recreation goods and services. Studies of regional economic impact do not measure the value of the project to the primary users of the recreation site but rather the value of the project to those who are involved in supplying the primary users with goods and services." ..."The Water Resources Council guidelines recommend the regional economic impacts should be treated as income transfers in a separate account to distinguish them from benefits which contribute to general welfare or national economic development. Conceptually, employment anywhere in the nation of otherwise unemployed or underemployed resources that results from a project represents a valid benefit. However, they are not counted because of problems of identification and measurement 23 and because unemployment is regarded as temporary. The guidelines allOW' one major exception to the rule. If the regional economy of a proposed project has substantial and persistant *unemployment of’ labor, then the benefits of the project may include the income (salaries and wages) of otherwise unemployed labor working onsite in the construction or installation of a project or a nonstructural improvement. ..."Most secondary gains to a particular region will be offset by actual or potential losses elsewhere. This means that outdoor recreation programs redistribute income to the regional economy of parks and other recreation sites from other regions and the nation. Whether such redistribution is desirable is a political decision beyond the scope of economics. The essential point is that these changes in the distribution. of income represent transfers of income and not social benefits, i.e., not real welfare gains to the nation. What is a gain to the local region may be a loss to another region, and the national economic welfare may not change. Economists refer to such transfers of income as pecuniary impacts to distinguish them from technological impacts where real national secondary benefits occur in regions with long run unemployment, immobility of resources, and economies of large scale." (pg. 373-376) This last. paragraph. contains several. misleading statements. It would have been more accurate to note secondary gains to a region mey be _p,,1eeep, partially offset by actual losses in other (subnational) regions, rather than qualifying the first sentence by acknowledging that the losses "will be offset by actual or potential losses elsewhere." Are the gains and. losses 'perfectly' equivalent, such that the gains merely "represent transfers of income" and nothing else? At issue is whether the gains are identically matched by losses elsewhere (within the nation, or system of regions being considered), or only partially so. If regional gains exceed losses elsewhere, the net gain. could rightfully be included in an impact analysis of the benefits for the system of regions. This point is not addressed directly, although later in the same paragraph the author does state that gains "pey be a loss to another region" (rather than will be), and "national economic welfare 24 mey not change" (rather than will not). Generally, it is not known whether regional secondary gains are offset or not by losses elsewhere. Economics has a role in aiding the understanding of the nature of income redistribution so as to allow for more informed political decisions on the desirability of such redistribution. It is within the scope of economics to objectively measure the redistribution and predict _ its impacts. Most (but not all) economists make a distinction between pecuniary and technological impacts, but the distinction is not well conveyed in the last sentence of the last quoted paragraph. Pecuniary effects are monetary (income or wealth) distributional effects stemming from market transactions and changes in prices. Dismissing pecuniary effects as merely distributional impacts rests on the assumption that there are exactly equal gains and losses. This, in turn, is dependent on perfect competition assumptions, or at least full employment of resources in the markets under consideration anui any related (complementary or substitute) markets. In contrast, technological effects imply "real" effects on preferences or technological opportunities, presumably affecting aggregate welfare. Resource allocation may be affected in either case. Technological impacts may occur from many sorts of economic activity, including travel and tourism. An important issue is what is the spatial unit of analysis? New economic activity may create beneficial technological impacts in. a depressed area by positively affecting the region's long run unemployment, immobile resources, or untapped. potential for economies of scale. In contrast, the same activity may well have only distributional consequences in regions with 25 full employment, perfectly mobile resources and no remaining economies of scale to tap. As described, new wealth is created in the depressed area whereas there is only income being transferred in the latter regions (assuming regions with such conditions exist). "Real" national secondary benefits may be said to occur in the first circumstance, as opposed to merely "pecuniary" benefits in the latter. It is clear that outdoor recreation can generate "real national secondary benefits" by this perspective, but it depends on the region in which the recreation occurs. Problems remain with this perspective, however. Neoclassical economics, with its emphasis on the forces of equilibrium and efficiency, tends to ignore situations exhibiting long run ("structural") unemployment and factor immobility. Belief in the workings of Adam Smith's invisible hand could lead one to argue that public or private recreation expenditures in areas of long run, high unemployment may be less economically beneficial than expenditures in areas of low unemployment particularly in terms of price distortions, but also in terms of productivity. Expenditures in low unemployment areas might provide further competitive incentives for resources to be allocated to where they provide their highest return and lead to (or force) technological innovation. In contrast, expenditures in high unemployment areas may not generate the same intensity of incentives and may distort price signals, leading to a loss in efficiency. The above argument is oversimplified as it does not address a number of social welfare concerns, such as those pertaining to Second Best, inflation, and welfare payments issues. The point is that determining what to consider in regional impact analysis is not as 26 straightforward as described by Walsh. The basis for counting impacts in areas of high unemployment rests not only on the idea that the nation gains from using resources that would have been wasted otherwise, but also on the concept that better economies of scale may be achieved in such areas, leading to gains in efficiency. (The corollary for low unemployment areas is that there are no further possible resources to be exploited -- this ignores the possibility of importation and the concept of comparative advantage -- nor are there any further economies to be achieved). If there is any basis to the saying that necessity is the mother of invention, then conditions of high resource utilization are likely to lead to technological progress. In contrast, idle resources and slack demand do not generate incentives for innovation. It is not the purpose here to draw final conclusions about these points of view, but only to contend that the rationale behind accepting or rejecting the legitimacy of impact estimates is not perfectly objective. Guidelines on when to count impacts versus not allowing them are arbitrary and more political than economic. b. Residents Versus Nonresidents A related recreation impact issue consists of which recreationists to count. Recreation expenditures within a region ‘by the region's residents are sometimes dismissed because they are assumed to contribute nothing to the regional economy. According to this view, such expenditures are simply a transfer of income from one part of the region to another part. The gain to one area is viewed as perfectly offset by a loss to another area resulting in no net impact. This view ignores distributional consequences relating to the marginal differences in 27 impacts between different types of activities occuring within the same region. While certainly less significant than entirely new economic activities introduced to a region, changes in economic activity within a region are not the same as economic leakages out of a region. Thus, for example, while more recreation in Michigan by Michigan residents may amount to a loss in other parts of the Michigan economy, the new recreation impacts are not likely to exactly equal reduced impacts elsewhere from the change in spending habits. It is unrealistic to believe they exactly cancel each other out. To contend the impacts are identical is paramount to believing there is no difference in sectoral multipliers. This would, in turn, eliminate much of the justification for differentiating between economic sectors and industrial targeting programs. Additionally, some recreation expenditures within the region by the region's residents may be a form of import substitution if they take the place of recreation expenditures outside the region. However, accurately differentiating such expenditures would be difficult for it would involve measuring incomes, costs, and preferences over time. It would be difficult to measure marginal differences in impacts from residents of a region engaging in more regional recreation and less of other activities. A primary problem with such measurement would be to identify the economic sectors of the economy which "lose" from more regional recreation expenditures. Several alternative situations exist. .Again using Michigan as an example, any additional Michigan recreation- related expenditure that occurs exclusively in place .of out-of—state ‘recreation spending is ‘pure economic gain to the state. From the perspective of the state, it would be legitimate to count any impacts 28 from such import substitution and resulting "leakage" reduction. However, it is very possible and perhaps likely that some portion of any observed increase in Michigan recreation-related spending takes the place of other spending in Michigan, or comes out of savings. Tradeoffs occur in such cases. A complete description would account for impacts associated with losses in areas and sectors from which the funds are transferred from, and contrast these to gains in areas and sectors benefiting from greater recreation activity. One possibility for assessing "true" gains from such transfers of spending within a region. would. be to offset any increase in new recreation spending by an equal amount in other sectors in proportion to typical personal consumption expenditure (pce) patterns. (Savings and taxes could be figured into these too.) This could provide a comparison of economic impacts associated with recreation relative 11) average or typical. consumer' expenditure. economic. impacts. However, this blunt approach measures average differences rather than true marginal changes in consumption expenditures that would likely occur as a result of increased recreation expenditures. Trend studies of personal consumption expenditures and leisure activity, delineated by income classes, might shed better light on the tradeoffs that are made. These might be used to develop weights of sectors likely to experience reductions from additional recreation activity. c. Comprehensive Impact Analyses Conducting more comprehensive impact analyses is related to the issue of examining net impacts. Computation of actual‘ local economic impacts stemming from recreation activity should include an analysis of costs, besides recreationists' expenditures. This would foster a better 29 understanding of impacts and who is affected by them. It might also improve impact projections which would be useful for planning purposes. However, as Keiner (1985) notes, this point is often ignored in many impact studies which address only expenditures of recreationists. There appears to be a lack of appreciation that costs need to be included in order to present a balanced impact assessment. Other reasons that costs are often ignored include funding constraints, study priorities, and factors relating to the nature of the costs. Millerd and Fischer (1979, p. 248) review secondary benefits and costs that should be taken into consideration when calculating local economic impacts. Their list provides an indication of the complexity a thorough recreation economic impact assessment would entail. The following description of secondary benefits and costs is an elaboration of Millerd and Fischer's list. Secondapy Benefits 1) public expenditures for initial construction, preparation, and operation of recreation facilities; and other public expenditures induced by these (schools, roads, etc.); 2) recreationists' trip-related expenditures (e.g. lodging, meals, travel, etc.); 3) increased private investment due to recreation facilities (stores, motels, roads, cottages, etc. - including construction and operation of these, and their induced impacts). (Note that tourist spending may make the difference between the success or failure of many marginal businesses, especially in more rural areas); 4) external effects from roads/ transportation facilities (better provision of goods and services - for example, the increase in size of a community permitting a large discount store to open, etc.); 5) major equipment purchases: boats, recreational vehicles, etc. 30 01:8 1) lost income opportunities (for example, to extraction industries -- agriculture, timber, and mining); 2) increased local government expenditures -- medical, fire protection and police services for tourists (however, these may be offset by stimulated nonresident recreation home development and the property taxes these provide); 3) "various external costs" - local residents may substitute other forms of recreation, causing loss of local income opportunities; 4) price effects on goods and services locally (higher mark-ups on goods during tourist seasons tend to apply to local residents as well) - also possible substitution of imported goods in place of locally produced goods (locally produced crafts replaced by imported crafts). 5) environmental costs from ‘recreational use of local environmental resources. These range from litter through soil erosion and noise pollution. Some of these costs may be obvious and have explicit market transactions associated with them; others may not be very discernible nor have any readily apparent monetary values. The level of analysis detail described above and by Millerd and Fischer is seldom approached in outdoor recreation or other resource economic impact studies. It would require multiple data sources and tools of analysis; an input-output table alone would not be sufficient. The list (and similar ones like it) may serve as an ideal to strive for and provides public agencies with reminders of impact considerations. Heroic assumptions are often necessary to complete such comprehensive profiles of recreation.economic impacts. The value of some recreation 'variables may ‘never ‘be unambiguously determined if they relate to goods which have joint production costs, are nonexclusionary, or otherwise are produced, traded or consumed in conditions which violate the perfect competition model. Conditions necessary to achieve perfect competition have been enumerated in many different ways. Broader descriptions sometimes refer to well-defined, enforceable property rights, the absence of market barriers, and the absence of Second Best 31 conditions. At a minimum, most lists include perfect information; many rational buyers and sellers operating as price takers; perfectly mobile factors of production and homogenous, perfectly divisible goods. In the case of joint production costs, the same factors are used for producing multiple goods. This makes it impossible to objectively allocate the costs for producing the goods. The problem is compounded by the goods often being produced at the same time and by indivisible factors (such as often associated with fixed costs). Nonexclusionary goods or services are those from which, due to prohibitive costs or simply the practical impossiblity, persons cannot be excluded. In such cases, persons who benefit cannot be made to pay for their use of a good or service. Outdoor recreation often involves such goods and services. Examples include scenic vistas, appreciation of the presence of wildlife, and multiple uses of waterways such as for fishing, boating, and swimming. Durable recreation equipment and public costs of providing for recreation experiences are examples of difficult-to-measure variables that affect outdoor recreation impact estimates. As an analog to joint production costs, the economic impact measurement problem with durable recreation equipment might be viewed as joint consumption (purchase) benefits. Such equipment is often used in multiple regions and sometimes even for nonrecreational purposes. Some analysts have elected to include a percentage of the equipment's costs in their impact studies, based on the percentage of recreation trips made to a region or amount of time the equipment is used in a region. . However, such approaches are usually quite arbitrary and may be improper depending on a study's objectives. 32 A few examples will illustrate the influence of a study's objectives on the appropriateness of different techniques for measuring the value of, or impacts from, recreation equipment expenditures. If the analyst is projecting changes in recreation impacts, marginal impacts may' well involve a different pattern of durable equipment expenditures than the existing average pattern. Change in recreation activity patterns involve not only a change in the types of equipment used, but also a change in use intensity of recreation equipment. The intensity-of-use issue relates to whether new equipment, used equipment, leased equipment, borrowed, or existing (already owned) equipment will be used. Projected changes in equipment purchases should take these alternatives into account, if durable equipment purchases are an important part of the analysis. The size of the tourism-recreation industry in a particular region may be defined to include all local durable recreation equipment sales, whether the equipment is used in the region or not. Alternatively, if the objective is to measure the influence of a state's tourism promotion campaign, it may be important to attempt to isolate the proportion of recreation equipment sales to nonresidents only. Again, the appropriate point of view can only be defined in light of a study's objectives. In addition to the difficulty between allocating public costs of providing for recreation experiences between residents and nonresidents and recreation and nonrecreation purposes, public costs are multifarious. The more obvious costs to include in impact studies are those related to constructing and maintaining recreation facilities. Roads, sewers, fire and police protection in surrounding areas also are necessary for the provision of recreational experiences. Less obvious 33 are other public functions performed at the state and local levels which affect the recreation experience. For example, these include, but are not limited to, many aspects of planning and management that occurs within. state bureaucracies that address natural resource, commerce, transportation, energy, and environmental concerns. Portions of office budgets for these state bureaucracies, including administrative and support staff salaries, reflect costs created by a desire to generate and monitor recreation activity. Public costs for recreation are not substantially different from public costs for other activities. For example, there are substantial infrastructure costs associated with maintaining agricultural activity in rural areas, not to mention agriculture extension and other agriculture-related government personnel costs. In a similar sense, the activities and associated costs of a state's commerce department may be partly responsible for the expansion of economic activity 1J1 a particular industrial sector. Whether public costs for recreation are substantially greater, more diverse, or qualitatively different from public costs associated with other forms of economic activity is not as important as attempting to identify them and linking them to measured benefits. Federal recreation expenditures are generally perceived as an inflow of funds and a pure gain at the regional level. Taxes (or federal budget deficit) required for the funds are ignored (due again, in part, to the joint production-allocation of costs problem). State expenditures may often similarly be viewed as an inflow of funds for substate regions. 34 Study Epemises and Assumptiops Regazdipg the Influences pf Variables pp Impact Estimates This research examines the variability of economic impact estimates associated. with outdoor recreation. expenditures. Different data and methods are two primary causes of variability in outdoor recreation economic impact estimates. Data used differ in terms of: 1) the degree to which they can be unambiguously defined or quantified, and their accessibility and ease of manipulation, and 2) perhaps more importantly, their influence on results, in this case, the magnitude of the impact estimates. The same type of data may well vary over space, activity, or time. Different methods and models may also be employed to derive impact estimates. The choices made as to which data, methods, or models are used in impact studies are also influenced by the study's objectives, funding, and expertise of the researchers. There are economic impact variables whose values could be accurately determined if enough effort is devoted to the task. Variables which fall into this category include industry output, the number of affected employees and their wages, industry-specific deflators, producer margins, and trade. Even with these variables, however, aggregation errors and other types of measurement error not subject to statistical analysis (unlike sampling error which can be estimated by statistical analysis) may occur. The premises guiding this research are that: 1) sources of high variability in outdoor recreation impact estimates exist and can be identified, 2) it is possible to assess the consistency of this variability and the ease in using ranges of these variables in a sensitivity analysis of impact results, and 3) impact estimates may be 35 refined by efficiently utilizing information displayed by patterns of differences in impact estimates. More specifically, in. the case of outdoor recreation economic impacts, it is assumed that: l) a source of variability in existing estimates of economic impacts stems from differences in definitions (and, hence, sectoral aggregations) used to classify and measure outdoor recreation. This variability stems not only from what activities to count as outdoor recreation, but also from what spending is considered (e.g. residents or nonresidents only, trip expenditures only or durable equipment purchases too, and costs of providing for the recreation experience). 2) Alternative producer' margins and deflators used to convert recreation participation and spending information into input data for an input-output model will generate results which will vary by less than a factor of one from each other. This occurs because alternative producer margins and deflator values fall within a narrow range. Producer shares are generally in the range of 50 to 100% of purchaser prices. Deflators similarly will be expected to fall within 50% of each other, unless the deflation takes place over several decades or is for a period of high inflation like the 19705. (For this study, the deflators used only cover the period 1982-1985). However, checking and improving the precision of margins and deflators used may be accomplished with relative ease. Substituting more precise margins and deflators to achieve even slight impact precision improvements may prove to be an efficient use of time in impact analysis. 3) The magnitude of RPCs and the nature of the sectorization scheme will generally tend to have larger influences on impact estimates than 36 deflators or margins. However, the influence of RPC trade estimates and errors due to very aggregated sectorization schemes are less tractable in terms of their effect on impact estimates. The extent to which they differ from the "ideal" ("true" values in the case of RPCs and fully disaggregated schemes in the case of sectorization) will influence their relative magnitude of influence on impact estimates. RPCs, in particular, can be expected to have inconsistent, but potentially large influences (greater than a factor of one) on impact estimates. This supposition is made on the basis of the author's research for the 1987 Lake States Governors' Conference on Forestry (Pedersen and Chappelle, 1988). Also, RPCs are used to convert matrices of intermediate demands (in addition to final demands). This is in contrast to deflators and margins being used to convert a vector of final demands only (representing spending distributed across I-O sectors); they are exogenous in a demand-driven model. Therefore, RPCs influence calculations of indirect and induced impact components, whereas deflators and margins do not. In other words, errors in RPC estimation may be viewed as subject to being compounded by the multiplier effect. Study Objectives Refining allocations of recreation spending to input-output sectors and examination of IMPLAN's trade estimates are primary objectives of this dissertation research. Special attention will be devoted to the issue of RPC trade estimates, as they affect all IMPLAN models used to generate economic impacts, whether the impacts relate to outdoor recreation or any other economic activity. Also, because of its applicability as a case study, a detailed description will be presented 37 of the steps followed in preparing 1985 National Survey of Fishing, Hunting, and Wildlife-Associated Recreation data for use with IMPLAN. Evidence that the trade estimates in the IMPLAN modeling system can be significantly improved and the development of an detailed bridge for allocating recreation spending to IMPLAN sectors are two major products of this research. The study also has implications for several other issues, including: how variability in participation and spending profiles across recreation studies affects resulting impact measurements; how sensitive economic impact estimates are to certain types of errors; where the greatest gaps in data occur; and where improvements in data collection could be made. The study utilized only secondary sources of data to develop alternative trade estimates and outdoor recreation economic impacts. The only trivial exception to this is subjective opinions obtained from Michigan State University Parks and Recreation Resources Department faculty on minor questions regarding the likely nature of certain types of recreation spending. Other secondary objectives and steps followed in the process of achieving the study's objectives are as follows: 1) Document problems with RPCs generated by version 2.0 of IMPLAN. Construct alternative RPCs from a "corrected" 1977 Multi-Regional Input- Output (MRIO) accounts database and contrast these with IMPLAN. This comparison will have the purpose of demonstrating that IMPLAN's trade and impact estimates can be improved if this alternative database is used to re-estimate RPCs for the modeling system. 2) Compile 1985 upper Lake State data on fishing, hunting, and wildlife-associated recreation activity and spending. Most recreation 38 participation data used for the 1987 Governors' Conference on Forestry (Pedersen and Chappelle, 1987) came from Michigan, Minnesota, and Wisconsin State Comprehensive Outdoor Recreation Plans (SCORP) reports. These data. were combined. with spending profiles from .Minnesota to calculate total regional recreation spending. These spending data are contrasted to similar spending estimates generated from the 0.8. Department of the Interior, Fish and Wildlife Service's I28: NaeIepal m at” fishing. limiting. and W Pragmatics. 3) Compile alternative spending profiles. Recreation spending categories are bridged to IMPLAN sectors. This process consists of disaggregating much of the data and converting them to producer prices to conform with the LO accounting format. They are then deflated to 1982 values and serve as vectors of recreation final demands for use with alternative Lake State IMPLAN models. A range of spending profiles are constructed, reflecting published statistical information and consideration of different types of expenditures. Low and high estimates of' spending; are developed for trip expenditures only, all spending including durable equipment, all spending except durable equipment, all recreationists, and. nonresidents only. These estimates take into account sampling errors only. Additional nonsampling errors which could compound ‘problems associated with sampling errors are usually not directly measureable. They are not addressed in this study. 4) Construct alternative IMPLAN models. The Lake State models will reflect different sectorization schemes and trade estimation assumptions. Alternative sectorization schemes allow one to measure the influence of aggregation error on estimated impacts. Models for the Lake State region (Michigan, Minnesota, and Wisconsin) are the primary focus 39 of the analysis. Models of the three individual Lake States and several counties in Michigan (Kalamazoo, Kent, and Ottawa) serve to test the general applicability of observations regarding IMPLAN's RPCs. 5) Estimate recreation economic impacts, using the final demands derived in step 3 with type I and type III multipliers derived from the Lake State IMPLAN models. Compile total output, personal income and employment economic impacts. 6) Contrast estimates of total economic impacts. Variables examined include activity participation, spending patterns, and LO model sectorization schemes, producer margins, and trade estimates. The objective is to measure the reliability of existing recreation and I-O model data, construct a range of estimates in which actual (real world) values likely exist, and derive a corresponding range of economic impacts. (True statistical confidence intervals for the recreation economic impacts cannot be constructed as they represent a synthesis of multiple sources of data without known probability functions). 7) Report results, consisting of: a) comparisons ‘between. alternative sets of activity estimates, spending profiles, margins, sectorization schemes, and RPCs, b) summary descriptions of resulting changes in multiplier-based impacts from the use of different values for these variables; b) estimates of economic activity associated with particular definitions of outdoor recreation in the Upper Lake States. Attention to alternative outdoor recreation definitions must be given, as this study relies on two alternative secondary data sources' with different delineations of recreation regions and activities. CHAPTER I I INPUT-OUTPUT ANALYSIS 6: IMPLAN Introduceien This chapter presents background material on input-output analysis and the USDA Forest Service's input-output model, IMPLAN. For a good reference text on 1-0, see Miller and Blair, Ipppt-Qtpue Analysis; Eeengegiep§_ egg, Extensions (Prentice-Hall, Inc., 1985). Major publications on IMPLAN include the IMPLAN fleep'e Geige and Apelysle G_ufie (U.S. Department of Agriculture, 1983 and 1985, respectively), More current IMPLAN materials may be available from the IMPLAN Development and Applications Group, Agricultural and Applied Economics Department, University of Minnesota, St. Paul, Minnesota. Input-Output Analysis Input-output analysis (LO) was developed by Wassily Leontief in the United States during the 19305. I-O can be used to measure effects felt throughout an economy when output of one or more sectors (industries) are increased or decreased. More precisely, this impact analysis tool allows for computation of direct, indirect, and induced effects associated with changes in final demand on an industry-by- industry basis. Final demands refer to consumption sectors of the economy and include government and household institutions and investment, inventory, and export accounts. They involve transactions after which there is no further processing within the region. LO tables are mathematical representations of economies. Through a system of linear equations, they serve as both accounting frameworks and impact analysis tools. Sales (receipts) of industries are recorded 40 41 across rows while purchases (expenditures) are recorded down columns. Sales are divided into intermediate and final demands, while purchases are divided into intermediate and final payment categories. Although different symbols are sometimes used, these relationships are often depicted by the following notations: i-ith row sector, j-jth column sector, nrnumber of sectors in model; Xi-total output (sales) of sector 1, Xj-total outlays (purchases) of sector j; xij is the output of sector 1 purchased by sector j; Yi-final demands of sector 1, consisting of C1 (personal consumption), Ii (here, investment, including inventories), G1 (government purchases), and Ei (exports). Final payment sectors may be depicted by: VJ-total primary inputs (value added and imports) of sector j, consisting of L (personal income or payments to labor), P (property income), T (indirect business taxes), D (depreciation), and M (imports). Rows of intermediate and final sales (receipts) may be expressed as followS' n z (x11+ xiz+ ... + x1" + c, + I1 + G, + E1 = z x1 i=1 i=1 n n n where x1 .121: J'l xlj +H: Y1’ and Yi =i§1(cl+ II + GI + E1) Columns of intermediate and final payments (expenditures) may be expressed as follows: n 2 (x13 + x21 + ... + x"J + 1.1 + I1 + G1 + E 1) . 2 xJ 3'1 - jsl n n n where X 1.1;1 1.1 xii + jflvj’ and V = 2 L + P + T + D + M 33.1(1 :1 .1 J 1’ A helpful accounting feature of an input-output table is that it is balanced; total gross outputs equal total gross outlays. Thus, 42 n 11X 2 X *2“. . 1:11 jglj The pattern of sector outlays depicted in the columns can be used to derive a set of fixed, linear production functions after all transactions are accounted for and recorded ‘between sectors. The portion of sector j's purchases attributable to sector i is called a technical coefficient and is noted as: The basis for input-output analysis can now be formulated, using the relationships and definitions presented. First, it should be noted that the level of intermediate purchases may be derived by taking the matrix of aij's and multiplying them by the vector of total gross outputs, Xi' Dropping the subscripts on vectors to allow for more convenient notation and beginning with the initial condition that total gross outputs are equal to intermediate and final demands, the following manipulations allow total gross output to be derived from knowledge of final demands and technical coefficients: X - AX + Y X - AX - AX - AX + Y X(I - A) - Y x - Y (I - A)’1 where I is an identity matrix with ones along its diagonal and zeros elsewhere. The (I-A)'1 matrix is called the Leontief inverse, after the pioneering economist who was in large part responsible for developing input-output analysis. Multiplying the Leontief inverse by a vector of final demands will produce estimates of output levels required throughout all sectors of an economy to exactly meet the final demands. 43 Subsequently, the output estimates can be used to generate projections of income, employment, value added or other economic variables by using historical information on sectoral ratios of these variables to output. Nppeppyey I-O; Adaptation pf National CoeffieIents for Regional Models Several techniques have been developed through the years to avoid the expense associated with constructing a complete survey-based I-O analysis. Most techniques adjust national level I-O coefficients to the region being analyzed. Employment data are often used to make extrapolations from the national to the regional level, despite problems associated with the practice, because of the ready availability and frequent reporting of employment data. Problems with such extrapolations include: 1) regional and temporal productivity differences exist; 2) employment data used are often based on employment for one date (in March for County Business Patterns), thus masking seasonal differences and. not necessarily representative of an annual average; 3) different mixes of full and part-time employment are reported for different sectors and exist in different regions, thus making extrapolations to sales (like the issue of productivity) questionable; 4) self-employed persons and certain other categories of workers (owners and administrative personnel) are not reported, underreported, or not reported as working within a particular sector in the same manner as other employees; and, 5) disclosure and aggregation issues, affecting comparisons of any data across regions and the nation, affect employment as well. Regional scientists are divided in opinion over the efficacy of nonsurvey techniques. The Brucker et al. article (1987) reviews five "ready-made" I-O model systems. The article, and subsequent comments on it, attests to the growing use of nonsurvey I-O models, despite 44 reservations about their use. Three nonsurvey techniques are briefly described below. See Miller and Blair (1985) or Richardson (1978 and 1985) for a more complete review of the numerous approaches available and perspectives on their performance. V’aTNLocation Quotients ' The location quotient (LQ) is generally specified as follows: ei / er e? / en where ei - regional employment in sector 1 er - total regional employment ”\F eis- national employment in sector 1 . J“ ef3- total national employment Alternatively, the mathematical equivalent of this is sometimes specified as: e: / a? er / en The location quotient for any particular industry indicates the relative share of local employment the industry accounts for vis-a-vis the national industry share of employment. If the industry accounts for a larger share of employment on the regional level than it does on the national level, the 101 will be greater than one. Conversely, if the industry comprises a smaller proportionate share of local employment than the national industry share, the LQi will be less than one. Many variants on the use of location quotients to transform national coefficients into regional coeffients have been devised. Fbr the Simple Location Quotient technique, if the L01 21, then the aij's for the ith industry are used. If the LQi <1, then the aij's for the 45 ith industry are adjusted downward by the value of location quotient. The‘basic idea behind this approachis that imports will have to occur if the industry is not as present on the local level as it is nationally. On the other hand, excess output is assumed to be exported (rather than consumed in intermediate production) if the industry comprises a larger share of local employment than nationally. This follows from interpreting the aij's as technical coefficients and the assumption that regional production processes are the same as national production processes. -\. ' ijSupply-Demand Pooling ' ./ iI/The supply-demand pooling (SDP) (or commodity balance) technique flows from the simple assumption that, given transportation costs, demand will first be met by local production. Simultaneously, sales of regional output will first go to meet local demand. Thus, imports will only occur after local production is exhausted and exports will only occur after local demand is met. The SDP technique is also sometimes termed a net trade approach. Both imports and exports will not be allowed to occur for the same industry, with the balance between local demand and production determining which will take place. I c. Regional Purchase Coefficients \ The original regional purchase coefficient (RPC) approach combined Census Transportation data with other secondary data econometrically to derive regional trade estimates. (The 1983 articLe by Stevens, Treyz, Ehrlich, and Bower, is one of the earliest descriptions of this nonsurvey technique; their approach will be referred to as the STEB approach.) A number of alternative specifications have been used by the 46 originators of the RPC approach to trade estimation. The changing of the specification for the RPC estimating equation and the resulting changes in estimated RPCs has caused some degree of consternation for REMI model users. On the one hand, it certainly is worthwhile to improve the specifications used, so as to enhance the reliability of the REMI models. On the other hand, it can be embarrassing to not be able to generate consistent forecasts and. impact estimates. An emphasis on determining RPCs on the basis of the relationship of regional to national values has remained throughout changes in the RPC specifications. The reduced log-linear form of the RPC equation indicates this emphasis: RPC§ - b0 / (FE'i/F?’1)bj; Fr or n,i - variable j, for commodity i in r or n where r - region under consideration n - U.S. b0 - a constant bj - elasticity of response of RPC? to a change in the ratio of variable j The first step in the STEB RPC technique involves deriving RPCs for manufacturing industries from the following equation: RPCi - (XE/DE) Pi where ix; - the amount of commodity i produced within the region (based on Census of Manufacturers data), D§ - the amount of commodity i demanded within the region (based on BEA I-O data and other demand calculations), Pi - the amount of commodity 1 produced within the region that is also shipped locally (based on the Census of Transportation Commodity Survey). 47 (Note that Pi can 'be expressed. as Xi’r/Xi, where Xi’r represents regional output: shipped locally (intraregionally). substituting this formulation of P1 into the RPC equation, the Xi's cancel, leaving Xi’r/Dr, the ratio of intraregional shipments to total demand, or RPC.) The derived RPCs are then used as dependent variables in a regression equation from which all other RPCs are derived. Richardson (1985, p.623) notes that the following equation has been used, based on fitting an initial sample set of RPCs: 1chi - K(w§/w‘i‘)b1 (eE/e?)b2 (W?/V§‘)b3 (LQi)b4 (Ar/An)b5 + e where w - wages, e — employment, W - tonnage of shipments, V - value of shipments, A - land area, i is a subscript for industry, e is an error term, and r and n are superscripts for region and nation. Fitting the equation for 2-digit SIC manufacturing sectors (and adding seven other dummy variables for particular 2-digit SIC sectors), Stevens et al. (1983, p. 279) report significant t-values for all the independent variables in the equation and an R2 of .679. IMPLAN IMPLAN is maintained by the Forest Service at the U.S. Department of Agriculture's Fort Collins Computer Center. The National Environmental Protection Act of 1970 provided the impetus for impact studies. The Rangeland Renewable Resources Planning Act of 1974 and National Forest Managment Act of 1976 provided further incentive for the development of IMPLAN. IMPLAN is an input-output (LO) model with associated data bases. Input-output analysis can be used to measure effects felt throughout an economy when output of one or more sectors (industries) are increased or decreased. More precisely, this impact analysis tool allows for computation of direct, indirect, and induced 48 effects associated with changes in final demand on an industry-by- industry 'basis. Final demands refer to consumption sectors of the economy and include government and. household institutions and investment, inventory, and export accounts. They involve transactions after which there is no further processing within the region. IMPLAN's data base contains a detailed national interindustry table and estimates of final demand, final payments, gross output and employment for each county in the U.S. A data reduction technique (the RAS method, which is an iterative, balancing process) is used to develop state and county estimates of value added and final demand. These data files can be combined with U.S. interindustry data to to form regional input-output models consisting of aggregations of counties and/or states. Appendix A presents a complete listing of IMPLAN's 528 sectors and their corresponding Standard Industrial Classsification (SIC) codes. The data in version 2.0 of IMPLAN represents 1982 economic relationships. The national interindustry table used in IMPLAN is based on 1977 U.S. input-output tables (U.S. Department of Commerce, 1984) updated to 1982 through the RAS method and related techniques. Use of more current national tables would be desirable to reflect changes that have occurred in the national economy since 1977; however, the 1977 tables are the most current detailed national I-O tables available (through April, 1990). Miller and Blair (1985, p. 266-316) review evidence regarding the stability of technical coefficients. They interpret the evidence as indicating that, while coefficients change over time,‘ "for aggregate kinds of measures ... the error introduced by using an ‘old' table may not be large." (p. 273). Most of the studies cited by Miller and Blair 49 compared I-O tables that are between four to ten years apart. Nonsurvey models (such as IMPLAN) using the 1977 U.S. I-O tables are facing a gap in excess of twelve years. This larger, more recent time period probably includes greater economic changes than in the smaller period covered by the I-0 tables in the comparison studies. Particular sectors have experienced very dramatic changes since 1977 (e.g., computers, service sectors generally, and foreign trade). In this light, Miller and Blair note that larger errors were often found when particular sectors were considered rather than aggregate measures. (p. 273). Thus, a need for a more current set of national tables exists for nonsurvey models such as IMPLAN, depending to some extent on which sectors are involved in its applications. However, IMPLAN and other nonsurvey modelers have little control over when new national I—O tables will be published. Use of IMPLAN involves running several of the model's modules, including Region, Accounts, Symmetric, Lister, Smash, and Invert. These are described in the IMPLAN User's M and AnaIysie gape (U.S. Department of Agriculture, 1983 and 1985, respectively). The titles of these modules are not all readily apparent (nor are they important) when operating the pc version of IMPLAN. The different steps allow the user to delineate a. region, estimate regional economic activity through combining regional and national data with the use of data reduction techniques, aggregate and name sectors, and derive estimates of multipliers and impacts for the sectors specified. Numerous descriptions of IMPLAN applications have been published. Two publications of interest here are ”Using Socioeconomic Data in the Management of Fishing and Hunting on Public Lands" (Alward et a1, 1985) 50 and "Opportunities for Analyzing the Economic Impacts of Recreation and Tourism Expenditures Using IMPLAN (Alward and Lofting, 1985). mmmmumm Earlier versions of IMPLAN used the supply-demand pooling approach for trade estimation. Version 2.0 incorporates a modified RPC approach. The IMPLAN adaptation of the RPC approach is described in an unpublished document by Alward and Despotakis (IMPLAN Version 2.0: Data Reduction Methods for Constructing Regional Economic Accounts, no date). Derivation of IMPLAN's RPCs will be described first, followed by a discussion of how the IMPLAN approach differs from the Stevens RPC technique and initial evidence of problems with the IMPLAN RPC values. a. Derivation of IMPLAN's RPCs For IMPLAN, the fitted model is given as: law’i‘r/ x?) - b0 + b11n(w‘i-’) + b21n(e§/ a?) + b3ln(LQi) + b41n(Ar/ A“) + e where Mgr - imports into region r from domestic (U.S.) sources, XEI - output produced and consumed in region r b0 - a constant (although different intercept terms are used for different sectors, to parallel Stevens, et a1.) w; - wage in region r for industry i ei/ e? - ratio of region to national employment by industry L01 - the location quotient for industry i Ar/ An - ratio of region r land area to total U.S. land area e - an error term A weight-to-value independent variable is not used as unique RPC equations are constructed for each separate commodity sector. Note too, that the lefthand, dependent term is neither the RPC nor the SDR, but rather a region's ratio of domestic imports to output shipped locally. 51 Also, 1982 data is used for the independent variables, while 1977 MRIO data is used for compiling terms for the lefthand side of the equation. The RPCs for region r and industry 1 are then calculated as: RPCE- 1 _WI'—TT—"MT_IT 1+Mi/Xi +Mi/x1 where Mwi - foreign (non-U.S.) imports to the region, and ng / Xir is assumed to be a constant b. IMPLAN Differences from the Stevens RPC Approach Two basic differences between the Stevens RPC estimation approach and the approach adopted for the IMPLAN modeling system are the different databases the approaches used to develop RPC values for their initial dependent variables, and the actual dependent variables being estimated. Other differences exist between the two approaches, such as the use of different independent variables, however, most other differences tend to relate to these two differences discussed below. Also, some other differences noted in the Alward and Despotakis paper have 'been. eliminated through evolution in the STEB RPC estimation technique (e.g. foreign trade is accounted for in Treyz and Stevens, 1985). l. Dependent Variable being estimated The IMPLAN regression equation actually estimates the natural log of the ratio of domestic trade to intraregional trade, which is a component of a subsequent RPC calculation. The STEB regression equation estimates the natural log of the RPC directly. 52 2. MRIO vs. Transportation Census Both databases are of 1977 vintage. The 1977 MRIO data constitute a complete set of U.S. multiregional accounts for the fifty states and the District of Columbia. It is based largely on secondary data, including the Transportation Census, but also numerous other sources that range from readily available to quite obscure. Other than the conversion of this data base from a port-of-entry to a contribution trade orientation, RPC estimates can be derived directly from it for use in a regression equation. (Foreign imports and exports are attributed to the state in which they first are unloaded or loaded in the port-of- entry approach. The contribution approach allocates total national exports and imports according to a state's proportionate share of demand for imports and output for exports.) STEB needed to use a number of data sources to compile initial RPC sample values. However, most of these data sources are well known governmental data sources that have track records and are published with descriptions of their statistical accuracy. The MRIO database, on the other hand, has not been extensively reviewed. It is known that budgetary constraints prevented some proposed data collection and reconciliation steps from being conducted, and that, as a result, it contains numerous gaps and inconsistencies (Multi-Regional Policy Impact Simulation Project, 1988). Therefore, the Stevens, et a1. database is suspected to be more reliable than the original 1977 MRIO database. There are a number of minor errors in the Alward and Despotakis draft report on IMPLAN's RPCs. These errors may be important because the report is the only documentation on the derivation of IMPLAN's RPCs and much of the report is devoted to critiquing the STEB RPC estimation 53 technique. For example, it is stated that equation 98 provides a ratio which could be used for comparison purposes with the STEB goodness-of— fit measure; however, the proposed ratio's numerator and denominator are exactly identical, making the ratio equal to one. The statement is made on the following page that, without an additional constraint, ”...the estimated RPC may indicate gross Ippeppe (underlining added) exceeding the production of a commodity in an application for a particular region." Gross imports may well exceed regional commodity production; an accounting problem arises when gross exporps exceed regional commodity production. Errors also exist in the report's "Appendix A: The Estimation of Regional Gross Trade Flows -- A Literature Survey." One of the more important of these errors is the contention that the STEB approach overestimates RPCs due to the manner in which local demand (Di) is estimated. It is correctly noted that 1972 BEA U.S. National I-O technical coefficients, incorporating, imports, are used by STEB to estimate demand, thus overestimating domestic requirements. This is in line with the conclusion to chapter two in the main text which states, "...the main weakness of the STEB approach is the inconsistency between the definition of RPC and the treatment of foreign trade." However, if Di is inflated due to the inclusion of imports in the aijs, then the RPCs are underestimated as Di appears in the denominator of the RPC equation (RPCi - (xi/Di)Pi ), and a larger denominator will reduce the RPC value. The conclusion to the STEB article states that the bias of the RPCs is towards underestimation: "In closing, it should be reiterated that most manufacturing RPCs for most states are somewhat underestimated by 54 the RPC estimating equation. As previously noted, the 1972 Census of Transportation failed to report shipments moving less than 25 miles. A preponderence of such shipments would be to destinations within the state of origin, so the percent of output shipped within each state is underestimated 'by an amount that will vary among commodities and states." (p. 284). The authors then note the 1977 Census will include "short shipments" and that future RPC equations will be based on the 1977 data. Ralston et a1. (1986) offer another view on the-RPC bias issue. They contend that the use of one RPC for an industry (per row in an I-O table), as is done in the STEB and IMPLAN models, leads to inaccuracies, including possible overestimates of multipliers. However, their evidence is a Delaware I-O model estimated by the supply-demand pooling method. More generally, Stevens (through RSRI) and REMI have adopted several changes in their approach which eliminate many of the concerns (including those about foreign trade) expressed in the Alward. and Despotakis report. about the STEB technique (see Treyz and Stevens, 1985). Finally, Alward and Despotakis report that it was assumed non- port. states (those ‘without foreign 'borders) have no foreign. import values in the MRIO database. This probably contributed to errors in calculations of IMPLAN RPCs. A review of the "corrected" MRIO data -- and tables prepared from them -- indicates many nonport states have foreign import values.) I c. Evidence of Problems with IMPLAN RPCs IMPLAN models generated by this author for the states of Michigan, Minnesota, and Wisconsin, a three-state model of these states, and 55 several Michigan county models, all contained dubious zero-valued RPCs as well as many near-zero RPCs, reducing the size of estimated impacts. The percentage of sectors affected was between ten to thirty. There appeared to be some consistency as to which commodity sectors were affected across the various IMPLAN models. Initial questioning of certain RPC values was on the basis of IMPLAN reporting a zero or near- zero RPC, despite output for the sector being sizeable relative to demand for the sector (as indicated in a large SDP value), and on the basis of comparisons with REMI RPCs. Nonsurvey Input-OuEput Apalysls Model Valldaeion Nonsurvey I-O model validation involves three issues: what is to be measured in the validation process, what it is to be measured against, and what measurement tools to use. How to interpret the measurement results could. be considered. a fourth issue. There is an extensive literature on the subject of nonsurvey I-O accuracy. Over two dozen regional science articles focused on this topic have been published in the last fifteen years. Articles which provide an overview of the work in this area or comment on alternative measurement tools include: Sawyer and Miller (1983); Jensen and McGaurr (1977); Morrison and Smith (1974); Butterfield and Mules (1980); Harrigan, McGilvray, and McNicoll (1981); Round (1983); and Richardson (1985). Comments accompanying the Brucker et a1. 1987 article on "ready-made" I-O models also reflect opinions on the topic of nonsurvey I-O accuracy. A primary choice of what is to be measured has been between cells of’ technical coefficients “versus multipliers (generally output multipliers are used). This choice is related to Jensen's (1980) S6 distinction between "partitive" and "holistic" accuracy, whereby the former refers to cell-by-cell accuracy and the latter the general accuracy of the table as a whole. Partitive accuracy is much more exacting, while achieving holistic accuracy would more modestly demand that an L0 model merely ”represent the size and structure of the economy in general terms" (Jensen, 1980, p. 143). (Jensen goes so far as to note that "Partitive accuracy in regional input-output tables, with existing data sources and research resources, is not an achievable goal." (Ibid., p. 143)). More recent attention to assessing the accuracy of trade variables, such as the size of imports and exports, or regional purchase coefficients, could presumably be characterized as falling between the extremes of partitive and holistic analysis. Such analysis of only one aspect (trade) of the I-0 table is generally conducted on an sector-by- sector basis, rather than cell-by-cell. This is in line with the fact that most nonsurvey methods have applied regional adjustments to national technical coefficients on a row-by-row basis. However, the need for partitive accuracy in trade estimates has been raised by Garhart (1985), Garhart and Giarattani (1987), and Ralston, Hastings, and Brucker (1986). These authors have contended that regional purchase coefficients should be determined on a cell-by-cell basis rather than applied across entire rows. Theoretically and intuitively their argument has appeal. It is highly unlikely regional demand for a particular commodity will be exactly equal across all sectors in a region, but that is the implication of "rows-only" RPCs applied by IMPLAN and the STEB approach. One reason mitigating against exactly equal RPCs for a particular sector 57 is linked to aggregation error. Any sector in an I-O table is actually comprised of different, but related entities. Various sectors' purchases from a particular sector are often actually purchases of different products and vary to the degree they are likely to be imported. One place to look for RPC differences is between intermediate demand and final demand sectors, just as there are often differences between goods produced for industrial use versus those for consumption in the home. For example, households purchase pick-up trucks while industrial sectors purchase a variety of other trucks, yet the LO model may have one aggregated truck sector. Also, the capacity and, therefore, propensity to import differs by buyer (as well as by size of region and other factors). Therefore, I-O sectors will naturally differ to the extent by which their demands are met by local supplies (as reflected in RPCs) for any particular sector. Garhart (1985) and. others have reported simulation and survey results which they interpret as lending support to their position that RPCs should be determined on a cell-by-cell basis. However, their results, to this author, do not appear overly compelling. Significant differences in RPCs across rows have been described, but the effect of these differences on multipliers does not appear to be very dramatic. For example, Garhart and Giarratani (1987) conducted simulations of errors introduced into a survey—based state of Washington LO model. They report mean absolute percentage errors under fifteen percent in multipliers from using rows-only RPCs instead of cell-by-cell RPCs. I believe accuracy within fifteen percent for measuring total impacts or individual sectoral impacts would generally be adequate for most LO applications. 58 IMPLAN software allows the user to change cell-by-cell RPC values. However, the issue of row variation of RPC values is not addressed in this study, as the basis for making cell-by—cell distinctions on a completely secondary basis is lacking and beyond the scope of this research. Further research in this area might involve adjusting the rows-only RPCs to cell-specific values by ad hoc assessments of industry characteristics or possibly using some nature of weights based on secondary data (such as the U.S. Transportation Census). With regards to what is to be measured in model validation efforts, it also should be noted that attention has seldom focused on evaluating the appropriateness or range of final demands that are used to "drive" the LO model. There have been many articles written on differences between survey and nonsurvey models but few on the variability of input data which serve as final demands for the models. It is a tenet of this study that model validation for practical LO applications involves examination of the input data at well as review of the LO model's accuracy. Whether the choice has been to measure nonsurvey I-O model technical coefficients, multipliers, or trade variables, the standard against which to measure these has been, with few exceptions, estimates from survey-based models. (The few exceptions involve simulation experiments that have measured the results of introducing varying percentages of change to trade or technical coefficients.) Concerns raised about such comparisons include that survey models are frequently rather aggregated, are out-of-date (approaching several decades in age), and contain data gaps or other sampling errors and problems. Despite 59 these concerns, most efforts at evaluating nonsurvey methods only compare nonsurvey model values against survey model values. The approach used here to evaluate IMPLAN's RPCs will assess the reasonableness of RPC values in light of alternative secondary data sources (Census data, for example) and their real world implications. "Reasonableness" will be imputed through both mechanistic means and by economic concepts. (The choice of the term "concepts" here is not arbitrary; most judgment on the RPCs will stem from common sense or have a basis in regional economic principles, but it may be argued that the latter are not well-developed enough to warrant "theory" status.) There are no perfect means to assure RPCs developed from secondary sources will conform precisely with true RPCs; in fact, "true" RPCs cannot ever be known with absolute certainty. However, there are several subjective means available to evaluate the overall reasonableness of RFC estimates. One such means is to compare them within and across regions in conjunction.‘with 'knowledge of the structure of different types of markets and how regional economies operate. As part of this process, it is helpful to contrast one set of RPCs with RPCs developed through alternative methods or from different data sources. This illuminates the implications of the RPC values and enables judgements as to which RPCs are more tenable both for specific sectors and as a set. A wide variety of measurement tools have been used to measure the accuracy of nonsurvey methods. A partial list of the more popular of these include the Theil inequality (or information) index, regression analysis, the chi square statistic, and correlation coefficients. Additionally, a number of simple comparison measures have been reported which relate absolute or relative differences between nonsurvey and 60 survey tables or multipliers. These are variations on simple percentage differences, including such measures as ~ the mean absolute percentage error (sometimes referred to as the average absolute percentage error or "MAPE"), the standardized mean absolute difference (or "SMAD"), and the root mean square error ("RMSE"). Opinions conflict as to which of these comparison measures are most appropriate, or if any of them are viable. For example, Miernyk (1976, p. 49), referring to a set of nonsurvey multipliers versus a set of survey-based multipliers, states "there is no way to statistically test the significance of the differences between the two sets of multipliers." A frequently cited problem in this regard is violations of assumptions necessary for the statistical measures (such as known population distribution frequencies and independent observations). For example, Boster and Martin (1972, p. 40) report results using the Wilcoxon signed-rank test, contending, "In analyses of this type, nonparametric techniques (as in the Schaffer and Chu study) have a clear advantage over parametric techniques." However, Round (1983, 19. 202) states, "Unfortunately, the Wilcoxon test is also inappropriate, again because the basic assumptions of the test are violated" (independence between the variates being measured). Similarly, Theil's inequality index is apparently chosen as superior to others by Stevens and Trainer (1980), Park et al. (1981), and Stevens et a1. (1986), but rejected by Garhart (1985) because of its questionable interpretation. Another problem with some statistical tools for measuring nonsurvey I-O accurécy is with zero cell values in 1-0 tables. There usually are a large number of cell values of zero, particularly in I-O survey tables. These create problems for measures that place such zero 61 values in a denominator (e.g., the chi square statistic and the standardized mean absolute difference) or that give equal weight to them as to other nonzero cells with the result that measurement of differences are arguably reduced (e.g., regression analysis). There are at least three reasons to use multiple statistical tools in this type of analysis. The first is, as several authors have noted (Harrigan, McGilvray, and McNicoll, 1980 and 1981; Butterfield and Mules, 1980; and Round, 1983), that there is no one "best" measure for contrasting the differences between two vectors or matrices. A related second reason is that different statistical tools will place an emphasis on different factors (for example, variance versus skewness), thus reporting multiple comparison measures will help avoid bias in the analysis. A third reason is to enhance comparisons with previous results reported in the regional science literature. Formerly a major reason not to use multiple comparison measures is the time involved in developing such measures; however, time is less of a factor with widespread use of computer spreadsheets and inexpensive statistical software. For these reasons, despite the slight redundancy due to similarity in the results from several statistical measures, a number of comparison tools will also be used for this study, specifically for measuring the differences between RPC estimates and the corresponding effects of these differences on multiplier estimates. Emphasis will be placed on measuring the relative influence of alternative RPCs on total outdoor recreation impact estimates. Measures of the differences in RPCs and multipliers for all economic sectors may not be the same as measures of differences between sectors affected by particular economic activity, such as outdoor recreation. In this 62 regard, it is important to measure how a model performs in terms of policy or study objectives. The particular objective here is accurate measurement of outdoor recreation economic impacts. a. Procedures for Evaluating IMPLAN's RPC Values This thesis devotes a disproportionate amount of space to consideration of IMPLAN's RPCs due to the topical nature of the RFC issue and the opinion expressed frequently in regional science literature (including by Richardson (1985), Stevens and Trainer (1980), and. Park et a1. (1981)) that trade estimates are critical to the accuracy of L0 models. RPCs can be evaluated on the basis of comparisons with other sets of RPCs and in terms of the RPCs' influence on impact estimates. They can also be evaluated on the basis of related primary or secondary data that provides implications as to what the magnitude of the RPCs should be. These alternative means of evaluation provide different kinds of information regarding RPC reliability, accuracy, and significance. In light of these considerations and the multiple objectives of this thesis, there were four major steps in the analysis of IMPLAN's RPCs: l) conduct comparisons across sets of RPCs from different sources (IMPLAN, REMI, MRIO); 2) check consistency of RPCs with well-established sources of data (specifically, 1982 Census and County Business Patterns data) and any available industry-specific studies; 3) measure consistency of RPCs from one source (IMPLAN) a) internal consistency across sectors within one region b) internal sectoral consistency across regions (states and substate regions); and 4) measure the impact of alternative sets of RPCs on resulting multipliers and estimated outdoor recreation economic impacts. 63 There are numerous statistical tools which could be used in these four steps, as indicated in the above descriptions of measures used for evaluating nonsurvey methods. For the most part, the analysis will employ relatively simple comparison measures (mean absolute percentage error, the goodness-of—fit measure - R2, or other standard measures of absolute and relative differences). Exceptions are that the Theil inequality index also is used to measure differences in RPCs and chi square also is used to measure differences in resulting multipliers. These exceptions are made to permit more extensive comparisons to past nonsurvey measurements. Detailed procedures for these steps are described further in the Methods chapter and results are presented in subsequent chapters. Measurement pi Outdoor Recreation Ecenopie Ippaets Economic impacts have ‘been. estimated. for ‘particular recreation activities, sites, park systems, and states. There have been numerous journal articles on outdoor recreation economic impacts, mostly appearing in the travel and tourism or recreation literature, and at least half a dozen Ph.D dissertations on some aspect of the subject. The approach here is relatively unique, in that it measures how outdoor recreation in the forested areas of a multistate region impacts the entire multistate region. Studies of outdoor recreation economic impacts using economic base theory or input-output analysis were first conducted in the 19505. Two of' the earliest: publications addressing ‘recreation. economic impacts remain important texts today: Clawson and Knetsch's Economice pf mtdoog Recreation and Tourism gig Recreatien, by Arthur D. Little, Incorporated. These publications compare the sectoral distribution of 64 site-specific outdoor recreation spending for nineteen and sixteen different studies, respectively, in addition to reviewing many other dimensions of recreation and associated economic impacts. At about the same time the Clawson and Knetsch and Little books were being written for a national audience, three books were published which address outdoor recreation economics issues affecting the Lake States. These include fie Developmept ef Qitdoor Reepeatiop 111 513 Upper Migwest, (Lodge, 1964), Resources epgi Recreatiop in ehe Noreherp Great Lekes Region (Minnesota Department of Agriculture, no date), and he Reopomice ef Quedoor Recreapiep ip the upper, Midwest (Sielaff, 1963). Several recent examples of state efforts to develop comprehensive profiles of travel and recreation are available. The Council of State Planning Agencies’ report, "The Contribution of Outdoor Recreation to State Economic Development" (Keiner, 1985) reviews a number of these. Holecek (1985) has proposed ongoing monitoring of tourist spending by county in Michigan and for the state as a whole based on extrapolations from lodging sales and use tax data. This approach forms the basis of O'Halloran's 1988 Ph.D. dissertation. Another recent study by Massoud Ahmadi, described in the report, "The Economic Impact of Tourism in Maryland: A Multiregional Analysis," (no date) used the same input- output model (IMPLAN, version 2.0) proposed for use here. Ahmadi's report describes the development of tourist profiles for eight subregions of Maryland as a result of participation in nine recreation activities and use of seven types of accommodatidns. The IMPLAN-based model developed was a 48-sector, interregional model, linking the eight subregions and a model for the overall state of Maryland. 65 In recent years, Michigan, Minnesota, and Wisconsin have had increased interest in the economic growth potential of recreation activities within their borders. New studies are being generated and there is more state-sponsored data collection. Two recent reports from Michigan are representative of this trend: menu MiebigepLs wmmlwfi- Meanwwea 0 Elaniaizeanflsfil Recreation in Etc—luau“. users. Mira. maternal and guaranties (no date) and Travel ml Touriem in Miehigep; A Statieticel Eroiile (Spotts, editor, 1986). The Recreation Rlep was prepared by Michigan's Department of Natural Resources as part of Michigan's 1985 State Comprehensive Outdoor Recreation Plan (SCORP). Travel e_I_1_d_ Tourism was funded by Michigan's Department of Commerce. Wisconsin and Minnesota also 'have recent SCORP reports indicating, type, frequency, location (state subregions), and. other aspects of recreation use. Michigan, Minnesota, and Wisconsin SCORP reports contain recreation data for the early 19805. The primary advantages to using the SCORPs is that there is some degree of conformity in definitions across the three states' reports and they contain recently compiled information by state agencies. This author has been involved in two previous research efforts related to estimating outdoor recreation impacts that used IMPLAN in conjunction with recreation data. The first was undertaken as a project for an MSU course, Resource Development 960, Simulation Models in Natural Resource Management. This effort involved using a computer spreadsheet to combine estimates of specific recreational activity spending patterns with sector specific deflators and multipliers for the State of Michigan. The sectoral spending patterns were based on a study 66 conducted by the Regional Science Research Institute (RSRI) for use with the U.S.D.A. Forest Service's IMPLAN model (version 1.1). (See "Tourism Expenditure Translators for 'Use in .Measuring the Regional Economic Impacts of Recreation on Forest Service Lands" by Benjamin H. Stevens, 1984.) The deflators and multipliers were generated using version 1.1 of the IMPLAN 'model. The computer spreadsheet allowed. the user to estimate recreation economic impacts based on selecting different levels and types of fifteen. different outdoor recreation activities. Adjustment in any of the program parameters (deflators, multipliers, and sectoral spending allocation) could also be performed by the model user. A major drawback to the spreadsheet model developed is that recreation participation data for the State of Michigan and most other states are not currently collected in as much detail as the categories developed by RSRI and incorporated in the spreadsheet. This limits the usefulness of the spreadsheet model as a practical tool for planners and decisionmakers. A second major effort by the author consisted of estimating ”wildland" recreation impacts in the upper Lake States for the 1987 Lake States Governors' Conference on Forestry (Pedersen and Chappelle, 1988). Following a literature search for recreation participation and spending data, total recreation spending was computed based largely on participation levels published in SCORP reports and reported spending patterns observed in Minnesota. These were used in conjunction with an IMPLAN (version 2.0) 48-sector model to generate economic impacts. This analysis of outdoor recreation economic impacts provided a profile of outdoor recreation in the more heavily forested areas of the Lake States. (The estimation. process and results are described in 67 Pedersen, et al, 1989.) However, as with most other recreation studies, there are several limitations to the analysis. Public costs associated with providing the recreation experience were not assessed, as these were not analyzed for the other two forest resource uses (forest products and wood energy) either. Little was done in the course of the analysis to check the consistency of estimates across data sources and overall sensitivity of results. Another concern with the resulting economic impact. estimates is the arbitrary' nature of' the “wildland recreation" definition (both spatially and activity-wise). A third concern regarding the economic impact estimates stems from the IMPLAN model's initial RPC values. During construction of this model, RPC values were changed for approximately 10% of the original 528 disaggregated sectors. Initially, these fifty-plus sectors, and others as well, had RPC values equal to zero. Changes in RPC estimates were made on the basis of REMI RPC estimates (Treyz, 1986), and subjective assessments regarding the likelihood a sector had an RPC significantly different from zero. Thus, for example, many service sectors with zero RPCs were changed to conform with REMI estimates, while some mining sectors with zero RPCs were left unchanged. As noted in chapter 1, the most troubling zero RPC values were for twelve pulp and paper sectors, which account for the largest share of forest products economic activity in the Lake States. Economic impacts would have been underestimated if the RPCs for the 50-plus sectors were not changed. It is not known how frequently IMPLAN generates unwarranted zero RPC values, or other RPC values that are significantly' at. odds ‘with actual real-world. RPC values. However, similar sectoral patterns of clearly erroneous RPC estimates were 68 observed for three individual state models generated at the same time the three-state regional model was constructed. Also, "Table 2. Observed RPCs for States," of Appendix G -- Regional Purchase Coefficients, in the Micro IMPLAN Software Manual (Alward, 1989) reports values for service and government sectors at the state level used in the pc IMPLAN model. A number of the reported RPC values have questionable zero values. If a pattern can be established, then the source of the estimation errors may be more easily identified and ameliorated. At a minimum, IMPLAN users may more readily avoid generating lower impact estimates than are warranted by knowing which particular sectors deserve attention. Secondary Sources pi Recreation Qeee fie; flee ip Impact Analyses The National Survey pf RishingI fluntipg, epe Wildlife-Aseociated Reereaeien (Fish and Wildlife Service, U.S. Department of the Interior) is a major source of recreation participation and expenditure data. The 1985 National Survey data will be used in this study and contrasted to SCORP data. Recreation data used in developing National Forest plans for the three states (U. S. Department of Agriculture, Forest Service, unpublished Recreation Information Management System (RIMS) data, no date) was rejected as being too unreliable for the 1987 Governors' Conference on Forestry. The quality of these data appears to differ from one National Forest to another and over time. Another source of information is Public Area Recreation Visitor Surveys (PARVS) data. However, the only nonproprietary published PARVS report as of summer, 1989 is a report of limited relevance to this dissertation, prepared for the TVA (Cordell, et a1. 1987). Two other national sources of data are 69 Charles R. Goeldner's Travel Trends ip pbe United States epel Cenada (Goeldner, Charles R. and Karen P. Duea, 1984) and Spatistice pp Outdoor Reepeatiop (Clawson, Marion and Carlton S. Van Doren, editors, 1984). The 1984 edition of Travel Irends is the seventh in a series of informative reports on state travel statistics. Some additional information is available from states and universities in the form of county level or specific recreation activity studies. Many of the latter are summarized for the state of Michigan in "Travel and Tourism" (Spotts, 1986). MSU's Park and Recreation Rescurces Department has been involved in a number of specific recreation activity studies, such as research into boating and marinas (Stynes, 1983). Examples of multifaceted, detailed county level reports include "The Economic Impacts of Recreation-Tourism: St. Croix County, Wisconsin" (Rose and Cooper, 1986) and Preliminary Results of Summer, Fall, and Winter (Recreation) Surveys: Tri-County Tourism Research Project (Spotts and Mahoney, 1985) . CHAPTER III METHODS Introductiop This chapter describes the variables examined and the methods used to analyze research results presented in chapter 5. Chapter four will present a detailed description of the steps followed to prepare the data and models used in the analysis. Rxepinatiop pf IMPLAN's Regionel Rprcpase Qoeiiiciepte The extent of problems with IMPLAN's RPCs were documented first. Alternative RPCs were subsequently constructed and compared to IMPLAN RFCs. Problems with IMPLAN's RPCs are documented through illustrating their inconsistency and dubious values within regions, across regions, and through comparisons with data both internal and external to IMPLAN. Much of the RPC evaluation relies upon subjective impressions of likely industry' and regional trade characteristics. These are guided by regional. economic. concepts. Also ‘personal communications ‘with individuals familiar with the database used to generate IMPLAN's RPCs have confirmed the database was flawed and likely to generate inappropriate trade values. MWQMEEELQLLM The consistenoy of dubious RPC values for particular sectors was examined through the construction of IMPLAN models for three counties in Michigan (Kalamazoo, Kent, and Ottawa), the states of Michigan, Minnesota, and Wisconsin, and a Lake State region model consisting of these three states. Confirmation of a pattern of zero or near-zero RPCs, 70 71 despite sizeable sector output relative to demand, was made on the basis of visual inspection of RPC estimates for these IMPLAN models. REMI RPC values were available for comparison purposes for the Kalamazoo County and Michigan models. A correlation analysis was conducted. between. REMI's and. IMPLAN's Michigan. RPCs. This analysis indicates the extent of correlation between the two sets of RPCs for various groupings of industries, both with and without suspect zero value IMPLAN RPCs. The groupings include major (SIC l-digit) industries and all industries for which RPCs were available. The goodness-of-fit measure (R2) was used to establish overall patterns in the comparability and consistency between the REMI and IMPLAN RPCs. Although evaluation of the goodness-of-fit measure is subjective, this author would interpret R2 values above 0.5 as indicating relatively good correlations between the sets of RPCs and R2 values below 0.25 as indicating little correlation. As it turned out, no R2 values exceeded 0.25, as indicated by Table 5, page 118, and the discussion of the correlation analysis in chapter five. Develpppent apd Qomperieop pi Alternative 32p; Miernyk. (1976) and. others have argued against overreliance on mechanical nonsurvey LO techniques. In their development of the RPC technique, Treyz and Stevens considered other approaches but adopted a "subjective" approach for estimating RPCs for non-manufacturing industries (Treyz and. Stevens, 1985). 1977 Census of' Transportation data, upon which their RPC estimation technique for manufactured goods relies, were only reported for manufactured goods. . More recently, Jensen (1988) and others have focused attention on "holistic" descriptions of economies using input-output tables. These 72 authors have contended that certain economic structures are predictable across regions, based on comparisons of different regions' I-O tables. Of particular relevance for the analysis of RPCs here is their finding that tertiary activities tend to be found across all regional economies and appear to be fundamentally universal in economic structures. If this is the case, then service sectors and their RPCs should be fairly uniform. Four approaches were considered which could generate more informed (yet still subjective) judgments regarding the parameters of RPCs. These four approaches would utilize sources of data other than REMI or IMPLAN. For example, it was thought that state tax data might possibly differentiate between in-state and out-of-state sales, such that estimates of exports (both domestic and foreign) on an industry basis could be derived. However, initial investigations indicated this "backdoor" approach to calculating RPCs probably would not be productive, at least in the case of Michigan. Sales and use taxes are not always attributed to the industry selling the product, but rather are sometimes reported by and attributed to the industry buying the product. According to Treasury Department officials, there are no tax records that reflect the level of sales or exports in any kind of systematic fashion across industries. Thus, sales or export extrapolations from Michigan Treasury Department tax records would be highly unreliable. Another approach examined whether the problem could be with output estimates rather than strictly the RPC estimation. . This approach contrasts IMPLAN and REMI estimates of output by sector with published 1982 Census and County Business Pattern (CBP) data. 1982 Census data was 73 not available in time for use in the 1987 Governors' Conference on Forestry study. Had the 1982 Census data been available, it would have permitted identification of the source and resolution of some REMI and IMPLAN RPC discrepancies. Use of Census and CBP data may establish on an industry-by-industry basis whether the source of RPC inconsistencies between IMPLAN and REMI stems from RPC estimation alone or arises largely from constraints imposed by REMI or IMPLAN estimates of supply. Differences in RFC values may be due to dramatic differences between REMI and IMPLAN output or demand estimates. This can be seen by recognizing total regional output for a sector, XE, is the ceiling value for intraregional trade, Xi’r. (Sales of local production to local demand cannot exceed local production sales.) In turn, by definition, the supply-demand ratio (SDR) functions as a ceiling value for the RPC. In this light, it would not be surprising if REMI and IMPLAN RPC estimates differ significantly if their estimates of output are dramatically different. Preliminary analyses indicated output estimates did play a role in the discrepancies between the two sets of RPCs for certain sectors. However, interpreting which output estimate was more accurate remained a problem due to aggregation and disclosure issues. This can be illustrated through a specific example, REMI and IMPLAN RPCs for water supply (SIC 494) and sanitary services (SIC 495). The 1983 REMI Michigan RPC for a ”Water Supply and Sewer Systems" sector was 0.03. This RPC value is constrained by a REMI estimate of the supply to demand ratio (SDR) for this sector being equal to .03. With demand being estimated at $75 million, this implies supply must be equal to approximately $2.25 million (-.03 x $75 ndllion). IMPLAN estimates 74 for what appears to be the same Michigan sector (#459, Water Supply and Sewerage Systems") are $442.245 million for demand, $365.32 for output, and an RPC of .6157, implying $272.29 million of demand is met by local production. Obviously, one or both of the output estimates ($2.25 million and $365.32 million) are grossly inaccurate (as may be the demand and RPC estimates). Census data for 1982 does not exist for the Transportation, Communication, and Utilities sectors (industries which fall within the SIC 40-499 codes, which includes water supply and sewer systems); however, 1982 County Business Patterns for the State of Michigan reports employment estimates for these sectors. The CBP estimates between 0 and 19 employees were employed in the water supply sector (SIG 494) and 1,835 persons were employed in the sanitary services sector (SIC 495), based on March 12, 1982 employment records. This does not include administrative and auxiliary personnel of which the CBP reports there were 1521 for all of SIC 40 through 499. Based on the water supply and sanitary service sectors accounting for less than 2% of the other employees (about 1850 of the 131,064 total), it is reasonable to assume between 15 and 150 (roughly 1% to 10%) of these administrative and auxiliary ‘personnel could. be associated. with the water supply and sanitary services sectors. Therefore, combined employment in the water supply and sanitary services sectors can be approximated to be between 1850 and 2000. Annual payroll for the 1835 sanitary services employees is reported to be $36.194 million, or very close to $20,000 per emplOyee. Including administrative and auxiliary personnel, and water supply employees would increase this figure slightly, to result in a rounded annual payroll 75 between $36.5 to $40 million. This payroll range would cast serious doubts upon the REMI production figure of $2.25 million; it is much more in line with IMPLAN's $40 million total income estimate for the water supply and sewer systems sector. However, the 1835 employees and $36.194 million includes all of sanitary services, SIC 495; the CBP data does not breakout sewerage systems, SIC 4952 which is the only portion of sanitary services REMI and IMPLAN include in their water supply and sewer systems sector. It is not readily apparent what portion of the SIC 495 employment or payroll should be attributed to SIC 4952. REMI and IMPLAN aggregate the remaining portion of sanitary services (that which is not in 4952) with other sectors, including steam supply, (SIC 496), irrigation systems (SIC 497). This becomes the "sanitary services, steam supply, and irrigation systems sector." The additional sectors have four to forty-four employees. REMI's RPC for this sector is .97; IMPLAN's .6157 RPC remains the same as for the water supply and sewer sector. In this case, the dramatic differences between REMI's RPCs casts doubt upon one or both of them. Consideration of the type of product (water supply and sewer service) associated with the sector also influences the evaluation. The IMPLAN .62 RPC is viewed as much more reasonable than the REMI .03 RPC for the water supply and sewer systems sector, as this author knows of no out-of-state projects responsible for meeting the vast majority of Michigan water and sewer services demand. If RPC values are imposed by erroneous output estimates operating as constraints on the RPC values, this may constitute a more serious problem for the I-0 model than when the RPC value is merely a product of the RPC estimation equation and output estimates appear sound. Other 76 facets of the model may be affected in those cases where it is found that output is the source of the RPC discrepancy. Multiplier, employment, and income estimates, and related ratios of output per worker, may be more seriously affected. (This has been illustrated with regards to IMPLAN sectors 461 (other wholesale trade) and 462 (recreational related retail trade). IMPLAN users and staff have noted these sectors acquired much lower output and value added estimates in all regional data files during the course of the development of the pc version from the mainframe version. The low values generate very high erroneous multipliers. This analysis avoided the problems by substituting original mainframe values for sectors 461 and 462 in all state data files.) Census and CBP data were used to develop "best guess" RPC estimates for both Michigan and the Lake States. The data were largely used to provide direction in choosing between alternative, widely divergent RPC estimates. Census and CBP data were taken into consideration more in the formulation of alternative Michigan RPCs than Lake State RPCs due to time constraints and the difficulty of working with three states' data. A third approach to generate more informed judgment on RPC values was to conduct a search for prior studies on industry trade flows and secondary trade data for specific industries. It was believed reliable information gained for even a few sectors could serve to establish a performance pattern between alternative sets of RPCs. Unfortunately, no industry-specific data. was found. that could provide reliable trade information. , ‘ Examples of important .Michigan industries which have ‘been the subject of extensive research include agriculture, forest products, and 77 automobiles. Michigan agricultural trade estimates were obtained from the Michigan Department of Agriculture and Michigan State University's Agricultural Economics Department. However, these proved to be based on simplifying trade assumptions rather than primary data. The same proved true of forest product industry data. The Michigan Commerce Department routinely revises REMI RPC estimates for automobile related sectors, based on their data collection and knowledge of the auto industry in Michigan. However, to this author, their revised RPCs appear to be as questionable as the unrevised estimates. Table 2 contrasts unrevised. Michigan motor vehicles and equipment sector 1985 REMI RPCs with 1983 REMI RPCs which have been revised by the Michigan Commerce Department: Table 2. 1983 and 1985 REMI Michigan Motor Vehicles and Equipment RPCs Sl_ Sector l983 l98§ 3711 Passenger Motor Vehicles .93 .44 3713 Truck and Bus Bodies .41 .48 3714 Motor Vehicle Parts & Accessories .90 .52 3715 Truck Trailers .23 .71 The revised 1983 RPC for passenger motor vehicles implies 93% of Michigan demand for passenger motor vehicles was met by Michigan production. This would probably be an overstatement even thirty years ago, let alone today with a larger foreign share of auto sales and auto plants distributed around the U.S. The same skepticism applies to the motor vehicle parts & accessories RPC. (It is noteworthy that the REMI unrevised 1985 RPCs are higher than the 1983 RPCs virtually across all sectors, with the exception of these two motor vehicle RPC values.) The lower truck trailer sector RPC is probably an improvement over the 1985 78 RPC of .71; one would expect it might be in line with truck and bus bodies (below .5). As there are fewer truck production plants, the RPC for both of these truck sectors may be more easily analyzed by a state commerce department than the other two sectors. The fourth approach was to closely examine the 1977 MRIO data used as the basis for constructing IMPLAN's RPCs. Alward et a1. (1989) report observed MRIO RPCs for service sectors that have been adopted in IMPLAN as state RPCs. These are reported for all fifty states and Washington, D.C. Means and standard deviations were calculated for each sector and each state. These were calculated both with and without zero- RPC values to examine whether problems were associated only with the zero-RPC values, or if problems existed with the remaining RPCs after the zero-RPCs were removed. ”Corrected" MRIO data was obtained for the purposes of examining the IMPLAN RPC estimates. The "corrected" MRIO data was prepared by the Multi-Regional Policy Impact Simulation (MRPIS) project of the Social Welfare Research Institute at Boston College (1988). Due to budget constraints, several gaps and inconsistencies are known.tx> exist with the original 1977 MRIO data prepared by Jack Fawcett Associates for the Office of the Assistant Secretary for Planning and Evaluation, U.S. Department of Health and Human Services (1983). The "corrected” MRIO data represent an effort by personnel at Boston College's Social Welfare Research Institute to eliminate MRIO data inconsistencies and balance the accounts. Adjustments to the MRIO data were made in consultation with the Jack Fawcett Associates staff who collected the original data. The MRIO data is compressed in seven computer files. It may be decompressed into seven ASCII files, between 2.3 and 3.3 megabytes each, 79 which contain 1977 use, trade, and make matrices for 50 states and the District of Columbia. The matrices have detail for 124 sectors which have a perfect correspondence, although at a more aggregated level, with IMPLAN sectors. The documentation accompanying the computer files is brief, but adequate. Crosschecks of accounting identities performed with the data have verified its consistency. The corrected MRIO data presented this study with several possibilities. An obvious task would be to retrace the estimation of the IMPLAN’ RPCs. Three difficulties prevent this, ‘however. The first problem. is that the input data used. in the IMPLAN RPC regression equations for the MRIO commodity sectors (84 of the 124) are not available. Coefficients (which apply across all 51 regions) are published for these sectors as are the actual "observed" service sector RPCs, but the input values of the independent and dependent variables are not. Therefore, it would not be known where mistakes occurred in the original RPC estimation process even if the regression equations were re-estimated. The second issue is that, regardless of the availability of the IMPLAN data, it would require a massive compilation and crosschecking of data to repeat the RPC estimation process. Essentially, it would involve manipulating approximately 10 or more units of data for 124 sectors in 51 regions (the fifty states and Washington, D. C.) (see the discussion of the IMPLAN RPC estimation process in chapter 2). The third problem is that the RAS procedure has been applied to both IMPLAN and corrected MRIO data, such that it would be difficult if not impossible to retrace exact values. It should be noted, however, that although the RAS procedure may change values, it should not be the 80 source of grossly conflicting, RPC values between IMPLAN and those derived from the corrected MRIO data. Fbr example, the RAS procedure cannot be held as the basis for the difference between RPCs for those service sectors where the corrected MRIO RPC is calculated as close to 1.0 and IMPLAN has a zero value. The RAS procedure may be viewed as a smoothing technique which does not drastically change overall patterns. For example, only minimum necessary modifications are made as the RAS procedure adjusts the A matrix to be in accordance with sectoral sales and purchase sums. (See chapter eight of Miller and Blair (1985) for detail on the RAS procedure). Thus, despite the RAS procedure being used for both sets of RPCs, one would expect there still should be a nonrandom similarity between IMPLAN's RPCs and RPCs calculated on the basis of the corrected MRIO data. Only limited uses of the corrected MRIO data were undertaken, due to time constraints. RFCs were calculated from the corrected MRIO data for all service sectors for the fifty states and the District of Columbia. The service sector means and standard deviations were compared. with the existing IMPLAN service sector RPC estimates. Average state service sector RPC values were used to rank order the states and evaluate the reasonableness of the RPC values. The multiregional nature of the MRIO data was also utilized to compile new' MRIO commodity and service sector RPCs for the three individual Lake States and the Lake States region based on summing Michigan, Minnesota, and Wisconsin intraregional trade and demand data. These were then contrasted. visually with IMPLAN RPC estimates to determine if IMPLAN's questionable zero RPCs were present in the corrected MRIO data set. Next, disaggregated IMPLAN and REMI data for 81 the state of Michigan were aggregated to conform with the MRIO 124 sector scheme and. contrasted. The aggregated data included output, demand, imports, exports, and. RPCs. Absolute RPC differences and revised MRIO percentage differences from IMPLAN RPCs provide the basis of the contrasts. Calculation of RPCs from the corrected MRIO data required adopting the contribution assumption towards foreign imports and exports made by Alward and Despotakis in their original IMPLAN RPC fbrmulation. According to the contribution assumption, for each MRIO sector, a share of total U.S. foreign imports and exports are assigned to a region (the three individual states or the Lake States) corresponding to the region's proportional share of national demand (for allocation of imports) and production (for allocation of exports). (In its unaltered state, MRIO foreign imports and exports are attributed to the "port-of- entry" state from which goods or services are shipped from the U.S. or through which goods first enter the U.S.) One of the issues that can be addressed in this process is whether it is legitimate to assume there are no foreign imports in the MRIO accounts for states which do not have foreign borders. This was assumed in the estimation of IMPLAN's RPCs. Foreign imports were supposedly attributed to states on the basis of their port-of-entry in the MRIO accounts; however, this does not appear to be the case across the board. As noted in the Alward and Despotakis draft paper on the IMPLAN RPC estimation process, after equation 90 (the text does not have numbered pages): "Special consideration was given to states with no foreign borders. For such states the diagonal elements of the MRIOA trade-flow xuf : Tug - xmf. matrices already correspond to Furthermore, interstate 82 flows between pairs of such states did not involve any imports to or exports from the US. Such flows therefore did not change as a result of the rebalancing process." The text following equation 72 makes a similar reference to the assumption that states without foreign borders have no foreign trade in the MRIO data base. However, a review of the corrected MRIO data by this author reveals many states without foreign borders have foreign imports credited to many sectors. The existence of foreign imports influences the size of RPCs. Therefore, the calculation of IMPLAN's RPCs may have been partially distorted by not taking into account foreign trade credited to states without foreign borders.) Lake Stage Outdoop Repreapiop Economic Impepts The major components of a recreation economic impact estimation process include the identification of: 1) activities (date, type, locality, and quantity) and recrea- tionists (e.g. nonresidents versus residents) to be counted, 2) recreation activity spending, bridged to the appropriate impact model sectors; 3) the appropriate sector-specific deflators and input-output model or multipliers to combine with the spending, in order to generate economic impacts. The identification of activities and recreationists to be counted should reflect the impact measurement objectives of the study and is done relatively independent of an L0 model, although the quantities estimated of these obviously influence the input data for the I-0 model. Data on recreation spending by sector form a bridge between recreation activity surveys and I-O models. However, spending categories included 83 in surveys are often developed independently of the I-0 model and rather aggregated spending categories are frequently used in the interests of survey brevity and achieving higher response rates. This tends to result in mismatches between recreation survey data and the Standard Industrial Classification (SIC)-based sectorization schemes of most I—O models. Thus, for example, surveys may not delineate between food purchased at restaurants and bars versus food purchased at grocery stores. The result of using conflicting classification schemes is that compromises are made to match the survey data to L0 sectors. The consequences of such compromises are largely unexamined and are generally not discussed in the reporting of recreation impacts. A number of other I-O variables also influence the ultimate impact estimates besides the sectorization scheme and its match with the survey spending categories. These include the deflators used to adjust the input data to the year of the 1-0 model and the trade estimation, or more broadly, the nonsurvey and reconciliation techniques the model employs. A related concern is how well the model corresponds with other secondary sources of data planners are likely to use. Veriablee Examinep Outdoor recreation economic impacts are sensitive to the influence 'of a large number of variables. Many considerations could be used for selecting which ones are the most important to test in a sensitivity analysis. The basis for selecting variables for this study were: 1) The variables had to be of potential importance in terms of ultimate impact results and of importance with regards to outdoor recreation impact analyses, and 84 2) The variables had to be either of general topical interest (RPCs) or could be relatively easily incorporated into the study (alternative sectorization schemes). Two objectives of this analysis are to examine the rangeof values _9:. alternatives within five sets of variables and to measure the relative influences of the variables on impact estimates. The five sets of variables examined are: recreation activity and spending levels, sectorization schemes, bridging of spending to IMPLAN sectors, and 1-0 model trade estimates. Much of the discussion and results relating to sectorization schemes, bridging of spending, and 1-0 model trade estimates apply to I-O analysis generally, although they receive particular attention here in the context of recreation impact analysis. Consideration was also given to examining the use of alternative deflators. However, the impact of alternative deflators is dependent on the number of years covered; for this analysis it is only three years, from 1985 to 1982. Hence, deflators were not extensively considered and will only be briefly addressed. The remainder of this chapter describes the approach taken for assessing the five sets of variables and their influence on Lake State outdoor recreation economic impact results. A variety of measurements and test statistics will be used in the analysis. Multiple means of measurement are necessitated by the difference in nature of the sets of variables and because both the range of data for the different variables is being examined as well as their influence on impact estimates. Also, there is a degree of arbitrariness as to which comparison measure to use, as some stress the importance of extreme values, while others stress the degree of ‘variance in the data or other factors. The 8S statistical measures used in this study are standard measures discussed in most statistics and econometrics textbooks (e.g., Pindyck and Rubinfeld, 2nd edition, 1981). Generally, relative differences between estimates will be expressed as percentages and unchanged IMPLAN parameters will be the basis for the measurement. A. number of summary comparison statistics will be employed for relating differences between RPCs and multipliers. In addition to means and standard deviations, other statistics referred to will include the mean absolute difference, the standardized mean absolute difference, the root mean square error, the correlation coefficient (r), the chi square statistic and Theil's inequality index. Formulas for these summary statistics are as follows: Mean Absolute Difference (MAD): where i - 1,2...n and x - variable of interest (RPC, multiplier,etc,) * Standardized Mean Absolute Difference (SMAD): %-2 Ixi - XII 1 n * Root Mean Square Error (RMSE): 1-2 x1 - xi {1 Z (x* _ x )2 "i "T, n 1' 1' Correlation Coefficient (r): COV (x?) (xi) ....1r________ * 2 ox ax Chi Square (x): 2 (x1 - xi) "1 * Theil's Inequality Index {1 2 (x1 - x.)2 (or information coefficient): n 1 $9212 ”F «92 The Theil index may be further subdivided into the following "proportions of inequality," which relate the source . of differences between two sets of variables: _* - 2 (x -X) 2 bias, or U : 1: * - n 2 (xi xi) variance, or U : * 2(1 - p) ox 0* covariance proportions, or U : 1 * 2 Z (Xi - Xi) n The ‘bias ‘proportion. is indicative of the amount of systematic error, as it measures how much mean values of the variables differ from each other; The variance proportion reflects the degree the two variates' variances are similar. The covariance proportion measures any remaining error. These proportions should sum to 1, where the ideal values would be U , U - 0, and U - l. The basis for including these is that they are relatively straightforward, well known statistical measures, they address both relative and absolute differences, they avoid bias stemming from reliance on any single comparison measure, and they have been used in regional science literature to contrast nonsurvey' trade estimation techniques. The use of these statistical measures does not assume that the variables measured have particular frequency distributions nor that the sets of variables measured are strictly independent. Astizisl and finsnéins Data The influence of alternative outdoor recreation definitions will be examined in terms of their influence on the size of estimated economic impacts. Outdoor recreation studies employ different definitions and measure different sets of activities, making comparisons difficult and reported impacts somewhat arbitrary. The Outdoor recreation analysis for the 1987 Governors’ Conference on Forestry was no different in this regard, reflecting fairly' unique objectives. It estimated impacts throughout the Lake States from outdoor recreation activity in a 87 subregion (the more heavily forested area) of the three states. It relied on recreation participation data from state government SCORP reports which use slightly dissimilar recreation classifications. Team sports and. attendance at manmade attractions were not included as outdoor recreation activities. It did not include durable equipment expenditures. The primary analysis relating to recreation activity and spending data. here was to measure impacts associated with alternative (restrictive versus inclusive) definitions of spending corresponding to fishing, hunting, and wildlife-associated recreation in the forested areas of the Lake States. The resulting differences in. estimated impacts ‘will. be contrasted to the earlier Governors' Conference on Forestry estimates. The level of correspondence in activity and spending patterns across the three states and within their respective forested substate regions will be briefly noted by contrasting the percentages of recreation activity reported in the PRU data. Percentage differences for the activities and spending will be reported. Participation ‘patterns in ‘various activities such as hunting, fishing, and wildlife observation are expected to be fairly consistent across the forested areas of the three states. (Indeed, assumptions to that effect had to be made to fill in some gaps for the original Governors'Conference analysis.) Climatic and ‘natural resource similarities, and geographic proximity, should promote relatively similar patterns of outdoor recreation. Evaluation of. the degree of inconsistency found will be subjective; evaluation of its reasonableness in terms of explanatory variables is beyond the scope of this thesis. 88 However, differences will be interpreted in light of available sampling statistics. (1985 National Survey of Fishing, Hunting, and Wildlife- Associated Recreation (FHW) reports include statistical parameters; SCORP reports generally do not.) The range of activity implied by the statistics plays an important role in determining the range of impacts. Temporal and Spatial Considerations Economic impact estimation must address the aggregation of time, space, and activity. Due to gaps in data, estimates for different time periods must often be employed. For example, recreation participation levels and spending patterns often will not coincide with the year on which the economic impact model is based (for the current version (2.0) of IMPLAN it is 1982). Even the use of sector-specific deflators may leave much to be desired in reconciling such data, as deflators are based on U.S. average sectoral inflation. These may differ from state inflation rates for sectors with the same SIC codes, but comprised of a different mix of industries. The spatial issue is usually fairly straightforward in terms of delineating the region to be studied based on the study's purpose or as specified by research funding sources. However, input-output models for a particular state or region often do not exist (although these are fairly easy to construct for states or collections of counties if there is access to a nonsurvey input-output model such as IMPLAN). Also, many regions, such as national or state forests, do not follow county political boundaries perfectly; therefore, impacts must be extrapolated from models for regions that do not match the region of interest. Additionally, as previously mentioned, different deflators may be appropriate for use in different regional models. 89 The primary region modelled here exactly follows the regional delineation used for the Governors' Conference, consisting of the states of Michigan, Minnesota, and Wisconsin. In addition, IMPLAN models of other regions were developed to examine the applicability of some of the observations and results from the three-state models. In particular, Michigan IMPLAN models permitted a closer examination of alternative RPCs, subsequent multipliers, and related model validation issues. Outdoor recreation final demand values should be converted to 1982 values for use with IMPLAN version 2.0. In turn, IMPLAN calculates further direct, indirect and induced impacts generated by the final demands, based on average 1982 economic relationships. Means of testing and evaluating how well IMPLAN's 1982 relationships represent actual relationships for 1982 have been alluded to in the first section of this chapter and will be further described below under the "RPC-based Trade Estimates" section. Different I-O Sectorization Schemes The IMPLAN modeling system contains a user-friendly aggregation module which made it convenient to test for the effects of aggregation error. Stevens (1987, p. 19) argues: "...it is desirable to maintain the highest possible level of disaggregation in regional I-O model construction if any survey data are to be used. At the same time, there are compelling reasons for avoiding aggregation in nonsurvey models as well because of errors generated in the calculation of impacts from aggregated models in comparison with disaggregated models..." "...error in the calculation of impacts increases systematically with the level of aggregation of the I-0 matrix... ...Furthermore, sectoral aggregation predominately leads to over- rather than underestimation of impacts and multiplier effects. " 90 The number of economic sectors that. will be viewed. must be determined before multipliers are derived or values are inserted into IMPLAN. In its most detailed, disaggregated form, IMPLAN could provide information on over 500 economic sectors for the three-state region. However, such level of detail may be cumbersome and is seldom utilized in input-output analysis. The tendency is to illuminate those sectors affected by the subject of interest. Thus, sectors that have little or no relation to the economic activities being studied generally are viewed as less important and are more likely to be aggregated. The 1987 Governors' Conference input-output study (Pedersen and Chappelle, 1988) delineated 48 sectors. In keeping with the study's objectives, the 48 sectors included a greater representation of forest product sectors relative to most other I-O studies. The same I-O sectorization scheme will be maintained in this study, except that the forest product sectors will be aggregated, resulting in a 31 sector model. In addition, four other sectorization schemes will be be devised and contrasted. One will be a more highly aggregated scheme, involving 16 sectors. This more highly aggregated scheme will also serve as a means to measure the influence of improper (or an absence of) margining. Another more disaggregated scheme will have close to 150 sectors and provide additional detail on sectors affected by recreation spending. A completely disaggregated model, with over 500 sectors, will reflect the greatest level of' detail IMPLAN 'provides for the Lake States' economy. A fifth model will be intermediate between the 150 and 500 sector models and will contain 308 sectors. Percentage differences between estimated Lake State impacts will be used to contrast multipliers generated by models with alternative sectorization schemes. 91 Allocation of Survey Spending Data to MPLAN §ectors Several outdoor recreation researchers (including Clawson and Knetsch, 1966, Arthur D. Little, 1967, and, more recently, Goeldner and Duea, 1984) have noted a consistent pattern in reported recreation or tourism expenditures (e.g., roughly 25% of spending goes to lodging, 25% to food, 25% to travel expenses, and 25% to other miscellaneous items). Archer (1973, p. 60-61) implies identifying tourist expenditure patterns is hindered by "the wide range and different number of expenditure categories used" in various tourist expenditure studies. He goes on to state, "A more useful picture emerges if the spending pattern in each survey is reduced to four headings: food, lodgings, transport and other purchases." He then cites Clawson and Knetsch and Arthur D. Little studies to support his position. Consideration was given to constructing a composite "average" spending profile which allocates the bulk of spending to sectors corresponding *with lodging, travel expenses, and food. Impact consequences from alternative distributions of the remaining spending for miscellaneous items (approximately 25% of total spending) could be estimated by conducting a sensitivity analysis. The sensitivity analysis would consist of combining the miscellaneous category of spending with high and low multipliers generated by LG models, thus providing a measurement of the likely range of impacts for spending on miscellaneous items. This approach was dropped in favor of devoting more time to refining detailed allocations of spending to IMPLAN sectors. One of the problems with the approach is that it may fit relatively well with some tourism expenditure measurement objectives, but it does not address 92 equipment and a wide range of other types of spending associated with outdoor recreation. The problem of translating the spending into an I-O model's sector classification remains, even if survey spending only covers food, lodging, transportation” and. "other" categories. Publications addressing the so-called "bridging" of spending from survey categories to IMPLAN sectors have previously been handicapped by an absence of data and have lacked precision. The availability of Bureau of Economic Analysis (BEA) detailed personal consumption expenditure worksheets has provided the opportunity to improve upon the precision in existing bridge tables (for example, the "Personal Consumption Expenditure Categories" table, Appendix E of the IMPLAN' Analysis Guide). More refined bridge tables of FHW spending categories to IMPLAN sectors are a major project of this dissertation research. Details on the bridging are presented in chapter four and appendices associated with the chapter. Most I-O tables, including IMPLAN, are based on producer prices with trade and transportation margins allocated to separate accounts. Data from recreation expenditure surveys should be converted from purchaser to producer prices in order to deve10p correct I-O input data. The influence on impact estimates of not using sector-specific pce data to bridge recreation spending data to IMPLAN sectors will be illustrated with a subset of recreation data. This subset includes most categories of recreation equipment. Differences will be examined between sector allocation of spending using the more aggregated pce I-O category versus more narrowly-defined pce items for the recreation equipment. This will illustrate what the potential influence on impact estimates may be if 93 the more aggregated. pce categories are used rather than the more detailed, sector-specific pce data. 112m nc 9.1.3. W 112% W Five different sets of RPC values were substituted into Lake State IMPLAN models. Five sets of aggregation schemes were used for each set of RPCs. In turn, two types of multipliers, reflecting three types of economic variables multipliers (output, personal income, and employment) were generated and subsequently used to contrast the relative influences of the RPCs at different levels of sector aggregation. The same exact process was repeated for the state of Michigan, except six sets of RPCs were used. Unmodified IMPLAN RPCs (and their corresponding model multipliers) served as the benchmark for statistical comparisons. This is different from most other analyses of nonsurvey techniques which have used a survey' or "true" estimates (RPCs or' multipliers) as the benchmark. Alternative sets of RPC values to be used include minimum, maximum, unchanged and "best guess" values. A vector of REMI RPCs were used as well for Michigan models. Summary statistics were calculated, including the mean absolute difference, the standardized absolute difference, and chi square statistic, as well as absolute and percentage differences. For purposes of comparison to Stevens et al. (1986), Theil's information index and the root mean square error will also be calculated for the disaggregated RPCs. CHAPTER IV PREPARATION OF INPUT DATA AND IMPLAN MODELS Introducgion This chapter has two purposes. The first is to describe the stages followed in preparing data from the 1985 National Survey of Fishing, Hunting and Wildlife-Associated Recreation (U.S. Department of the Interior, Fish and Wildlife Service, 1988) to be used as inputs for IMPLAN models. The second purpose is to describe the construction of alternative IMPLAN models using the recreation data. The chapter is divided into two major subsections in line with these purposes. Converting Fishing‘, Hunting, and Wildlife-Agsociaged Data for Use with IBM The 1985 National Survey of Fishing, Hunting and Wildlife- Associated Recreation (FHW) contains both participation and spending data associated with fishing, hunting, and nonconsumptive wildlife recreation; the latter refers to the observation, photographing or feeding of wildlife. The 1985 FHW Survey, conducted by the U.S. Bureau of the Census, consisted of an initial telephone screening sample of almost 111,000 households. Subsequently, over 28,000 interviews were completed with fishermen and hunters and in excess of 26,600 interviews were completed with nonconsumptive recreationists (U.S. Department of the Interior, 1988, p. vii). Survey results are published in individual state reports and a national report. Data used in this analysis were from:the Michigan, Minnesota, Wisconsin, and national reports. 94 95 The FHW Survey has several advantages as a source of data for outdoor recreation economic analysis. It is the seventh in a series of surveys dating back to 1955 and provides detail for all fifty states. This allows for trend analysis and consistency checks of the data. The substate data the survey provides further enhances its applicability and was of critical importance in this analysis. The Survey employs a classification of spending which enables the segmentation and analysis of particular types; for example, trip-related expenditures. Generally speaking, spending, categories in the survey are more numerous and detailed than in most recreation surveys. This reduces the amount of error an analyst may make in translating survey spending categories to an impact model's economic sector classification scheme. The Survey differentiates between spending on used and new equipment, and includes durable equipment spending. These are spending categories some recreation surveys do not address. Finally, the presentation of statistical parameters for use with the reported data provides analysts with one means of measuring the reliability of the FHW estimates. The 1985 FHW Survey also has limitations. It does not report all spending associated with outdoor recreation. The Survey only applies to wildlife-related recreation and does not address other forms of outdoor recreation in the absence of wildlife-related recreation. Examples of activities not covered by the FHW independently of wildlife recreation include skiing, boating for pleasure, and hiking. Another problem is that many FHW spending categories are too vague or general to use directly with an impact model, as will be further explained below in the discussion on disaggregating FHW spending categories. 96 Sample sizes for a few categories of FHW data are so inadequate that some values are not reported due to reliability problems. This is especially true for substate region estimates. Other reported values based on small sample sizes may be misleading in light of sampling errors associated with them. For example, for some Michigan, Minnesota, and Wisconsin spending data, one standard error's difference from the spending estimate was equivalent to or greater than the estimate itself. (This finding points to the need to utilize the statistical parameters published with the FHW data to gauge its sampling precision.) Also, a few estimates are reported that are highly suspect in light of other estimates. Problems with dubious estimates might not have been detected had this analysis used only one state's FHW data. The 1985 FHW Survey remains a valuable source of data for measuring recreation impacts despite its limitations. Alternative sources of secondary data on recreation tend to have equal or greater limitations and. primary data collection. is often ‘prohibitively expensive. Recognition of the FHW Survey as a valuable source of recreation data is indicated in the IMPLAN Analysis Guide (1985, p. 4-15) and by Alward, Sullivan, and Hoekstra (1985). FHW data was converted into vectors of final demands and combined with IMPLAN multipliers to measure outdoor recreation economic impacts for the upper Lake States. This process will be described in terms of four major stages. Stage 1: Compile 1985 Recreation Activity Levels for More Heavily Forested Areas of the Lake States Step 1. Identify the more heavily forested subregions reported in the FHW state reports corresponding to the region analyzed for the 1987 Governors' Conference on Forestry 97 Individual state reports for Michigan, Minnesota, and Wisconsin of the 1985 National Survey of Fishing, Hunting and Wildlife-Associated Recreation (FHW) were used to construct recreation activity estimates. These state reports contain recreation activity data for subregions of the states in Table 23, "Fishermen and Hunters, Trips, and Days of Participation, by State Wildlife Management Region: 1985." The table provides estimates of' resident and ‘nonresident fishing and. hunting activity within each state subregion, in terms of numbers of participants, days of participation, and trips. State subregions were identified which closely corresponded to those in the "wildland" region used in the 1987 Governors' Conference on Forestry. The resulting FHW upper Lake State area differed from the recreation region analyzed for the Governors' Conference. These differences stem from the subregions in the FHW being composed of different combinations of counties than those contained in SCORP data which were the basis of the "wildland" regionalization in the original Governors' Conference recreation study. The differences between the two regions of analysis in terms of counties contained within them are as follows (with 1980 population figures in parentheses): Additionnl Cguntieg in Eng EEK Enginn MI: Isabella (54,110), Midland (73,578), and Bay (119,881) MN: Benton. (25,187), Kittson (6,672), Marshall (13,027), Norman (9,379), Pennington (15,258), Polk (34,844), Red Lake (5,471) Sherburne (29,908), and Stearns (108,161) anngigg in Eng vaegnogs' Conference fiegign, fin; n2; in‘ghe EB! Regign WI: Fond Du Lac (88,952), Green Lake (18,370), Marquette (11,672), and Sheboygan (100,935) 98 Thus, the FHW study area is larger, containing approximately ten percent more land area and population than the Governors' Conference area. FHW' recreation activity' would, therefore, be expected to be approximately ten percent greater in the area considered in this study. Step 2. Compile recreation activity data for the identified state subregions and utilize FHW statistical parameters to develop low and high activity estimates The object of this step was to develop estimates of the proportion of 1985 state recreation activity which took place in the FHW study region. This necessitated compiling FHW data for both the state subregions to be included in the FHW study area and state totals. Statistical parameters for each state from "Appendix B: Sample Design and Statistical Reliability" were used to construct standard errors of the participants, activity days and trips for both the FHW state subregion and each state. Low and high estimates were made on the basis of taking two standard errors from the initial participant, day, and trip data. Estimates were then summed for the three states. Low and high ratios of the FHW study area totals to the state totals were then constructed. The low ratio values were based on subtracting two standard errors from the FHW study area data and dividing it by the three-state sum data plus two standard errors. The high ratio was formed on the basis of adding two standard errors to the FHW study area data and dividing it by the three-state sum data minus two standard errors. The resulting range of estimates is shown in Table 3, below. This table represents the range of percentages used with state spending data to develop low and high estimates of spending in the FHW study region. Extremely low or negative values for certain cells in the table 99 reflect small sample size. Data in the table imply that, proportionately fewer trips and days are spent in the study region than average, relative to the number of participants. Table 3. FHW LAKE STATE STUDY AREA % OF 3 STATE RECREATION ACTIVITY DAYS OF PARTICIPANTS PARTICIPATIOI TRIPS (RIGINAL MIGINAL GIGINAL ACT IVITY LN ESTIMATE HIGH LCM ESTIMATE HIGH LCM ESTIMATE HIGH TOTALS: FISHING 61.92% 74.35% 89.06% 32.60% 45.60% 63.37% 27.82% 38.78% 53.73% HUNTING 59.88% 78.07% 101.34% 32.17% 54.65% 92.08% 28.43% 47.95% 80.36% RES: FISHING 59.67% 74.07% 91.62% 28.58% 42.43% 62.36% 24.75% 36.63% 53.72% IIJNTING 58.63% 77.41% 101.70% 30.44% 52.62% 90.06% 27.32% 46.88% 79.84% Big Game 56.41% 75.95% 101.64% 36.65% 63.82% 109.83% 32.27% 54.87% 92.69% Small 6812 40.13% 60.78% 90.16% 21.79% 45.37% 90.20% 20.80% 43.20% 85.75% Migratory 25.82% 53.31% 103.00% 12.29% 39.61% 110.88% 13.08% 39.76% 107.98% Other 3.65% 21.10% 64.73% -1.48% 13.26% 87.51% -1.22% 11.76% 77.58% NOIRES: FISHING 53.00% 75.09% 105.53% 36.25% 66.35% 118.59% 35.21% 61.46% 105.63% HUNTING 20.75% 56.58% 144.19% 9.98% 51.49% 246.64% 8.16% 35.66% 144.55% Stage 2: Estimate Study Area Recreation Spending Totals Next, statewide spending data was compiled and used in conjunction with the ratios in Table 3 to estimate the amount of recreation spending in the study subregion. Data for this step was drawn from FHW tables 19, 20, "Expenditures in the U.S. by State Residents for Fishing: 1985;" "Expenditures in the U.S. by State Residents for Hunting: 1985;" 21, Expenditures by State Residents for Special and Auxiliary Equipment Purchased Primarily for Fishing or Hunting: 1985;" 22, "In-State Trip- Related Expenditures for Fishing and Hunting: 1985; 30, "Expenditures in the U.S. by State Residents for Nonconsumptive Wildlife-Related Recreation: 1985;" 45, "Trip-Related Expenditures for Fishing, Hunting, and Primary Nonresidential Activities, By State Where. Spending Took Place;" and tables in Appendix B: Sample Design and Statistical Reliability. 100 The following steps were followed to develop spending estimates for the FHW study area: Step 1. Compile reported statewide estimates for fishing, hunting, and nonconsumptive recreation spending for each state Data for this step was drawn from FHW tables 19, 20, 21, 22, and 30. For each spending category where it applied, the reported "percent of equipment expenditures for new rather than used items" were used to derive estimates of new equipment purchases. Estimates of the percent of new equipment purchases are not reported for nonconsumptive spending for the U.S. or states. New equipment estimates of 60% for nonconsumptive special equipment and 90% for nonconsumptive auxiliary equipment were adopted, based on the percentages for similar fishing and hunting equipment in the region and U.S. averages. For special and auxiliary equipment, the percentages of new equipment purchases for hmnting are 59% and 95%, for fishing they are 64% and 83%, respectivelyu For the region, fishing and hunting special equipment are reported together as is the auxiliary equipment for the two types of recreation. New special equipment amounted to 50% of total regional special equipment purchases, while 85% of auxiliary equipment purchases were new. Step 2. Construct low and high estimates of spending Statistical parameters from Appendix B in the state FHW reports were used to estimate standard errors. In turn, these were used to develop low and high estimates for each spending category reported in the state tables. Low and high estimates were made on the basis of adding and subtracting one standard error from initial spending estimates. The decision to use one standard error here was a relatively arbitrary decision. It was partially predicated on the earlier decision to employ two standard errors for the range of recreation activity. Use 101 of two standard errors for spending estimates would have generated inappropriate negative low estimates for some spending categories. Step 3. Estimate the ratios of in-state trip-related spending to resident spending and use these to convert spending estimates to in-state resident and nonresident spending Detailed expenditure tables for fishing, hunting, and nonconsumptive recreation (tables 19, 20, and 30) include expenditures by residents in other states. Trip-related estimates were adjusted by data contained in tables 22, "In-State Trip-Related Expenditures for Fishing and Hunting: 1985" and table 45, "Trip-Related Expenditures for Fishing, Hunting, and Primary Nonresidential Activities, By State Where Spending Took Place." Due to a lack of data, the assumption was made that all reported equipment expenditures were made within the state of residence. Step 4. Sum the three states' spending estimates to derive low and high spending estimates for the FHW study region Statistically, this results in a greater range between low and high estimates than the true range would be for the three state region. This is because standard errors for the three-state region would be based on the sum of the three states' sample sizes, reducing the probable sampling error. However, FHW reports do not present sampling sizes, so standard. errors for the three-state region can not be calculated. Adopting the sum of the three states' low and high estimates as the three-state region's estimates may be viewed as a conservative approach to the estimation. Spending categories were first matched with a ratio of FHW region to total Lake State activity measurement (participation, days, or trips). Thus, for example, equipment and food spending categories were 102 matched with the ratios of participants, whereas transportation was matched with the ratios for days. This was an attempt to reflect that the categories of spending would be more closely associated with particular types of activity measurement. Transportation could have been linked with the FHW region- to-Lake State trip ratios, however the trip ratios are only slightly lower than the day ratios. The measurement ratios were then multiplied by the three states' summed low and high spending amounts to derive low and high estimates of spending by category for the FHW region. Nonconsumptive recreation activity is not delineated by substate region in the FHW state surveys. Nonconsumptive spending in the study area was assigned low and high ratios of total three state spending slightly lower than the fishing and hunting ratios. Nonconsumptive recreation activity refers to feeding, photographing, or observing fish, birds or other wildlife. The slightly lower ratios reflect the author's general impression that there is likely to be more nonconsumptive activity outside the FHW region relative to fishing and hunting. This impression is partly based on knowledge of bird, fish and wildlife sanctuaries with the three states, but outside of the FHW region. Auxiliary and special equipment spending also required an extra step. Lump sum amounts for these two categories appear both in the fishing expenditures table (table 19) and in the hunting expenditures table (table 20). Table 21, "Expenditures by State Residents for Special and Auxiliary Equipment Purchased Primarily for Fishing or Hunting: 1985" further delineates subcategories of spending which fall within auxiliary and special equipment classifications, but the table combines both fishing and hunting expenditures for these types of equipment. 103 Fishing and hunting spending estimates for the FHW study area were calculated with respect to their separate ratios of activity in the study area to total three-state activity. Low and high amounts for special and auxiliary equipment spending from the separate state estimates of special and auxiliary expenditures for fishing (table 19) and hunting (table 20) were used with FHW study area fishing and hunting participant ratios to weight the FHW study area estimates for these spending categories. These steps generated a complete set of low and high estimates of 1985 spending associated with resident and nonresident fishing, hunting, and nonconsumptive recreation in the FHW study area. Table 4, "FHW Lake State Study Area Spending Totals," presents this data. Table 4. FHW Lake State Study Area Spending Totals 5 SPENDING CATEGORY TOTAL SPENT TRIP-RELATED food&lodging food lodging transportation other rental& use fees boat fuel boat maintenance bait ice EQUIPMENT & OTHER guns & rifles ammunition other hunting equip field glasses binoculars film & dev. other photo equip carrying cases and clothing bird seed other bird items other noncons equip rods reels lines, hooks,etc. lures & flies tackle boxes creels, nets bait containers scales & knives other fishing equip Licenses Stamps,Tags&Permits F&H AUXIL EQUIP Total Camping Equip Foul Weather Gear Spec Clothing Rubber Boots/Waders Equip Maintenance Fish or Hunt Boots 6 Other Aux Equip 104 1985 SPENDING ESTIMATES (1985 $) LOW RESIDENTS & NONRESIDENTS HIGH 1656137192 4606529506 780847326 1761962757 457326683 379102799 78223883 156244689 5578302 28810119 56092174 18248782 49908306 8638272 875289865 50133145 16850942 41133645 1295999 7993398 11662630 10359849 3067522 26427582 4940832 4761586 28483013 15147069 16963921 25659365 3130878 390674 704546 1053558 31155186 43378888 6502729 54980767 10825992 5071973 22056471 6345510 938077 5117635 4625110 963742975 770519833 193223142 456017744 26622643 56182109 105726637 48701439 89470997 15498212 2844566750 117295350 32745642 86426819 6165415 47309958 63944055 . 69644146 19035050 142923478 29509077 24423050 55381930 30457776 30236818 47701426 6294312 735984 1267743 2168250 64759735 77834072 14293593 149843373 29830741 12560638 60699838 16195281 4386746 13939228 12230900 NONRESIDENTS ONLY LOW 166681860 166681860 97784729 79653205 18131525 38294574 1413705 8723119 9683197 1220628 7775963 1785944 HIGH 453756719 453756719 242900499 189635349 53265150 142605518 10204019 17738881 19032785 3396978 14536675 3341365 105 Table 4. (cont'd.) SPENDING CATEGORY RESIDENTS & NONRESIDENTS FGH SPEC EQUIP Total 265050068 959249150 Boats & Canoes 129217752 389197274 Boat Accessories 30245150 78929673 Boat Trailer&Hitch 9968098 41603430 Travel or Tent Trailer, 67686868 292177683 Pickup, Van, Motor Home, or Cabin Off-Road Vehicles 25028944 150961533 Ice Chests 2903255 6379557 7 Magazines,Dues,Leases 152477996 378370350 Nonconsum Magazines 4128112 21959082 Nonconsum Membership Fees 3941310 21625035 8 Nonconsum Spec Equip 41676626 330230391 9 Nonconsum Aux Equip 1838029 12735692 Table 5 Notes 1) 2) 3) 4) 5) 6) 7) 8) 9) includes nonconsumptive and. hunting expenditures for "equipment rental (boats,camping equipment, etc.) and fees for guides, pack trips, public land use, and private land use." "includes bows, arrows, archery equipment, telescopic sights, decoys and game calls, equipment or game cases or carriers, handloading equipment, hunting dogs, hunting knives, and other unspecified hunting equipment." includes "nest boxes, bird houses, bird feeders, and bird baths" undefined in FHW survey "includes electronic fishing devices (depth finders, fish finders, etc.) rod holders and belts, spear fishing equipment, ice fishing equipment and other unspecified fishing equipment." "Includes binoculars, field glasses, snow shoes and skis, processing and taxidermy costs and other unspecified equipment." ”includes magazine subscriptions, membership dues and contributions, and land leasing and ownership" fishing and hunting expenditures. "Includes travel or tent trailers, off—the-road vehicles, pickups, campers or vans, motor homes and other unspecified equipment." "Includes tents, tarps, frame packs and other backpacking equipment, other camping equipment, snowshoes and skis." 106 §£ege 1: Disaggregate Spending Totals for Bridging with IMPLAN Data on recreation spending by sector form the link between recreation activity surveys and I-O models. It is beneficial for the sake of precision to have spending categories as narrowly defined as the input-output model with which the spending estimates are to be used. This avoids aggregation error. However, survey spending profiles are often developed independently of the I-0 model and rather aggregated spending categories are frequently used in the interests of survey brevity and a higher survey response rate. This sometimes results in mismatches between recreation survey data and the Standard Industrial Classification (SIC)-based sectorization schemes of most I-O models. Thus, for example, surveys may not delineate between food purchased at restaurants and bars versus food purchased at grocery stores. The result of using survey classification schemes which ck) not correspond with impact model classifications is that compromises must be made to match the survey data to I-O sectors. A consequence of such compromises may be impact estimates having poor reliability. Substantial effort was devoted to bridging FHW spending data to the appropriate IMPLAN sectors. Publications addressing the bridging of recreation spending data to IMPLAN model economic sectors have tended to treat the subject superficially or in a relatively crude fashion. Much like improved RPCs, refinements in bridging of spending data will lead to improvements in ‘both total impact and specific sectoral impact estimates. They may also help improve survey design to meet impact measurement objectives. While a number of spending categories were adequately disaggregated, most of the spending items as they appeared in the state 107 reports could not be directly bridged to IMPLAN sectors. Five approaches were taken to further disaggregate the spending. The first consisted of utilizing available local studies. A second technique was to use ratios of spending from the U.S. FHW report which contains more detail than the state reports for certain spending categories. A third approach involved using extrapolations from national averages cited in other sources. A. fourth approach involved reviewing the aggregated spending categories with Michigan State University Parks and Recreation Resources Department personnel for expert opinion as to the possible division of items within these categories. Finally, ad hoc judgment was used for some items. The further disaggregation used the initial spending categories as control totals and subsequently allocated these totals among subcategories of spending items. FHW state and U.S. reports provide indications of what is included in various spending categories. Specifically, these reports contain a number of footnotes elaborating on the items contained in many spending categories. For example, a footnote for the hunting category, "Trip—related Other," indicates the spending in this category includes hunting expenditures for "...equipment rental (boats, camping, equipment, etc.) and fees for guides, pack trips, public land use and private land use." In a number of cases, spending for these items could be identified directly in the corresponding national FHW report and their national ratios adopted for the study here. FHW "food" spending was the major category for which there was substantial regional data that provided guidance as to how it could be further disaggregated. Four regional recreation studies were identified 108 that divided food between food for off premise consumption (groceries) and food purchased at restaurants and bars (eating and drinking places). The four studies included studies of recreation in three counties of Michigan's upper peninsula (Spotts and Mahoney, 1986), Michigan state parks (Fridgen, et a1, 1986), Michigan state forest campgrounds (Nelson, 1988), and the more heavily forested areas of Minnesota (Kelly and Becker, 1985). Transportation was also divided between private automobiles and other forms of transportation (airlines, railroads, and buses) based on the three states' travel data (Gouldner and Duea, 1984). National averages were used to allocate private transportation spending between automobile repairs and petroleum and 011. These averages were derived from 1977 U.S. Bureau of Economic Analysis (BEA) detailed personal consumption expenditure (pce) worksheets (1984). The difference between 1982 U.S. personal income and personal consumption expenditures was used to allocate lease payments to regional households' personal consumption expenditure. The "FHW Bridge Table" (Appendix F) includes a column, "Basis of Allocation" which spells out the basis for sector allocations of all spending that required further disaggregation from the original FHW spending categorization. SEege g: Allocate Spending to IMPLAN Sectors Step 1. Margining Most I-O tables, including IMPLAN’s, are based on producer prices. Separate I-O economic accounts distinguish trade and transportation margins from producer goods. Margining refers to the process of converting purchaser prices to producer prices through. separating the trade and transportation margins and allocating the remainder to 109 appropriate producer sectors. Data from recreation expenditure surveys must be margined in order to develop correct IMPLAN input data. Close attention to margining may be very important if an objective is to measure sectoral impacts rather than merely total impacts. A relatively narrow range of" multipliers may lead to only a ‘modest difference in total impacts regardless of sector spending allocation. However, there will be dramatic differences in the size of spending for particular sectors -- especially trade sectors -- depending on whether the spending data is properly margined or not. Margining is also particularly important for any sectors which do not exist in a region being modeled. Spending on items produced by sectors without regional output should still have their margins, representing the proportional amounts of distribution costs, allocated to corresponding regional trade and transportation sectors. In contrast, the producer's share of an expenditure for an item not produced in a region would be considered imported and it would not be included in the measurement of impacts. It is not perfectly reflective of the real world to allocate all distribution margins to the regional economy in the case of sectors without regional production. For example, the major portion of transportation costs may occur outside of the region. However, the bulk of distribution costs are normally associated with the retail trade sector, and the majority of these costs occur within the region's retail trade industry. There are numerous other spending allocation considerations. Some service sectors, such. as hotels and lodging 'places and. eating and drinking places, do not have trade and transportation margins associated 110 with them. For these, the recreation spending should be completely allocated to the corresponding sector. .Used equipment purchases pose several complications. They do not represent current production and, therefore, no portion of used items should be allocated to the sector corresponding to new production of the equipment. Furthermore, purchases may be made through retail outlets or from other households. If from retail outlets, IMPLAN has specific margins associated with its used and secondhand goods sector which may be appropriate. If the purchase is from other households, the treatment of it should depend on whether the puchase is made by regional residents or nonresidents. Appendix E of the IMPLAN Version 1.1 Analysis Guide provides national margins for most of 100+ personal consumption expenditure (PCE) I-O categories derived from detailed U.S. Bureau of Economic Analysis (BEA) worksheets. (Appendix B of "Use of IMPLAN with Public. Area Recreation Visitor Survey (PARVS) Pretest Data: Findings and Recommendations" (Propst, 1988) contains similar data.) Computerized files of the detailed BEA worksheets are also available indicating the detailed margins for the 1700+ PCE items within the 100 1-0 categories. Use of the margins from the detailed BEA worksheets are too difficult and time consuming to wade through for most IMPLAN users. The detailed BEA worksheets were utilized to bridge fishing, hunting, and wildlife-associated recreation spending in the FHW study area. The bridging included the process of converting from purchaser to producer prices. The resulting detailed bridging tables and associated computer spreadsheets may prove especially helpful for IMPLAN users measuring recreation impacts based on FHW data. However, the tables may 111 also be useful for other recreation I-O analyses. They cover a broad range of spending categories, incorporate standard SIC and PCE code numbers and names, and address issues such as used equipment spending. The two major bridge tables produced are Appendix F: "FHW Bridge Table," and Appendix G: "FHW Used Equipment Sector Allocation Table." There are qualifications to the usefulness of the tables. They have not 'been. extensively reviewed, sector allocations indicated in the tables apply to Lake State FHW recreation spending and may not be appropriate for other regions, and the allocations of FHW spending were sometimes made on the basis of ad hoc judgement with little information. Two examples of the latter include the disaggregation of the FHW spending categories of "camping equipment" and "hunting dogs and associated costs." Step 2. Develop four different types of spending: trip-related, total, total minus special equipment, and nonresidential Each of these types of spending contained spending subcategories with low and high estimates. A further distinction was developed on the basis of only allowing for the measurement of nonresidential payments to households. Thus, low and high estimates were further subdivided by a distinction as to whether payments by residents in the region to other residents of the region were counted. Step 3. Deflate spending I-O input data and models for different time periods must often be employed because of analysis objectives or gaps in data sets. Recreation participation levels and spending patterns often will not coincide with the year on which the economic impact model is based (for the current version (2.0) of IMPLAN it is 1982). Sometimes analysts have defaulted to using singular price deflators such as the Consumer Price Index (CPI) 112 to using singular price deflators such as the Consumer Price Index (CPI) to adjust their spending data to I-O models. Such an approach involves aggregation error as different economic sectors often experience significantly different rates of inflation. Prices consumers pay for goods will also vary from one region of the country to another. A standard set of U.S. Bureau of Labor Statistics (BLS) deflators distributed with IMPLAN materials and often adopted by IMPLAN users covers price trends at a 105-sector level of aggregation (see, for example, Appendix D of the IMPLAN Version 1.1 Analysis Guide). Use of these deflators may still involve significant aggregation error as the IMPLAN model database contains 528 sectors. Of course, the potential error grows as the difference between input data and model year increases. Sector-specific deflators for manufacturing industries are published in the BLS' monthly Producer Price Index (PPI) journal. The PPI provides up to seven-digit SIC detail which is much more than is required to match the maximum level of disaggregation in the IMPLAN database. These may be available in electronic form from the BLS and may be very easy to substitute for a majority of the more aggregated 105-sector deflators. Admittedly, even the use of more precise sector-specific deflators may leave much to be desired because they will still be based on average inflation rates for the U.S. These may differ from regional inflation rates for sectors with the same SIC codes because of differences in local conditions from U.S. averages. However, regional deflators are also published by the Federal Government and there may be possibilities to use these in conjunction with more detailed sectoral deflators. In 113 distributed with the IMPLAN model warrants further attention. The 105- sector BLS deflators distributed with IMPLAN materials were used in this study as it was assumed the deflation of spending from 1985 to IMPLAN's 1982 base year would not introduce substantial aggregation error. Step 4. Eliminate spending for sectors not in region Comparisons of the deflated spending vectors were made to the vectors of sectors IMPLAN reported as existing in the three-state region. Spending for sectors which did not exist in the region were subsequently dropped from the analysis. This is equivalent to assuming the producer's share of expenditures for these sectors was wholly imported into the region, and thus had no impact. Inn contrast, the distributional margins associated with those sectors found in the region were included in the impact measurement. Step 5. Aggregate spending to match sectorization schemes The spending was then aggregated according to the five different sectorization schemes described in the next chapter through the use of computer spreadsheets specifically developed for this purpose. Step 6. Multiply spending by multipliers to derive impact estimates The final step was to take the alternative low and high estimates for the four types of recreation spending and combine them with the three sets of multipliers (output, personal income, and employment) for each aggregation scheme. This provided FHW recreation economic impact estimates for the forested study area. Chapter five describes the further analysis of these impact estimates in terms of the influences of variables on the size of the estimates. 114 . Censernction efi g3 m M Models Once the FHW data had been converted to 1982 final demands as described in the steps above, it was used in conjunction with a variety of multipliers derived from alternative Lake State IMPLAN models. The discussion below focuses on “how the alternative models were developed based on different sets of regional purchase coefficients (RPCs) and alternative sectorization schemes. The following chapter will contrast the alternative I-O models and present outdoor recreation impact results. Develenment ei Alternative RPC-Based Trade Estimates Lake State region IMPLAN models were generated using five different sets of RPCs. These are presented in their entirety in Appendix C: "Alternative Lake State RPCs and RPCs used as Guides for the Estimation of the Lake State RPCs." They are identified as "SDP," "FLRLK," ALTFLK," "BSTLK," and UNCHLK" in the Appendix. Additionally, the Appendix table includes five other sets of RPCs which influenced the construction of the Lake State region RPCs. These other five include two sets of Michigan RPCs (one from IMPLAN and one from REMI), Minnesota and Wisconsin IMPLAN RPCs, and a set of RPCs for the Lake States generated from the corrected MRIO data set. Each set includes 528 RPCs, one for each sector in the fully disaggregated IMPLAN model (including about twenty sectors not present in the Lake State region). Two of the sets of RPCs are standard IMPLAN outputs and required no significant user changes. One of these was a set of unchanged RPCs (UNCHLK) derived directly from IMPLAN and used without any modification. Another set of RPCs was based on the supply-demand pooling (SDP) trade estimation technique. The SDP values serve as ceiling RPC estimates 115 (given the levels of IMPLAN-estimated regional demand and output). The other three sets of RPCs required sector-by-sector examination and modification of RPC values. A set of RPCs representing minimal or floor values (FLRLK) was generated based on adopting the minimum RPC values from among unchanged Michigan, Minnesota, Wisconsin, and Lake State IMPLAN models, a Michigan REMI model and calculated Lake State RPCs based on the corrected MRIO data for the three states. REMI RPC estimates were not available for all 528 sectors and the RPCs were for a different base year (1983 in contrast to 1982). A limitation on the MRIO data is that it is at a 124 sector level of aggregation. A fourth set of RPCs was developed as an alternative set of floor RPC values (ALTFLK). The intent behind this set of RPCs was to address the questionable nature of many of the zero and near-zero RPC values found in the IMPLAN models. Alternative minimal RPCs were substituted for FLRLK RPCs for approximately fifty sectors; otherwise, the ALTFLRLK vector of RPCs is identical to the FLRLK vector. The substitution was done on the basis of a sector's FLRLK RPC appearing unreasonably low (generally less than .01) in contrast to the majority of other RPCs, especially the REMI, MRIO, and SDP estimates. While the value was selected on an ad hoc basis, the new ALTFLK RPC value adopted tended to be half the SDP value or the lowest RPC from among the remaining nonzero RPCs. The fifth set of RPCs (BSTLK) represented the author's best judgement as to the probable magnitude of the RPC. ‘ This set was constructed on the basis of subjectively weighting the other sets of 116 RPCs. In a few instances, consideration was also given to 1982 U.S. Census data for the three states in the region. 2mm; I_-Q W911 Schemes The five different sets of RPCs were used with five different sector aggregation schemes. One of the schemes involved no aggregation and resulted in Lake State models with 502 sectors. A second aggregation scheme involved an intermediate amount of aggregation resulting in models with 308 sectors. Part of the motivation behind this sectorization scheme was to reduce the total number of sectors to just below IMPLAN's upper limit (310 sectors) whereby the model can invert theLeontief matrix without partitioning it. IMPLAN's inversion routine slows down for models with more than 310 sectors, although by standards of a few years ago, the speed of the inversion is still quite respectable. Two rules of thumb were followed in aggregating sectors. The first was to aggregate sectors as they appeared in IMPLAN version 1.1 to enable comparisons of recreation impacts with earlier analysis conducted with version 1.1. Aggregation of sectors is generally more viable than disaggregation. Version 1.1 of IMPLAN had 464 sectors whereas version 2.0 has 528. Version 2.0's 528 sectors can be aggregated in a straightforward fashion to identically match version 1.1 sectors. The second rule of thumb was to aggregate only relatively small and similar SIC sectors which did not have any, or only minimal, recreation spending associated with them (mining sectors, for example). A third aggregation scheme represents the extreme case of aggregating all less important: sectors into a "miscellaneous" (or "other") sector. This third sectorization scheme has 159 sectors, 158 of 117 which are relatively disaggregated. Sectors were first segregated on the basis of whether or not there was recreation spending associated with them. Those that had recreation spending associated with them remained relatively disaggregated. All sectors not associated with recreation spending were aggregated together. The only aggregation for the relatively disaggregated sectors was for the purpose of matching the aggregation scheme of the IMPLAN version 2.0 sectors to the aggregation scheme of version 1.1. The other (159th) sector is "all other". It is an aggregation of 326 sectors which had no recreation spending allocated to them in the 1987 Governors Conference study. The intention behind this aggregation scheme was to examine the issue of what happens when the analyst combines all other sectors together which are not of immediate interest. The other two sectorization schemes are highly aggregated. One has 31 sectors. This scheme closely approximates a 2-digit SIC delineation which is probably the most common sectorization scheme seen in regional economics literature. It is quite similar to the original Governors' Conference on Forestry model (Pedersen and Chappelle, 1988), except that it lacks sectorization detail for forest product industries. The other aggregation scheme has only sixteen sectors. The major difference between it and the 31 sector model is that most manufacturing sectors and many service sectors have been further aggregated. The exceptions are ‘recreation-related..sectors (food, clothing, petroleum, and transportation services). Hotels and Lodging Places (IMPLAN sector 471) and Eating and Drinking Places (IMPLAN sector 491) were left unaggregated for both of these sectorization schemes. 118 The sector aggregation schemes and sector names for the four aggregated model types are presented in Appendix D: "IMPLAN Aggregation and Sectorization Schemes." CHAPTER V RESULTS Introduction This chapter describes results of the IMPLAN RPC analysis and subsequent analysis of alternative estimates of Lake State outdoor recreation economic impacts. The RPC analysis will be discussed first, followed by results pertaining to Lake State outdoor recreation economic impacts. WMWQEWMWW Concerns regarding IMPLAN's RPCs were first encountered during the construction of a Lake States IMPLAN model (Pedersen and Chappelle, 1988). Between 50 to 60 of the original disaggregated sectors had obviously inappropriate RPC values equal to zero. Initial questioning and subsequent adjustment of RPC values was on the basis of IMPLAN reporting a zero or near-zero RPC, despite output for the sector being sizeable relative to demand for the sector, and on the basis of comparisons with RPCs from REMI models maintained by the three Lake States. Thus, for example, many service sectors with zero RPCs were questioned and subsequently changed, while most mining sectors and some manufacturing sectors with zero RPCs were left unchanged. Also, zero or near-zero RPCs tended to come in clusters and were relatively easy to identify. It was later observed that the basis for the cluster pattern was that IMPLAN's RPCs are initially estimated at a more aggregated level, based on MRIO 124 sector data. IMPLAN models were subsequently generated for several Michigan counties and the states of Michigan, Minnesota, and Wisconsin. All 119 120 models contained dubious zero-valued RPCs as well as many questionable near-zero RPCs, reducing the size of estimated multipliers. For all the models, between ten and thirty percent of sectors with regional output had questionable RPCs at the fully disaggregated level. There appeared to be a high degree of consistency as to which commodity sectors were affected across the various models. Michigan and all Michigan counties had thirteen service sectors (IMPLAN sectors 478 through 490) with RPCs equal to zero, whereas Minnesota and Wisconsin RPCs were all above .5 for the same sectors. Two other categories of sectors had dubious negligible RPCs in all the models, in addition to the commodity sectors listed in Table l (p. 17). These were SIC 364, Electric Lighting and Wiring Equipment (including IMPLAN sectors 386, 387 and 388) and SIC 32, Stone and Clay Products (including IMPLAN sectors 257 through 279). Unchanged IMPLAN RPCs for Michigan, Minnesota, and Wisconsin are presented in Appendix C: "Alternative Lake State RPCs and RPCs Used as Guides for the Estimation of Lake State RPCs." The number of sectors with questionable zero-valued RPCs was higher for the State of Michigan than it was for Michigan counties on an absolute basis. This is because the zero RPC estimates for certain manufacturing sectors at the county level coincided with those sectors not being present in the county, making the zero value the correct RPC value. At the state level, in contrast, the same manufacturing sectors often had substantial levels of output that made a negligible RPC estimate questionable. However, the Michigan county models examined did have a higher percentage (closer to thirty percent) of sectors with questionable zero RPC values than the state (which had between ten and fifteen percent). 121 This observation provides an indication that IMPLAN's RPC estimation problem may be more serious for substate or small models with relatively few sectors. IMPLAN models may be developed for single or multiple counties, in addition to individual states or multiple states. Smaller regions being modelled with IMPLAN will tend to have more distortion in their trade estimates in terms of the percentage of total model sectors affected. As Jensen (1988) has suggested, service (or "tertiary") sectors appear to be fairly universally found across most economies, large or small. Therefore, service sectors often comprise a greater proportion of the total number of sectors in small economies than in larger economies. Furthermore, RPCs for service industries are generally higher than those for manufactured goods. This reflects demand for service industries often being met by local suppliers in contrast to demand for manufacturing and other goods often being met by production outside a region. Therefore, the magnitude of error for service sectors with zero RPCs is generally greater than manufacturing sectors with zero RPCs. These problems compound. the existing nonsurvey I-O problem of extrapolating from national technical coefficients (the "A” matrix) to the small area economy. Generally, the smaller the region, the less likely' the industry mix and ‘production functions will reflect the national average. In sum, if service sector RPCs are in error, these may create worse problems for small economies being modeled than large economies. A comparison between disaggregated 1983 REMI and IMPLAN (1982) RPCs for the State of Michigan revealed little overall correlation in the pattern of RPCs between these two models. Disaggregated REMI RPCs were 122 not available for agricultural, construction, or government sectors. The coefficient of multiple determination, R2, goodness-of—fit measure was used to indicate the amount of correlation between major sector groupings of disaggregated RPCs. The regression results are shown in Table 5. Table 5. Comparison of IMPLAN and REMI RPC's for Major Sector Groupings Number of Sectere Qgggeieted SeeEeze E2 * All 440 .1578 Forestry & Fish Prdcts,& Mining 42 .0218 Mining 38 .0183 Manufacturing 370 .1527 Nondurable Manufacturing 158 .2295 Food & Kindred Products 49 .1900 Durable Manufacturing 206 .1063 Transportation, Communication, and Utilities 14 .0259 Finance, Insur, & Real Estate 7 .0180 Other Services 45 .0598 * "All" does not include agriculture, construction and government sectors, as REMI RPCs were not available for these sectors The above comparison does not indicate IMPLAN's RPCs are of poor quality, but rather simply that there is little correspondence in the overall consistency between REMI and IMPLAN values. While some inconsistency should be expected in light of the different techniques used to derive their RPCs, the low level of correlation was less than what was expected. Alward et al. (1989) report observed MRIO RPCs for service sectors that have been adopted by IMPLAN. Commodity sector coefficient values for independent variables in the RPC regression equation are published as well, but not corresponding commodity RPC values. The service sector RPCs are reported for all fifty states and Washington, D.C. Close 123 inspection indicates most states have two or more MRIO service sectors (corresponding to aggregations of IMPLAN sectors) with RPC estimates of zero. Many of these zero RPCs appear dubious in light of: l) the same state having RPCs close to or equal to 1.0 for most other service sectors, 2) most other states having RPCs close to or equal to 1.0 for the same sector, and 3) service sectors are usually oriented towards supplying local markets. A possible alternative to accepting zero RPC values for a particular service sector would be to substitute the average MRIO RPC observed for that sector in other states. This approach would assume the nonzero service sector RPC values are correct and it does not address questionable nonzero RPC values observed for commodity sectors. .Also, an analysis of all service sector RPCs has revealed that there are extensive problems with the nonzero values as well. For example, New York has the lowest overall service sector RPCs -- less than half the average value of all other states -- despite the state being an important regional and national center for finance and other services. It is not known whether the reported zero RPC values are the result of legitimate estimates from the original MRIO data, udscalculations, or data entry errors. In any case, many nonzero service sector RPCs are suspect as well. The questionable nature of the current IMPLAN service sector RPC estimates will be illustrated in tables 6 through 9. The pattern of IMPLAN's average state RPC values that emerges is not what one would expect to actually occur. For example, the economies of New York, Illinois, South Carolina, New Jersey and Indiana have average service sector RPCs less than 0.60, while Minnesota, Wisconsin, California, Nevada, and North Dakota have averages above 0.95. There is 124 no apparent basis for this pattern in terms of the size of states or their spatial or economic positions relative to surrounding states. Substituting higher values for the zero RPC service sectors only slightly improves the resulting RPC estimates; a peculiar pattern of RPC values still remains. For example, even after dropping sectors with RPCs equal to zero, New York still has the lowest average service sector RPC. Its average service sector RPC value would remain under 0.37, less than half the average value of all other states. New RPC estimates were calculated for the service sectors of all fifty U.S. states and Washington D.C. based on corrected MRIO data (see chapter 3, p. 78-81). These are presented as "Revised RPCs" in Table 6. The revised average service sector RPCs for the states have a much smaller standard deviation (0.0361) than the IMPLAN average service sector RPCs (0.1508). Mathematically, the smaller standard deviation for the revised RPC averages stems from there being few zero, near-zero, and unit-value RPCs among the revised service sector RPCs. The narrower range of the revised state service sector RPCs corresponds with observations regarding the similarity of service sector markets across states -- service sector demands are usually met locally. The revised RPCs also more closely reflect what one would expect from regional economic theory in terms of the magnitude of the average state RPCs relative to each other. This is illustrated by Table 7 which lists the ten states with the lowest and highest revised average service sector RPCs. IMPLAN and Revised State Average Service Sector RPCs and Rankings By State Table 6. 125 5.21:2 WWW WEBER IL SC NJ IN AK WV NC PA OH HI MS AL FL CT MI 0 3’ §858§§§§§§§53 Averagez. Standard Deviation: 0.3487 0.4471 0.5468 0.5497 0.5591 0.5825 0.5905 0.5950 0.5978 0.6070 0.6314 0.6418 0.6421 0.6449 0.6487 0.6666 0.6790 0.6981 0.6981 0.7099 0.7109 0.7162 0.7283 0.7539 0.7575 0.7681 0. 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 l 1 0 0. 7751 .7833 .7843 .7932 .8051 .8271 .8291 .8485 .8671 .8671 .8807 .8860 .9007 .9040 .9124 .9194 .9215 .9280 .9389 .9438 .9511 .9700 .9722 .0000 .0000 .7594 1508 H OOQNO‘LflthNH mU‘k46>?9t?9k§§¢wwwwwwwWUUNNNNMNNNNNF-‘HHHHHHHH HO‘DQNO‘U‘tfiWNO-‘OQONO‘U'#UNHOOQNGU‘#UNHOOQVO‘U‘9UONH OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOGOOOCOCO .8299 .8255 .7777 .7938 .7635 .7982 .7240 .7992 .8206 .7639 .8044 .7665 .7649 .8242 .7912 .7787 .8310 .7518 .8162 .8077 .7377 .8249 .7581 .7526 .8295 .7937 .7277 .8553 .7809 .8213 .7006 .7637 .7301 .8158 .7804 .7463 .8061 .8034 .7556 .8259 .8042 .8178 .8398 .7311 .7828 .7852 .8380 .7735 .8515 .7944 .7492 .7884 .0361 43 19 27 13 29 2 30 39 15 33 17 16 41 25 20 47 9 37 35 6 42 12 10 45 26 3 51 22 40 l 14 4 36 21 7 34 31 11 44 32 38 49 5 23 24 48 .18 50 28 8 126 ,Table 7. Ten Lowest and Highest Revised MIRO Average Service Sector RPCs owest Averege RPCs Highesg Average RPCs Rank State RPC Rank State RPC 1 Rhode Island .7006 42 Arizona .8249 2 West Virginia .7240 43 Illinois .8255 3 Wyoming .7277 44 Virginia .8259 4 South Dakota .7301 45 Texas .8295 5 Delaware .7311 46 New York .8299 6 New Hampshire .7377 47 Georgia .8310 7 Washington D.C. .7463 48 Minnesota .8380 8 North Dakota .7492 49 Missouri .8398 9 Vermont .7518 50 California .8515 10 Kentucky .7526 51 Colorado .8553 Table 7 illustrates a discernible, logical pattern to the revised RPCs based on the corrected MRIO data. Small revised average service sector RPCs tend to be characteristic of small and/or relatively isolated states, while higher average RPC values tend to be associated with larger and more highly developed state economies that are regionally dominant. Table 8 illustrates how revised estimates of RPCs based on the corrected MRIO data compare to current IMPLAN estimates for particular service sectors for the states of Michigan, New Jersey, New Ybrk, and North Dakota. There are substantial differences between the two sets of RPCs. Table 8 illustrates the severe distortion an analyst might encounter using version 2.0 of IMPLAN to measure tourism or recreation impacts. Note for example that IMPLAN's current New York Hotel and Lodging sector RPC is .182 (see Table 8). This is less than one-fourth of the average for the fifty states and implies that 80% of demand for temporary lodging within the State of New York is met out-of-state. However, despite consistently appearing to be inappropriately on the low side, IMPLAN's New York service sector RPC estimates at least have a pattern compared to New Jersey's service sector RPC estimates. Consider 127 the transportation and tourism implications of IMPLAN's New Jersey RPCs. Except for private automobiles, IMPLAN's RPCs imply railroads are the only form of transportation New Jersey residents and businesses use. The zero RPCs for MRIO sectors 98 (eating 8 drinking) and 106 (hotels and lodging) imply state residents never patronize New Jersey restaurants and bars or accomodations. Table 8. Comparison of Recreation-Related Service Sector RPCs for Michigan, New Jersey, New York, and North Dakota MRIO MI RPCs NJ RPCs NY RPCs 1D RPCs it SECTOR NAME IMPLAN REVISED IMPLAN REVISED "PLAN REVISED IMPLAN Revised 85 Railroads 0.4440 0.5593 0.9574 0.8362 0.5005 1.0000 1.0000 1.0000 86 Local Transit 0.5506 0.4673 0.0000 0.8000 0.3061 0.7262 1.0000 0.5187 87 Motor Freight 0.3911 1.0000 0.0000 1.0000 0.3574 0.9034 1.0000 0.8477 88 Water Transportation 0.4019 0.6727 0.0000 1.0000 0.3035 1.0000 1.0000 0.0000 89 Air Transportation 0.4228 0.2527 0.0000 0.3496 0.2219 0.4801 1.0000 0.2686 97 Wholesale Trade 0.6044 0.6274 0.2755 0.9704 0.5003 0.9949 1.0000 0.5789 98 Eating 8 Drinking 0.6020 0.8901 0.0000 0.8084 0.5001 0.8145 1.0000 0.9002 * RETAIL INDUSTRY (A000) 0.5596 0.9500 1.0000 0.9500 0.2404 0.9500 1.0000 0.94 106 Hotels & Lodging 0.6252 0.7232 0.0000 0.5205 0.1822 0.5230 1.0000 0.8000 111 Amusements 1.0000 0.8379 0.0000 0.8500 0.2848 0.8500 1.0000 0.8222 Ave All Service Sectors 0.6666 0.7787 0.5497 0.7938 0.3487 0.8299 1.0000 0.7492 STD DEV “ “ 0.2847 0.2095 0.4358 0.2216 0.1431 0.1789 0.0000 0.2571 * revised retail RPCs are based on aggregating four retail sectors New York and New Jersey are admittedly extreme illustrations of the problems with IMPLAN's current RPCs. However, all IMPLAN models suffer from poor RPC estimation for some sectors. As illustrated by tables 6 through 8 and more extensive analysis conducted for the Lake States, the RPC estimates could be improved by using the corrected MRIO data to re- estimate IMPLAN's RPCs. How large are the consequences of errors in IMPLAN's RPC estimates on impact estimates? If New York's "true" service sector RPCs are only the average of the other fifty states, then direct service sector 128 impacts may be estimated by IMPLAN at well under half of what they should be. In other words, more than two-thirds of New York demand for services is probably being met by New York service suppliers, rather than approximately only the one-third share IMPLAN currently estimates. In fact, it is likely that New York’s true RPCs are higher than average due to its economic size, diversity, and national importance. The revised average based on the corrected MRIO data (.8299, versus IMPLAN's unadjusted estimate of .3487) is much more in line with these factors. Indirect effects may be subject to worse distortions as the underestimation (or overestimation as is the case for North Dakota and certain other states) is compounded with multiple rounds of spending. At a minimum, indirect and induced portions of multipliers should be expected to have the same sign and somewhat reflect the average differences between RPCs based on the original MRIO data and the corrected MRIO data. Table 9 provides a contrast of State of Michigan multipliers for recreation-related sectors generated by IMPLAN using the two alternative sets of trade estimates. As expected, the resulting relative differences between revised and unchanged multiplier values displayed in Table 9 reflect differences between IMPLAN and revised Michigan RPCs presented in Table 8. For example, the mean Type III service sector output multiplier for the revised RPCs (1.81) is approximately sixteen percent larger than the mean Type III service sector output multiplier for the unchanged RPCs (1.56). This is very similar to the difference between the IMPLAN and revised average RPC for all service sectors (.6666 and .7787, respectively, as indicated in Table 8). 129 Table 9. Michigan Output Multipliers for Recreation-Related Sectors REVISED RPCs UNCHANGED RPCs SECTOR NAME TYPE I III TYPE I III 85 Railroads 1.38 1.74 1.29 1.55 86 Local Transit 1.19 1.61 1.12 1.43 87 Motor Freight 1.27 1.64 1.17 1.44 88 Water Transportation 1.47 1.68 1.32 1.45 89 Air Transportation 1.33 1.58 1.25 1.42 97 Wholesale Trade 1.28 1.44 1.14 1.24 98 Eating & Drinking 1.46 2.23 1.33 1.89 * RETAIL INDUSTRY (AGGD) 1.23 1.99 1.11 1.67 106 Hotels & Lodging 1.50 2.43 1.30 1.97 111 Amusements 1.34 1.97 1.21 1.66 Mean for All Service Sectors: 1.37 1.81 1.24 1.56 * revised retail RPCs are based on aggregating four retail sectors A. more complete RPC comparison is presented in .Appendix. B: "Comparison of IMPLAN, REMI, and Corrected MRIO RPCs." This appendix presents RPCs for all sectors in the three Lake States and the Lake State region based on the corrected MRIO data. It also presents IMPLAN and REMI RPCs based on the MRIO 124—sector aggregation scheme. These RPC calculations confirmed that the corrected MRIO data does not have the zero or near zero RPCs observed in the IMPLAN models and reported for IMPLAN service sector RPC values. Absolute and percentage differences were calculated between all IMPLAN and MRIO Michigan sectors and are presented in Appendix B. The average RPC difference across all sectors on an output-weighted basis was approximately .14. The corrected. MRIO Michigan. RPCs were 31.8 percent higher than the IMPLAN RPCs on a weighted average basis. Regressions were also conducted on the Michigan RPCs to determine the correlation between them at this level of aggregation. Quite surprisingly, the REMI and. corrected. MRIO RPCs had. an R2 of approximately .54 (with 94 degrees of freedom), while the IMPLAN and 130 corrected MRIO R2 was close to .23 (with 113 degrees of freedom). Differences in the degrees of freedom stem from RPCs not being available for REMI agriculture, construction, and government sectors. IMPLAN's R2 with the corrected MRIO data for only those sectors which also had REMI RPCs was .22. These regression results were unexpected as both the IMPLAN RPCs and corrected. MRIO RPCs were developed from the same original MRIO database, whereas the REMI RPCs were developed largely from. the Transportation. Census. These results do not confirm the accuracy of either the REMI or corrected MRIO RPC estimates, but they do imply that these models have similar RPC patterns that are currently not present in the IMPLAN model. Sector Allocation pi Outdoor Reopeation Spending Information from the U.S. Bureau of Economic Analysis (BEA) computer tape "Personal Consumption Expenditures and Gross Private Fixed Investment Item Detail" (1984) was used to bridge FHW recreation spending in this study's forested area of the Lake States. The detail contained on the BEA tape allows much more precise allocation of spending to IMPLAN sectors than the sector allocation bridges in the IMPLM Analysis Guide (USDA Forest Service, 1985) and published in conjunction with PARVS (Propst, 1988). These two bridge tables are based on approximately one hundred aggregated personal consumption expenditure (pce) categories. In contrast, the BEA tape contains 1,790 different consumption items and their corresponding SIC codes, total 1977 U.S. expenditures, and the breakdown of these expenditures between producer and distribution shares. 131 The pce I-O category #9400, "pce wheel goods, durable toys, sport equipment, boats & pleasure aircraft," will be used. to illustrate problems associated. with adopting aggregated I-O pce categories to allocate recreation survey spending. Pce I-O category 9400 contains the majority of equipment items associated with outdoor recreation. Table 10 presents a listing of the 108 items contained within the category and their corresponding pce item numbers. Trade and transportation margins and producer shares of expenditures for all 1790 pce items have been bridged to IMPLAN sectors by IMPLAN personnel. Additionally, these shares have been used to derive the percentage distribution of purchaser spending among producer and margin shares for the pee I-O categories and some subtotals within the categories. Table 11 presents the percentage distribution of spending on specific items within pce I-O category 9400 under the column heading "% OF PCE I-O CATEGORY 9400." Table 11 also shows the distribution of spending among IMPLAN sectors for three subcategories of pce I-O category 9400: small arms (pce item #1481), optical goods (pce items #1534-6), and boat building and repair (pce item #1507). The percentage distribution among IMPLAN sectors for these items was calculated by this author based on the detailed BEA national expenditure data. Table 11 also presents Lake State Type I and III output multipliers, based. on revised (BSTLK) IMPLAN RPCs, for the IMPLAN sectors associated with pce I-O category # 9400. Table 12 shows the results of combining the percentage distribution estimates and output multipliers for the four types of industry groupings (I-O category 9400, and the three pce item subcategories) 132 Table 10. Pce I-O Category 9400:Wheel Goods, Durable Toys, Sport Equipment, Boats 8 Pleasure Aircraft ITEM 8 1481: 1482- 1483 1484 1485: 1486: 1487: 1488: 1489: 1490: 1491: 1492: 1493: 1494: 1495: 1496: 1497: 1498: 1499: 1500: 1501: 1502: 1503: 1504: 1505: 1506: 1507: 1508: 1509: 1510: 1511: 1512: 1513: 1514: 1515: 1516: 1517: 1518: 1519: 1520: 1521: 1522: 1523: 1524: 1525: 1526: 1527: 1528: 1529: 1530: 1531: 1532: 1533: 1534: PCE ITEM DESCRIPTION SMALL ARMS TEXTILE BAGS CANVAS PRODUCTS, NSK CAMPING TENTS OTHER CANVAS PRODUCTS FABRICATED TEXTILES, NEC, NSK SLEEPING BAGS PARACHUTES BICYCLE CASINGSGSINGLE TUBE TIRES BICYCLE INNER TUBES BOATS, PONTOONS, LIFE RAFTS LEATHER GOODS N.S.K. LEATHER NOVELTIES SADDLERY B HARNESS G ACCOUTERMENTS DOG COLLARS, LEASHES 6 OTH PET ACCESS. OTHER LEATHER GOODS BEA COVERAGE ADJUSTMENT CUTLERY,SCISSWS,SHEARS,TRIIIIERS,& saws POCKET mvss aomsn FOLDING-BLADE mvss MARINE HARDWARE STEEL WIRE CAGES OUTBOARD ENGINES STORAGE BATTERIES-SLI-AIRCRAFT B MARINE STORAGE BATTERIES-SLI-AIRCRAFT 8 MARINE COMPLETE AIRCRAFT, PERS 8 UTILITY YACHTS, UNDER 65 FT BOAT BUILDING B REPAIRING, NSK OUTBOARD MOTORBOATS INBOARD MOTORBOATS, NSK INBOARD RUNABOUTS INBOARD CABIN CRUISERS, UNDER 26 FT INBOARD CABIN CRUISERS, 26 FT AND OVER HOUSEBOATS INBOARD-OUTDRIVE BOATS, NSK INBOARD-OUTDRIVE BOATS, UNDER 20 FT INBOARD-OUTDRIVE BOATS, ZOFT 8 OVER BOATS, NEC, NSK SAILBOATS OTHER BOATS, NEC BOAT REPAIR, NONMILITARY BEA COVERAGE ADJUSTMENT MOTORCYCLES, BICLES B PARTS, NSK BICYCLES, CGPLETE BICYCLE PARTS MOTORCYCLES AND TRAIL VEHICLES TRANSPORTATION NEC. NSK. PARTS FOR SELF-PROPELLED GOLF CARTS SELF-PROPELLED GOLF CARTS SELF°PROPELLED SNOWMOBILES PARTS FOR SELF-PROPELLED SNOWMOBILES BOAT TRAILERS ALL‘TERRAIN VEHICLES PARTS FOR ALL-TERRAIN VEHICLES WTICAL INSTRIHENTSBLENSES, NSK ITEM 3 1535: 1536: 1537: 1538: 1539: 1540: 1541: 1542: 1543: 1544: 1545: 1546: 1547: 1548: 1549: 1550: 1551: 1552: 1553: 1554: 1555: 1556: 1557: 1558: 1559: 1560: 1561: 1562: 1563: 1564: 1565: 1566: 1567: 1568: 1569: 1570: 1571: 1572: 1573: 1574: 1575: 1576: 1577: 1578: 1579: 1580: 1581: . USED ° USED PCE ITEM DESCRIPTION WTICAL INSTRLHENTS, CUP” LENSES, NSK BINOCULARS,OPTICAL ALIGNMENTEDISPLAY INSTRMNTS HAND HELD STILL CAMERAS FLASH UNITS-ELECTRONIC 8 NONELECTRONIC PROJECTORS, SLIDE 8 STRIP STILL PICTURE EQUIP,PARTS,ATTACH,GENLARGERS 8816 MM MOTION PICTURE CAMERAS 16MM SOUND 8 SILENT PROJECTORS ALL OTHER BMM PROJECTORS 8104 SILENT PROJECTMS LESS THAN S100 PROJECTION SCREENS 8&16MM MOTION PICTURE PARTS,ATTACH, ETC BABY CARRIAGES AND CHILDREN'S VEHICLES SPORTINGBATHLETIC GOODS,NEC. NSK GOLF BALLS AND GOLF CLUBS GOLF BAGS 8 OTHER GOLF EQUIPMENT HOME PLAYGROUND EQUIPMENT HEALTH,PHYSICAL FITNESSBEXERCISING EQUIP BILLIARD AND POOL, TABLES B SUPPLIES BOWLING BALLS TEAM SPORTS EQUIPMENT SIDEWALK AND RINK ROLLER SKATES ICE SKATES WATER SKIS AND SURFBOARDS SNOW SKIS WINTER SPORTS EQUIPMENT UNDERWATER SPORTS EQUIPMENT BEA COVERAGE ADJUSTMENT MISC. FABRICATED PRODUCTS, NEC., NSK OTHER MISC. FABRICATED PRODUCTS, NEC., NSK. MISC. FABRICATED PRODUCTS, NEC. WELDING REPAIR HOUSEHOLD FLIGHT INSTRUMENT REPAIR SNOWMOBILE REPAIR LOCK SMITHS, GUNSMITHS REPAIR SERVICE FOR OPTICAL GOODS MOTORCYCLE REPAIR HOUSEHOLD TENT REPAIR SPORTING GOODS INCLUDING BICYCLE REPAIR CAMERA AND PHOTO EQUIP REPAIR USED OPTICAL GOODS USED OPTICAL GOODS USED SPORTING GOODS USED SPORTING GOODS USED MOTORCYCLES THRU USED MERCH. STORE USED MOTORCYCLES THRU USED MERCH. STORE USED WHEEL GOODS SPORTS EQUIP, ETC USED WHEEL GOODS SPORTS EQUIP, ETC USED WHEEL GOODS SPORTS EQUIP, ETC USED WHEEL GOODS SPORTS EQUIP, ETC WHEEL GOODS SPORTS EQUIP, ETC PLEASURE BOATS PLEASURE BOATS THRU RETAIL BABY EQUIPMENT USED USED 133 presented in Table 11. For each of the four groupings, this process involved multiplying the two types of output multipliers by any proportional share of spending in each multiplier's corresponding sector. Thus, for example, railroads and related services (IMPLAN #446) multipliers of 1.45 and 1.92 were multiplied by .00069 for I-O category #9400, .00011 for small arms, 0 for optical goods, and .00065 for boat building and repair, resulting in the corresponding output impacts shown in Table 12. The estimates shown in Table 12's columns may be viewed as depicting the distribution of output impact among sectors per average dollar spent for the four pce groupings of industries. The sums of the columns at the bottom of Table 12 indicate the total output impacts for each of the four pce industry groupings according to type of multiplier (Type I or III). The corresponding percentage differences in total output impacts from pce I-O category #9400 are shown below the sums for the three pce subcategories. 134 Table 11. Contrast of Sector Allocations Based on I-0 Category #9400 * Versus Small Arms, Optical Goods, and Boat Building 8 Repair PCE Items Within 1-0 Category #9400 LAKESTATE %G %G BSTLKIM’UN %G %G PCEMS34-6 P631507 OJTFUT PEI-O ”#1481 CPIICAL swam BEA INISTRY MLTIPLIERS UTEGRY SILL ANS ms 8 REPAIR HRS N am IIPUN IMRY ME "PLAN # TYPE I III “N * (II? #79) ("P M) (IA? W) RAIL HRGIN mmos ND ELATE) saw-s 446 1.45 1.92 0.N9% 0.011% 0.00m 0.065% Tm MRGIN AUTO! FREIGIT 112nm 448 1.41 , 1.95 0.5” 0.” 0.09m: 0.717% MTER MARGIN ENTER TRANSPGITATIOI 449 1.6 1.91 0.156% 0.013 0.000: 0.229: AIR MRGIN AIR TRANSPCRTATICN 450 1.45 1.77 mm 0.000% 0.090% 0.130% PIPE MRGIN PIPE LIIES, EXCEPT NAT GAS 451 1.46 1.64 0.000% 0.” 0.110% 0.013 lllSALE ARIN-REC REGEATIOl-ELATE) UQESALE TRADE 460 1.36 1.61 6.%9% 11.55“ 5.750% 0.000% “GEM MRGIN OTIIER HQESALE TRIM 461 1.36 1.93 1.37.94 0.0007: 0.000% 7.54% ETAIL HEN-REC RECREATIOl-RELATE) RETAIL was 462 1.30 1.61 27.445% 34.&2% 34.N1% 0.0002: RETAIL MARGIN OTIER RETAIL m 43 1.30 2.31 2.65% 0.0002; 0.000: 0.0002: INSM MRGIN INSM CARRIERS 467 1.8 2.50 0.000% 0.” 0.000: 0.000% 13NN ML AIDS 79 1.32 1.91 3.4002; 53.536% 1911501 TEXTILE BAN 154 1.5 LE 0.“ 190302 WAS PROILTS 155 1.21 2.13 1.1m 1m FABRICATB) TEXTILE ms 159 1.24 1.62 0.629% 3N1N TIRES AN) IIIER TIBES 240 L3 1.73 0.130): 320.302 FABRICATB mm was 243 1.42 2.03 0.N3% 340505 LEATI'ER ms, N.E.C 54 1.50 2.77 LWX 4201N OJTLERY 319 1.57 2.07 0.791% 43300 WARE, N.E.C. 322 1.36 1.72 0.156% 4ZNN MISGLLNEIS FABRICATED WIRE PNC 38 1.40 1.93 0.032% 43mm INTERNAL m I01 EIKSIIES 31 1.57 2.14 1.51% 1.461.84 0.024% 1.411.84 3.114% 5mm m BATTERIES 600100 AIROTAFT 610100 SHIP anumo no REPAIR 1.34 1.78 0.1m: am am anmxm no mm 1.66 2.06 19.058: 91.5% 611500 mass, amass, no was 411 1.59 1.97 10.0567; 610700 TRANSPORTATIOI mum, N.E.C. 415 1.53 1.91 1.8.9: 630100 0mm msmmns no mm 425 1.23 1.53 0.504% 59.209): 6303(1) W1C scum no 61me 425 1.35 1.65 536% 640501 ones, TOYS, no cum mamas 431 1.48 2.09 1.544: 640400 mum no ATllETIC toms, N.E.C 433 1.40 1.89 6.840% 641200 mmm Imusmss, N.E.C. 445 1.42 1.98 0.100% §§§§ 730101 HIM REAIRSIKPS 478 1.18 1.50 L%7% 81m MAIDENDHNDGIDS 534 LN LN - 0.486% 9.118: LN LN LN LN l * I-0 category #940) is advised of poe meal goats, drdale toys, sport aqfiplant, boats so please aircraft 135 Table 12. Contrast of Lake State Output Impacts Based on I-0 Category #9400 Versus Small Arms, Optical Goods, and Boat Building 8 Repair PCE Items Within I-O Category B9400 IiDCAflEORYIBND SILLIIMS OPHCN.GEDS BOMIBUXIBIEHUR Oflflfl'flflflfl (UHUWIMWCT OflHfl'flENfl OUWUTIMWET IMHAN BMED(N WEEJON BNED(N BNED¢N IM’UNIMISTRYNNE 8 TYPEITYPEIIITYPEITYPEIIITYPEITYPEIII TYPEITYPEIII RAIUWS ND RELATE EVI‘ 446 0.N10 0.N13 0.00:2 OJ!!! 0 0 0.01” 0.N13 AUTO! REIGN 172nm: 448 0.032 0.0113 0.N11 0.N16 0.N13 0.N18 0.0101 0.0140 “TB! WATIOI 449 0415 0.0030 0 0 0 0 0.N37 0.N44 AIR WTATICN 450 0.” 0.” 0 0 0.N13 0.N16 0.N19 0.- PHEELHES,ENEPTAMT(HS 451 0 0 0 0 0 0 0 0 RECREATIOl-RELATE) WESAIE mos 4N 0.®2 0.1N7 0.1571 0.155 0.0m 0.0904 0 0 OUR HQESALE was 461 0.01N 0.66 0 0 0 0 0.09m 0.14M REOEATIOJ-ELATE RETAIL was 462 0.565 0.4410 0.4524 0.55% 0.4529 0.5% 0.0000 0.“!!! OTIER RETAIL TRNE 465 0.0345 0.N15 0 O 0 0 O 0 IERRNIEIURRHIS 467 0 0 0 0 0 0 0 0 “LL A106 79 0.0448 0.w1 0.757 LN44 TBOIUEUMS 154 (LOGS OJINB CNAMSFROIIJS 155 0JM04 (LEE! FAHUCAED'EKTHEEHKDUCEB 159 OJIWB DINO? TIRES ND "DER TLBES Z40 0.N18 0.” HBRHIWEJRUEERFROIIWS 243 OJIO9 OJIM3 LEATIR .6, N.E.C 54 0.0164 0.032 OJTLERY 319 0.0124 0.0164 mm, N.E.C. 322 0.091 0.027 MIW FABRICATED WIE Pill: 3 0.“!14 0.” INHRNM.OOEUSUONENGHES 331 (LUIS OJBB4 SKRNEEUNTEUES I55 OJIO4 (LOOK AIROZAFT 45 0.0439 0.572 SHIP NILDIIE AN) REPAIR 4N 0.0118 0.0156 EMT NILDIIB ND REPAIR 4N 0.3161 0.” 1.5191 1.0903 muss, amass, no ms 411 0.1595 0.1906 immune: acumen, N.E.C. 415 0.- 0.- O’TICAL INSTIUENTS no LENSES 4.8 0.0065 0.0179 0.7597 0.9336 WIC 00.1mm no arms 45 0.078. 0.0961 ones, 1073, no cums VEHICLE 431 0.- an mm no ATIIETIC ms, 77.2.0 433 mm 0.1292 mmm "names, N.E.C. 445 0.0014 0.0172!) MISILINEIS REPAIR SOS 478 0.- 0.” (E) All) ssmouno 530 0.N46 0.N46 8173: 1.43 1.84 L9 1.” L29 L59 LGS 2.5 % of I-O Category “N Input Sun: 92.27: %.4% N.6% $.5% 114.5% 111.7% 136 The percentage differences shown at the bottom of Table 12 indicate that some items within pce I-O category #9400 are associated with sectors that generate larger output impacts than the category as a whole, ‘while, other items are associated.'with sectors that generate smaller impacts. Specifically, 'boat 'building and repair has output impacts that exceed #9400's by more than ten percent, while the output impacts for small arms and optical goods are between five and fifteen percent below those for #9400. Thus, if only the pee I-O categories are used to bridge spending data, the error involved in measurement of total impacts will depend on what specific items are being considered. Pce I-O categories may also include producer sectors not represented. in. the regional economy. 'This is especially true for smaller regions. Sectors not represented in the regional economy will have no multiplier impact associated with them beyond what impacts are associated with their distribution margins. Bridging spending to such sectors results in the expenditure being eliminated from the subsequent impact analysis. Use of the pee I-O categories can err in both directions. Spending may be distributed to a range of sectors in the region when, in reality, the spending should go to a sector not in the regional economy (i.e., imports). Alternatively, spending may be distributed among a range of sectors not represented in the region when, in reality, it should all be allocated to an existing regional sector. Using the pce I-O categories to allocate spending invites errors of this nature. In contrast, use of the detailed pce items will enable more precise allocation of spending to sectors, whether they exist in the region or not. It should be noted, however, that bridging based on the detailed pce items assumes national margins are appropriate for the 137 sectors and region being modelled. This may not always be the case, but regional data on margins is seldom available to use in place of the national margins. Tables 10, 11 and 12 illustrate that it is essential to use the detailed pce item information if measurement of differential impacts among sectors is important. Different items have different margins associated with them and the pattern exhibited by the overall pce I-O category may vary greatly from individual items within the category. For example, there is no boat building and repair retail margin, while the margin for pce I-O category #9400 as a whole is above twenty-seven percent for recreation-related retail trade and over two percent for other retail trade (see Table 11). The direct impact to the boat building and repair sector would be understated by close to thirty percent if the I-0 category's margins were used in place of the BEA detailed margins for boat building and repair. Total impacts for the sector would be understated by an even greater amount, as the boat building (and ‘repair sector has larger output multipliers than the recreation-related retail trade sector. (See Table 11). Allocating the producer's share of expenditures to the proper sectors is especially important from the standpoint of sectoral impacts. The producer's share of consumer expenditures usually exceeds the combined share of distribution margins. (Table 11 indicates the approximate producer's share of small arms is 54%, for optical goods it is 59%, and for boat building and repair it is 92%.) The use of pce I-O categories distributes this producer's share of spending among a range of sectors which may not be at all associated with the particular spending under consideration, or worse, not exist in the region. 138 Appendices E and F, ("FHW Bridge Table" and "FHW Used Equipment Allocation Table," respectively) represent the work devoted in this research to more narrowly define the sector allocations of FHW spending to IMPLAN sectors. These tables and the computer spreadsheets they are based on may help to make the task of bridging spending to I-O sectors more accurate and less time-consuming. They need to be further reviewed and refined for greater ease of use however, as time constraints prohibited extensive review and refinement prior to the completion of this research. Lake State Outdoor Recreation Economic Impacts Differences fietweeg Lake State IHELAE flgggl BBQ; Five sets of RPCs were used to develop Lake States IMPLAN models. The models and what they represent are as follows: SDP - trade estimates are based on the supply-demand pooling trade estimation technique; these represent ceiling values for the IMPLAN RPCs FLRLK - RPCs are based on the minimum RPC values for the three Lake States, including questionable low RPCs ALTFLK - RPCs are the same as for FLRLK except for those RPCs which appeared highly contradictory to SDP, MRIO, REMI, and. Census estimates. .ALTFLK RPCs represent a more accurate set of minimum RPCs. BSTLK - this author’s best judgment of the approximate value of Lake State RPCs, based on MRIO and unchanged IMPLAN RPC values for the three individual Lake States and the Lake States as a whole, REMI RPC ‘values for' Michigan, and Lake State SDP and ALTFLK RPC values. UNCHLK IMPLAN (version 2.0) RPC estimates, without changes I It should also be noted that, with the exception of UNCHLK, all the models required adjustment to IMPLAN sectors 461 (other wholesale trade) 139 and 462 (recreation-related retail trade). These sectors have inappropriately low value added and output estimates in the microcomputer version of IMPLAN. (The problem with sector 461, which is of greater magnitude than the problem with sector 462, has been noted by IMPLAN personnel and has been communicated to IMPLAN users in IMPLAN News, September, 1989, p. 3). Following changes to sectors 461 and 462, some SDP RPC values were below the UNCHLK RPCs. This was particularly true for’ service sectors, where SDP ‘values subsequently"were often several percentage points below the UNCHLK RPCs. In this light, the SDP RPCs represent ceiling values for Lake State models with modifications to sectors 461 and 462. A more inexplicable occurrence was the generation of a few unchanged Lake State RPCs that were lower than any of the RPCs for the states of Michigan, Minnesota, and Wisconsin. This was true for sectors 162, 224, 401 and 402, although the differences for sectors 224 and 401 are only slight and can be discounted as possible rounding errors (see Appendix C). Theoretically, this should not be possible. Demand for the region can not be greater than the sum of the individual state's demands, whereas the region's intraregional trade can be greater than the sum of each state's intraregional trade. The intraregional trade value comprises the numerator of the RPC, while regional demand is the denominator. Thus, with the only possibility of change being an increase in the numerator (intraregional trade) value, Lake State RPCs should not be smaller than all three of the individual state's RPC values. This can be represented for any sector 1 as follows: Lake State intragregional trade - X}’1, Michigan intraregional trade - Xmi,mi’ 140 xmn,mn Minnesota intraregional trade - 1 Wisconsin intraregional trade - xgi’W1, The first superscript represents the producing region and the second represents the purchasing region; 1 - Lake States, mi - Michigan, mn - Minnesota, and wi - Wisconsin. Lake State intraregional trade includes trade that occurs between the Lake states, in addition to the intraregional trade within the individual states depicted above. Thus, 1,1 Lake State intraregional trade, Xi , encompasses: mi,mi mi,mn mi,wi x1 x1 x1 mn,mn mn,mi mn,wi x1 X1 x1 wi,wi wi,mi wi,mn X1 x1 x1 In contrast, there is no additional demand (D1) for the Lake States region beyond the sum of the individual states' demands: 0} - D91 + 0?“ + 0Y1. The RPC for any Lake State sector is equal to Xi’l/Di. As the sum of any positive proportions cannot be less than the least of those proportions, no Lake State RPC should be less than the least of the three RPCs for the individual Lake States. There were only four Lake State sectors exhibiting the lower RPCs. The sectors were not important sectors for this study and the discrepancies were not large. A number of statistical tools were used to measure differences betweeen the RPCs. These were described in chapter three (pages 80-81). Results of these measurements are presented in Table 13. 141 Table 13. MEASUREMENT OF LAKE STATE MODEL RPC DIFFERENCES Dependent variable - UNCHLK (unchanged IMPLAN RPCs) Model: SDP FLRLK ALTFLK BSTLK UNCHLK Easy: MEAN 0.6467 0.1699 0.2266 0.4435 0.4362 SQUARE ERROR 70.4155 73.6130 59.8931 11.1668 MEAN SQUARE ERROR (MSE) 0.1334 0.1394 0.1134 0.0211 ROOT MEAN SQUARE ERROR (RMSE) 0.3652 0.3734 0.3368 0.1454 CORRELATION COEFFICIENT (r) 0.6497 0.6791 0.6716 0.9133 STANDARD DEVIATION 0.3573 0.2240 0.2332 0.3150 0.3558 ROOT MEAN SQUARE (RMS) 0.7389 0.2812 0.3252 0.5440 0.5629 MEAN ABSOLUTE DIFFERENCE (MAD) 0.2168 0.2695 0.2491 0.0649 THEIL'S INEQUALITY INDEX (U) 0.2805 0.4424 0.3792 0.1314 Um, bias 0.3323 0.5086 0.3872 0.0025 Us, variance 0.0000 0.1245 0.1324 0.0785 Uc, covariance 0.6677 0.3669 0.4804 0.9190 U sum (-1.000) 1 l 1 l The measurement indices reported in Table 13 provide several indications of the differences between the alternative RPCs and unchanged IMPLAN Lake State (UNCHLK) RPCs. These indices consistently indicate the vector of BSTLK RPCs is the most similar to the UNCHLK RPCs. This is indicated by the close proximity of their means (and RMS values) to each other relative to the other three sets of RPCs. Their means fall almost exactly between the mean of .2266 for ALTFLK and the mean of .6467 for SDP. The small difference between UNCHLK and BSTLK RPCs is further indicated by several measures of distance between the alternative sets of RPCs and UNCHLK. These distance measures, including the MSE, RMSE, and MAD, have values for BSTLK RPCs that are less than half the other RPC sets' values. The similarity 'between. BSTLK. and 'UNCHLK is (also indicated by their correlation coefficient value of .9133 versus correlation coefficients closer to .65 for UNCHLK and the other sets of RPCs. The SDP, FLRLK, and ALTFLK RPCs appear quite similar in terms of 142 their absolute differences versus the UNCHLK RPCs on the basis of their MSE, RMSE, and MAD values. Theil's Inequality Index and its components conform to the results provided by the other measures. The closer to zero the U index is, the lower the inequality between the variables being measured. The BSTLK RPCs appear distinctively closer to the UNCHLK RPCs than the other sets of RPCs on the basis of its Theil Index. The BSTLK U value (.1314) is less than half of the SDP value (.2805), which is the next lowest U value. The Um component reflects differences in mean values; FLRLK has the greatest difference in mean value from UNCHLK and also the highest Um value. The US component reflects. variance differences; SDP and UNCHLK standard deviations are virtually identical and, therefore, the SDP Us value is zero (rounded to the nearest ten-thousandths). Finally, the BSTLK value for the Uc component would be expected to be highest as this component relates random error ‘not accounted for by ‘bias and variance sources of differences. Differences Between Lake State IMPLAN Model Multipliers It is important to what extent the differences between the sets of RPCs are reflected in subsequent multiplier and impact estimates. Differences between sets of multipliers were examined across the five different Lake State aggregation schemes used and two types of multipliers (IMPLAN's type I and III) for each of three types of economic variables (output, personal income, and employment). .Additionally, two other' measures, chi square and standardized. mean absolute differences, were used in measuring multiplier differences. Overall results were quite similar to those found for the RPCs. Only mean values and general indications of the differences are reported 143 here; a complete set of the measurement values generated is presented in Appendix H: "Lake State RFC and Multiplier Analysis." Table 14 presents mean multiplier values for the five different aggregation schemes, five different sets of RPCs and three different economic variables. BSTLK multipliers tended to be the most similar to the UNCHLK multipliers across all aggregation schemes and types of multipliers. Means and root mean square values were similar to the results for the RPCs. The BSTLK and UNCHLK values were closer to each other than values for the other models and fairly equidistant between the two minimum sets of multipliers (FLRLK and ALTFLK) and the SDP multipliers. Measures relating the distance between the vectors of UNCHLK multipliers and the corresponding sector multipliers for the other RPC models further indicate BSTLK multipliers were closest to UNCHLK multipliers. This is true also for the two additional measures not used with the RPCs, chi square and SMAD. BSTLK multipliers generally had distance measures less than half the value of the other models. There were inconsistencies. For example, the ALTFLK type I personal income multipliers have slightly lower MAD, SMAD, and chi square values than BSTLK for the completely disaggregated model, but not for any of the other aggregated models. Sometimes the SDP values were closest to BSTLK's distance measure values, more often the ALTFLK values were. Most correlation coefficients between, the other sets of multipliers and UNCHLK are higher than they were for the RPCs. They tend to be close to or above .90 with BSTLK correlation coefficients being the highest. This represents a significant increase for the UNCHLK, FLRLK AND ALTFLK models over their RPC correlation coefficient values which were in the neighborhood of .65. Table 14. Mean Values of Lake State Model Multipliers Multiplier: Model: Aggrggatign 502 Sector 308 Sector 159 Sector 31 Sector 16 Sector Multiplier: Model: Aggregation 502 Sector 308 Sector 159 Sector 31 Sector 16 Sector Multiplier: Model: Aggregation 502 Sector 308 Sector 159 Sector 31 Sector 16 Sector SDP OUTPUT TYPE I FLRLK ALTFLK BSTLK 1.19 1.17 1.17 1.17 1.18 144 UNCMLK PERSONAL INCOME TYPE I FLRLK ALTFLK BSTLK EMPLOYMENT TYPE I FLRLK ALTFLK BSTLK UNCNLK UNCMLK 50? 2.31 2.44 2.51 2.37 2.45 SDP 2.53 2.84 3.28 2.57 2.81 SDP 2.63 2.98 3.20 2.75 2.89 OUTPUT TYPE III FLRLK ALTFLK BSTLK 1.44 1.54 2.07 1.39 1.48 2.03 1.40 1.50 2.06 1.38 1.48 1.99 1.44 1.53 2.09 UNCHLK PERSONAL INCOME TYPE III FLRLK ALTFLK BSTLK 2.34 2.32 2.61 2.13 2.34 EMPLOYMENT TYPE III FLRLK ALTFLK BSTLK 2.38 2.45 2.57 2.27 2.41 UNCMLK 2.10 2.13 2.38 2.07 2.38 UNCHLK 2.20 2.33 2.43 2.20 2.39 145 As would be expected, multipliers generated from the two sets of minimum RPCs had the lowest standard deviations, reflecting their lower RPC values and resulting lower multiplier values. Theil's U Index results conform with the results provided by the other measures, indicating BSTLK multipliers were closest to the UNCHLK multipliers. However, there 'were inconsistencies across the 'various sectorization schemes and economic variables. For example, the Um value (reflecting bias) varied from a high of .5585 for the disaggregated BSTLK type I output multipliers to a low of .2766 for the same multipliers at a 16-sector aggregation scheme. For corresponding multipliers, FLRLK and ALTFLK Um values moved in the opposite direction, increasing from .7414 to .8462 and from .5681 to .7720 respectively. Similarly, the BSTLK Um value went from .3536 for the disaggregated type I employment multiplier to .0345 for the same multiplier at the 16- sector aggregation level, while the SDP Um went from a disaggregated value of .1779 to a 16-sector aggregation value of .4603. These results and other Um estimates are presented in Table 15. In summary, the measures used to analyze the sets of multipliers tended to conform with the RPC results. Differences between the RFC and multiplier measurements were more of quantity than quality. BSTLK multipliers appear to be closer to UNCHLK multipliers than the other multipliers. However, they do not appear uniformly across all sectorization schemes and sets of multipliers to be as dramatically close to UNCHLK multipliers as BSTLK RPCs are to UNCHLK RPCs. Table 15. Multiplier: Model: e 502 Sector 308 Sector 159 Sector 31 Sector 16 Sector Multiplier: Model: W 502 Sector 308 Sector 159 Sector 31 Sector 16 Sector Multiplier: Model: Aggregation 502 Sector 308 Sector 159 Sector 31 Sector 16 Sector SDP .6778 .7415 .8007 .6336 .6367 .7414 .7566 .7454 .8618 .8462 OUTPUT TYPE I FLRLK ALTFLK .5681 .6201 .6320 .7781 .7720 146 BSTLK .5585 .5161 .5808 .4585 .2766 PERSONAL INCOME TYPE I SDP .5281 .5744 .5056 .5760 .5181 SDP .1779 .4474 .4330 .5132 .4603 FLRLK .4728 .5067 .3933 .4357 .3832 .2989 .3795 .3553 .3067 .2358 ALTFLK .2861 .3799 .2947 .3138 .3173 FLRLK ALTFLK .1888 .2967 .2778 .2234 .2008 BSTLK .4219 .4332 .3865 .3981 .2158 EMPLOYMENT TYPE I BSTLK .3536 .2864 .2328 .1978 .0345 SDP .7967 .7505 .7413. .2463 .0675 .6800 .7957 .7581 .5391 .5232 Lake State Multiplier Um (Theil Index Bias) Measurements OUTPUT TYPE III FLRLK ALTFLK .5829 .7299 .6898 .4614 .4667 BSTLK .1583 .1988 .1247 .0008 .0355 PERSONAL INCOME TYPE III SDP .5495 .5986 .5449 .4057 .2470 SDP .2240 .5204 .4867 .5712 .5133 .5525 .5911 .5330 .5425 .5549 .4589 .5124 .4810 .4607 .3551 FLRLK ALTFLK .4448 .5138 .4697 .4627 .5124 EMPLOYMENT TYPE FLRLK ALTFLK .3848 .4671 .4313 .4005 .3273 BSTLK .3586 .3857 .3391 .0171 .0043 III BSTLK .3899 .2529 .2337 .1499 .0162 147 Theil Index results, along with similar changes in other measures, indicate the level of aggregation can influence results from the measurement indices across models, across economic variables, and between RPCs and their corresponding multipliers. This implies it would be appropriate to use a. battery of statistical measures (as most nonsurvey I-O researchers have) at different aggregation levels and with different economic variables to avoid spurious measurement results and to cover a wide range of possible I-O model applications. Esgimates a; Lake State Qutdoor Recreation Ecoaomic Impacta A complete listing of estimated economic impacts are presented in Appendix E: "Lake State Outdoor Recreation Impacts." The discussion here will summarize the estimates and differences between them according to variables considered in the analysis. Those variables include: the range and types of spending, I-O model sectorization schemes, I-O model RPCs, comparison with spending estimates generated for the 1987 Governors' Conference on Forestry, resident versus nonresident impacts, and recreation sectoral multipliers versus average multipliers. Range of FHW Impact Estimates by Major Spending Category Table 16 presents Lake State disaggregated BSTLK economic impacts for different categories of FHW spending. The impacts result from combining recreation spending profiles for the spending categories with BSTLK disaggregated model multipliers. The resulting pattern shown in Table 16 reflects the relative magnitudes of impacts across spending categories for other models as well. Spending categories other than "Nonresidents Only" include both resident and nonresident spending. Table Hunting, 148 16. Lake State Economic Impacts from Upper Lake State Fishing, & Wildlife-Associated Recreation, Based on Disaggregated, Adjusted RPCs (BSTLK) IMPLAN Model * Estimate: Multiplier: Spending Categopy Nonresidents Only Trip Spending Only Total less Spec Equip Total Estimate: Multiplier: Spending Category Nonresidents Only Trip Spending Only Total less Spec Equip Total Estimate: Multiplier: Spending Category Nonresidents Only Trip Spending Only Total less Spec Equip Total * OUTPUT (millions of 1982 Low Type I Type III Type I 241 355 659 1,180 1,664 2,547 1,720 1,638 4,046 2,078 2,915 5,906 PERSONAL INCOME (millions of Low Type I Type III Type I 59 91 158 278 427 621 451 655 1,120 595 831 1,714 EMPLOYMENT (thousands of full and Low Type I Type III Type I 5 6 12 21 30 47 29 41 70 34 48 89 included in any of the spending categories $) High Type III 950 3,703 5,779 8,117 1982 $) High Type III 242 947 1,608 2,337 part-time jobs) High Type III 17 66 99 126 Direct payments by region residents to other residents are not 149 Low impact estimates were consistently less than half the size of high impacts. This was true for all four spending categories and is in accordance with the differences between the initial low and high FHW spending estimates shown in table 4. Low to high ratios differed little by model aggregation, economic variable, or type of multiplier (I or III). The ratios differed somewhat by spending category. Across all economic variables, model aggregations, and types of multipliers, they averaged about 37% for nonresident spending, 46% for trip spending, 43% for total spending less special equipment, and 37% for total spending. It is significant that the high impact estimate exceeded the low impact estimate by more than the size of the low estimate. This was not an unexpected finding in light of the difference in the initial spending range estimates discussed in chapter 4; however, the impact estimates make the ramifications of the spending range more apparent. For example, estimates in Table 16 imply between 3 billion and 8 billion dollars of sales (in 1982 dollars) were generated in the Lake States by 1985 FHW recreation in the more heavily forested areas of the region, based on IMPLAN's type III multipliers. Decisionmakers are not likely to be satisfied with such a broad range and people may be hesitant to report it. There was not an overly conservative bias in the use of FHW statistical parameters; the range of impacts reflects underlying data sample sizes and associated uncertainty in activity and spending levels. These should not be ignored. For the activity and area examined in this study, more intensive sampling is required or, at a minimum, crosschecks with other studies and secondary data sources would need to be utilized, to reliably report sales estimates in a narrower range than the 3 to 8 billion dollars reported here. 150 Relationships Among Categories of Lake State FHW Recreation Spending Several points may be highlighted regarding relationships among the four categories of Lake State FHW recreation spending. Table 17 presents the percentage impacts for the three subcategories of spending comprise of total (resident and nonresident) Lake State FHW spending impacts. These are based on the disaggregated BSTLK model and are representative of the pattern for other models. An important point illustrated by Table 17 is the degree to which impacts will differ depending upon what spending is being measured. Comparisons across studies are made difficult by the studies not including the same types of spending in their analyses. Substantially different impact measurements may be developed depending upon the spending being considered. For example, one objective may be to measure gains to the economic base of the Lake States as indicated by nonresident FHW expendituress. Table 17 indicates such impacts are only between nine and fourteen percent of the total impacts associated with both resident and nonresident FHW recreation expenditures. (A qualification on this is that, due to lack of data, equipment expenditures are assumed to occur in recreationists' state of residence. Therefore, estimates of nonresident expenditures are biased downward.) If only trip spending is considered, then the nonresident share of spending is between twenty and twenty-five percent. This can be derived from Table 17 by dividing nonresident percentages by trip spending only percentages. The analysis for the 1987 Governors' Conference on Forestry did not include major durable equipment expenditures (such as boats and vehicles). This is quite comparable to Table 17's "Total less Spec Equip" (total spending 151 Table 17. Lake State Subcategories of Recreation Spending as a Percent of' Total Upper Lake State Fishing, Hunting, & ‘Wildlife- Associated Recreation, Based on Disaggregated, Adjusted RPCs (BSTLK) IMPLAN Model * OUTPUT (millions of 1982 $) Estimate: % of Low Disag BSTLK Sum % of High Disag BSTLK Sum Multiplier: Type I Type III Type I Type III Spending Categpry Nonresidents Only 11.64% 12.19% 11.10% 11.70% Trip Spending Only 54.53% 57.09% 43.12% 45.61% Total less Spec Equip 78.84% 81.00% 68.51% 71.20% Total . 100.00% 100.00% 100.00% 100.00% PERSONAL INCOME (millions of 1982 $) Estimate: % of Low Disag BSTLK Sum % of High Disag BSTLK Sum Multiplier: Type I Type III Type I Type III Spending Category Nonresidents Only 9.95% 10.97% 9.26% 10.34% Trip Spending Only 46.68% 51.44% 36.24% 40.51% Total less Spec Equip 75.78% 78.78% 65.33% 66.81% Total 100.00% 100.00% 100.00% 100.00% EMPLOYMENT (thousands of full and part-time jobs) Estimate: % of Low Disag BSTLK Sum % of High Disag BSTLK Sum Multiplier: Type I Type III Type I Type III Spending Category Nonresidents Only 13.56% 13.56% 13.30% 13.30% Trip Spending Only 63.45% 63.45% 52.27% 52.27% Total less Spec Equip 86.35% 86.35% 78.37% 78.37% Total 100.00% 100.00% 100.00% 100.00% * Direct payments by region residents to other residents are not included in any of the spending categories 152 minus special equipment) category, of which nonresident spending comprises about fifteen percent. Measured impacts also differ depending upon what economic variable and type of multiplier are being considered. For example, trip spending accounts for less than 40% of the high estimate for personal income type I impacts, in contrast to it accounting for over 60% of the low estimate for employment type I impacts. Influence of Sectorization Scheme on Lake State FHW Impacts Table 18 expresses as percentages the ratios of BSTLK impact sums for the various sectorization schemes relative to the impact estimates for the completely disaggregated, 502 sector BSTLK model. Percentage figures presented in Table 18 indicate the various sectorization schemes used in this analysis had little effect on most impact estimates. Part of the differences may stem from the aggregation of recreation trade sectors with other trade sectors to conform with the 1987 Governors' Conference study. This was done for all aggregation schemes other than the fully disaggregated scheme. The unchanged IMPLAN model has errors associated with its "other wholesale trade" and "recreational related retail trade" sectors, as mentioned previously. The adjustments to these may have affected the resulting aggregated wholesale and retail trade sector multipliers relative to the disaggregated multipliers. The only other minor note on the influence of the sectorization schemes is the difference between the low and high estimates. In particular, the 159 sector model, which consists of many relatively disaggregated sectors and one large "all other" sector, has lower "high" output and personal income multipliers. However, overall, the different aggregation schemes tended to produce quite similar results. 153 Table 18. Impacts of Lake State FHW Recreation Spending for Different Sectorization. Schemes as a Percent. of the Disaggregated, Adjusted RPCs (BSTLK) IMPLAN Model * OUTPUT (millions of 1982 $) Estimate: % of Low Disag BSTLK % of High Disag BSTLK Multiplier: Type I Type III Type I Type III Aggragation eme 308 Sectors 98.81% 101.21% 98.87% 101.88% 159 Sectors 98.78% 101.54% 86.35% 90.62% 31 Sectors 99.90% 102.10% 92.31% 96.37% 16 Sectors 100.67% 101.20% 91.42% 94.09% PERSONAL INCOME (millions of 1982 $) Estimate: % of Low Disag BSTLK % of High Disag BSTLK Multiplier: Type I Type III Type I Type III Aggregation Scheme 308 Sectors 99.39% 101.55% 99.35% 102.13% 159 Sectors 97.53% 101.87% 86.01% 91.39% 31 Sectors 101.38% 102.90% 98.83% 100.84% 16 Sectors 94.17% 96.77% 90.48% 93.56% EMPLOYMENT (thousands of full and part-time jobs) Estimate: % of Low Disag BSTLK % of High Disag BSTLK Multiplier: Type I Type III Type I Type III Aggpegation eme 308 Sectors 108.91% 108.25% 111.72% 111.05% 159 Sectors 107.67% 107.00% 101.34% 100.71% 31 Sectors 109.41% 107.95% 109.04% 107.58% 16 Sectors 104.31% 103.13% 103.01% 101.85% * Direct payments by region residents to other residents are not included in any of the spending categories 154 Influence of RPCs on Lake State FHW Impacts RPC influences were very similar across the four major spending categories. Table 19 shows the percentage differences of total impact estimates for the alternative RPC disaggregated models from disaggregated, unadjusted IMPLAN (UNCHLK) total impacts. The most extreme percentage differences for any of the four spending categories were incorporated into the table. Thus, Table 19 indicates BSTLK impact estimates ranged between ninety percent and one hundred and seven percent of the UNCHLK impact estimates across all spending categories, economic variables, and multipliers. The percentage figures in Table 19 indicate impacts associated with supply-demand pooling (SDP) trade estimates were consistently the largest, while impacts associated. with FLRLK ‘were consistently the lowest. This is what was expected in light of the RPC and multiplier statistical analysis (e.g. SDP means were highest, while FTRLK's were lowest). ALTFLK model impacts were slightly higher than FLRLK impacts, reflecting increases in approximately ten percent of the RPCs from their zero or near-zero FLRLK values. BSTLK impacts were slightly higher than the unchanged IMPLAN impacts, except for type III output and personal income impacts. (The lower type III BSTLK output and personal income impacts are due to the errors associated with unadjusted IMPLAN trade sectors described previously.) The relative closeness of BSTLK and UNCHLK impacts could have been predicted on the basis of their close RPCs and multipliers. In this regard, it is important to point out that IMPLAN's unadjusted RPCs for the Lake States are not as suspect as they are for many other states and for smaller regions within the Lake States. For example, Michigan, 155 Table 19. Impacts of Lake State FHW Recreation Spending for Different Sets of RPCs as a Percent of the Disaggregated, Unadjusted (UNCHLK) IMPLAN Model * OUTPUT (millions of 1982 $) * Direct payments by region residents to included in any of the spending categories other residents are not Estimate: % of Low UNCHLK % of High UNCHLK Multiplier: Type I Type III Type I Type III RPCS: SDP 116.85% 109.37% 119.11% 111.15% FLRLK 79.50% 61.07% 79.55% 60.75% ALTFLK 83.13% 64.81% 83.21% 64.49% BSTLK 103.44% 91.25% 103.79% 90.14% PERSONAL INCOME (millions of 1982 $) Estimate: % of Low UNCHLK % of High UNCHLK Multiplier: Type I Type III Type I Type III RPCS: SDP 117.97% 110.68% 120.18% 112.51% FLRLK 84.17% 63.71% 83.80% 62.89% ALTFLK 89.71% 68.97% 89.42% 68.15% BSTLK 105.47% 92.80% 105.48% 91.52% EMPLOYMENT (thousands of full and part-time jobs) Estimate: % of Low UNCHLK % of High UNCHLK Multiplier: Type I Type 111 Type I Type III RPCS: SDP 114.93% 120.18% 117.89% 123.28% FLRLK 88.48% 77.77% 87.72% 77.11% ALTFLK 93.10% 82.91% 92.96% 82.79% BSTLK 105.45% 106.29% 106.25% 107.09% 156 and subregions within the state of Michigan, will have greater differences in impacts for adjusted RPCs.than were found for the Lake States, corresponding with proportionally greater numbers of sectors having erroneous (zero and near-zero) RPCs. Comparison to Outdoor Recreation Impact Estimates Prepared for the 1987 Lake State Governors' Conference on Forestry a. Updated 1987 estimates Profiles of' nonresident and combined resident and nonresident outdoor recreation spending were used for the 1987 Governors' Conference on. Forestry (Pedersen, Chappelle, and Lothner, 1989). These were converted from their original disaggregated IMPLAN version 1.1 format to four of the version 2.0 aggregation schemes used in this study. The profiles were not converted to the fully disaggregated sectorization scheme as to (k) so would have required making several arbitrary assumptions regarding how to disaggregate the IMPLAN version 1.1 sectors to the version 2.0 sectors. The translation of the profiles to the aggregated schemes was based on a perfect correspondence between the version 1.1 sectors which were aggregated and the resulting version 2.0 sector. Changes in the IMPLAN modeling system and different sectorization schemes resulted in modest changes in total estimated impacts from those reported in Pedersen et al. (1989). The original estimates and BSTLK re-estimated estimates are presented in Table 20. The impacts reported are based on IMPLAN's type III multipliers. The re-estimated BSTLK impacts were quite Similar across the four sectorization schemes. They differed from each other by less than five percent; Table 20 presents their average. 157 Table 20. Comparison of Estimated Type III Impacts Based on 1987 Recreation Spending Profiles OUTPUT PERSONAL INCOME EMPLOYMENT (million 1982 $) (thousand jobs) Nonzeaidenpa Only Original 1987 Gov. Conf. Est. 1223 340 25.5 Re-estimated Average 1399 369 25.6 12531 Original 1987 4245 1143 85.2 Gov. Conf. Est Re-estimated 4729 1274 88.1 Average b. Comparison of Governors' Conference and FHW Impact Estimates Table 21 relates the four FHW spending impacts to the re-estimated 1987 Governors' Conference impacts in terms of the farmer's percentage of the latter. The FHW nonresident category is stated as a percentage of the Governors' Conference nonresident impacts, whereas the other three categories'impacts are related as a percent of the total resident and nonresident Governors' Conference impacts. There are several differences between the FHW recreation economic spending data and the 1987 Governors' Conference data. The latter incorporated many more types of recreation than the FHW data, but for an area in the Lake States that was about ten percent smaller than the FHW study area. Major durable equipment expenditures were not included in the 1987 data; however, there were some nonresident equipment expenditures in the area that were included. Also, the 1987 analysis used point estimates rather than a range of estimates. 158 The range of FHW impact estimates appear reasonable relative to the 1987 estimates, given the differences between the 1987 analysis and the FHW data used here. FHW nonresident impacts range between 25% to about 65% of the 1987 nonresident spending. The lower FHW estimates are due to three factors. The 1987 data included more recreation activities; it included some nonresident expenditures on equipment; and the nonresident share of FHW activities was lower than the 20% to 25% share of recreation activities attributed to nonresidents in the 1987 study. The FHW spending category which best mirrors the spending items included in the 1987 analysis is the category of total spending minus special equipment. (The FHW "special" equipment category is largely comprised of major durable equipment expenditures.) This FHW category ranges from a low estimate of about 50% of the 1987 total recreation economic impact estimate, to a high estimate above the 1987 estimate. This illustrates that fishing, hunting, and wildlife-associated recreation comprise a large share of outdoor recreation in the forested areas of the Lake States. Table 21. BSTLK % of 1987 Lake State Outdoor Recreation Impact Estimates OUTPUT PERSONAL INCOME EMPLOYMENT Low High Low High Low High Nonresidents Only 25% 67% 25% 65% 25% 64% Trip Spending Only 35% 77% 34% 75% 33% 73% Total less Spec Equip 50% 123% 53% 130% 49% 116% Total 63% 176% 68% 193% 59% 158% Comparison of Recreation Multipliers to "Average" Multipliers Multipliers associated with recreation spending were compared to average multipliers in order to relate recreation spending impacts to average sectoral impacts. .Average sales, employment, and. personal 159 income multipliers for the BSTLK, 31-sector model were derived and then multiplied by four categories of recreation spending. The four categories included.'both types of spending considered in the 1987 Governors' Conference analysis (nonresident spending only and both resident and nonresident spending) and low and high estimates of nonresident FHW spending. The resulting impact estimates were then compared to the sums that resulted by multiplying sector-specific spending by sector-specific multipliers for the same four recreation spending categories. Table 22 presents a comparison. of the impact estimates. The comparison includes absolute impact magnitudes and the impacts derived from sector-specific spending and multipliers expressed as a percentage of the impacts derived from total spending and average multipliers. This ratio of impacts provides an indication of the size of outdoor recreation impacts relative to the average for the rest of the economy. The results conform with the expectation that outdoor recreation generates many more jobs than the average sector, but lower amounts of personal income. This is due to the disproportionate amount of outdoor recreation spending in service sectors which typically employ more people, but at lower wages. It is worth noting, however, that the . personal income type III impacts are higher than the average sector impacts (see Table 22). This implies the induced effect associated with recreation sectors more than offsets the relatively lower direct and indirect personal income. Slightly more output than average was generated by the sector-specific outdoor recreation impact estimates, but the difference was not not appreciable. 160 Differences between the impacts associated with using sector- specific ‘multipliers versus average multipliers also illustrate the error that may be introduced by borrowing "average" multipliers from other studies, or applying the same multiplier across different economic variables (e.g. using a sales multiplier for employment or income impact estimates). Table 22. Model and Revised (BSTLK) IMPLAN RPCs 1987 GOV CONF NONRESIDENTS ONLY TYPE I TYPE III (1982 million 3) MULTIPLIERS USED AVERAGE: SECTOR SPECIFIC: RATIO OF SECTOR- SPECIFIC T0 AVERAGE:106.9% OUTPUT 896 958 1224 1404 114.7% (1982 million 8) MULTIPLIERS USED PERSONAL INCOME AVERAGE: 267 358 SECTOR SPECIFIC: 245 370 RATIO OF SECTOR- SPECIFIC TO AVERAGE: 91.8% 103.1% (jobs) EMPLOYMENT MULTIPLIERS USED AVERAGE: 13471 18792 SECTOR SPECIFIC: 18310 25544 RATIO OF SECTOR- SPECIFIC T0 AVERAGE:135.9% 135.9% 1987 GOV CONF TYPE III OUTPUT 4184 4741 3062 3195 104.3% 113.3% PERSONAL INCOME 911 1225 852 1285 93.5% 104.9% EMPLOYMENT 46035 64222 63476 88554 137.9% 137.9% TYPE I 108.3% FMU LOU EST 80TH RES 8 NONRS NONRESIDENTS ONLY TYPE I TYPE III OUTPUT 306 356 224 243 116.2% PERSONAL INCOME 67 90 60 92 90.4% 102.4% EMPLOYMENT 3368 4699 4632 6463 137.5% 137.5% Contrast of Impacts from Using Average Vs. Sector-Specific Multipliers for Recreation Spending, Based on a 31 Sector FMU MIGM EST NONRESIDENTS ONLY TYPE I TYPE III OUTPUT 613 657 107.1% 838 955 114.0% PERSONAL INCOME 183 245 164 247 89.6% 100.7% EMPLOYMENT 9221 12864 12240 17075 132.7% 132.7% CHAPTER VI SUMMARY, IMPLICATIONS, AND CONCLUSION t c o This study has addressed issues relating to nonsurvey I-O models, estimation of outdoor recreation economic impacts using IMPLAN, and specific factors influencing the size of lake State outdoor recreation economic impacts. Implications of this study for nonsurvey I-O models and specifically for the IMPLAN modelling system will be summarized first. This is followed by a presentation of conclusions regarding the measurement of outdoor recreation impacts, as illustrated by the measurement of impacts for the Lake States. Several further research needs are summarized at the end of the chapter. Use 0 Nonsurvey I-O Mpaala The nature of the debate regarding nonsurvey models is conveyed in Jensen's comments regarding the five "ready-made" modeling systems reviewed by Brucker et al. (1987, p. 21): "In the evaluation of any method of economic model compilation there can, at the bottom line, be only two fundamental questions of concern, namely does the method produce a :model ‘which is representative of reality within professionally acceptable limits and do the results of the ,model have professionally acceptable levels of integrity in the real world?" Jensen goes on to pose a series of eight questions specifically aimed at nonsurvey models. None of these eight questions, nor his "two fundamental questions of concern" address whether nonsurvey models will ‘be used -- let alone experience an increase in use -- or whether they can be improved. Most literature on nonsurvey I-O does not address these questions. The answer to both of these unaddressed questions is "yes." 161 162 Nonsurvey models will be increasingly used partly because of advancements in computer hardware and partly due to improvements in input-output model software. The speed at which computer developments have occurred was not envisioned even as recently as 1985, when Richardson predicted the future lay with hybrid (partial survey) I-O models. The high costs, in terms of both time and money, for conducting surveys are additional negative factors mitigating against their use. It now appears the greatest number of LG applications will not entail developing I-O models using a mix of survey and nonsurvey methods. Future I-O applications largely will be done entirely with nonsurvey models due to the advent of high speed, powerful computers and inexpensive, user-friendly modeling systems. Relatively inexpensive, "ready-made," input-ouput modeling systems are now available that operate on personal computers. A prime example is the USDA Forest Service's IMPLAN model which is receiving wider and wider use. At a cost of only a few hundred dollars, regional economic models can often be generated with IMPLAN software in under a few hours. In the face of a proliferation of nonsurvey models, the practical issue for regional economists is how to improve these models and their use, not whether survey I-O tables are superior to nonsurvey tables. The issue of survey model superiority is essentially moot if funding is not available for them and nonsurvey models and nonsurvey applications are proliferating. The gap is likely to increase between the costs of conducting primary surveys and the costs of simply using a nonsurvey model. At the same time, decisionmakers generally will not be concerned or knowledgeable enough to distinguish between the sources of impact 163 estimates, nor be able to differentiate between experienced I-O analysts and LO novices. The paramount role regional economists can play in this situation is to help refine the reliability of LG models that will be used. Advice from regional economists is also appropriate on how to best utilize scarce survey funding to supplement or crosscheck nonsurvey model estimates. For example, subject to funding constraints, priorities for surveying should include sectors which are the focus of the study (e.g.- hotels and lodging or eating and drinking establishments for recreation and tourism studies), sectors which account for large proportions of the region's economic activity, and sectors which have peculiar values in the nonsurvey model relative to other sectors in the model or relative to the same sector in other regions or studies (e.g. the zero-RPC sectors in the current IMPLAN model). It is also appropriate to encourage surveys in the case of modeling small regions, such as cities or counties. It is generally more feasible to survey a smaller region's establishments and there is a greater likelihood of error in nonsurvey I-O extrapolations from national averages to small regions. However, advice on expeditious surveys should be accompanied by research and advice on other means to improve nonsurvey models and their use. Questions such as Jensen's "do the ready-made methods satisfactorily fulfill our professional obligations and standards in producing reliable and high quality advice?" (Ibid., p. 21) miss the point that nonsurvey models and methods are not static. The current performance of nonsurvey models is not the only practical concern; there is evidence that the models may be made more reliable. The focus of 164 attention should be on how the models and their use can be improved, especially if it is true that many or most I-O applications will be carried. out on ready-made models by persons without extensive I-O knowledge. Regional economists' "professional obligations," to the extent they exist, are not static. Nor do they exist in a vacuum. Regional economists should be responsive to and help mold the direction applications are proceeding. This would be a more substantial contribution to furthering I-O analysis than comments on the current status of I-O models or bemoaning the fact that certain ideals (e.g. all I-O models be based on surveys) will not be achieved. MEL-LN Nonsurvey methods are being refined. Richardson (1985, p. 623) noted that the RPC approach was "a welcome change from endlessly repetitive and mechanical location quotient methods..." The adoption of an alternative RPC estimation procedure in the IMPLAN modeling system represented a further advancement in terms of breaking new ground. There are many other developments associated with the IMPLAN modeling system that have enhanced its use as an impact measurement tool. The personal computer version incorporates a wide array of user-friendly features. To name just a few of these, they include: the ability to generate a vast array of user-selected reports; a fast, user-friendly sector aggregation module and data editing software; and the disaggregation of sectors that are particularly useful for specific applications (recreation-related retail and wholesale trade, for example). In addition, there are continually-improving training workshops and materials for first-time users. These include the provision of sector- specific deflators and personal consumption expenditure data. 165 Another development with the IMPLAN model is that it is being converted to a social accounting matrix (SAM) format. Adelman and Robinson (1986) explain the distinction between SAM and I-O accounting formats in the following manner: "A. standard input-output model includes the intersectoral flows of intermediate inputs, and so captures one major source of linkages in the economy. However, the input-ouput model ignores the flows from producing sectors to factors of production (value added), and then on to entities such as government and households, and finally back to demand for goods. A Social Accounting Matrix (SAM) expands the input- output accounts to include a complete specification of the circular flow in the economy." (p. 4) Adoption of the SAM format will permit better illumination of transfers between institutions and facilitate analyses that are constrained by the traditional I-O accounting framework. Along with continually increasing enhancements in computing power and storage capabilities, the IMPLAN modeling system has made it relatively easy to generate economic impact estimates in very short periods of time. However, there are many problems remaining with the reliability and interpretation of such estimates for regional economists to address. These problems stem from both the accuracy of input data used with IMPLAN and the construction of the modeling system itself. Input data pitfalls are substantial, especially for outdoor recreation applications. Problems include the very basic delineation of activities that constitute outdoor recreationl and identification of sector-specific spending associated. with. those activities. Accurate recreation participation and spending profiles need to be constructed before reliable recreation economic impacts may be estimated. Spending profiles may then be converted into final demand vectors which, in turn, "drive" input-output (I-O) models. Thus, the levels of direct, 166 indirect, and induced sales, value-added, income, or employment impacts generated by a recreation I-O analysis critically depend on participation and spending estimates. I As noted in -the first chapter, the reliability of recreation participation levels and spending profiles are probably at least as important and in need of improvement as the input-output model used to generate the economic impact estimates. The examination here.:3£:FHWI activity participation and spending support this contention:;:: The. greatest range in economic impacts for the Lake States (besidesnthe issue of what spending categories were included in the impacts) stemmed from low and high estimates of spending which reflected limited surveyir sample sizes. This is an acute problem for many recreation studies. 0n the other hand, problems with an LG model sometimes may;be identified and rectified, resulting in universal improvements forrfuture users of the I-0 model system. It is argued here that potentiallyslargen: improvements in the accuracy of IMPLAN impact estimates can be achieyedxi by making more detailed sector allocation bridge tables available taria“ IMPLAN users and by re-calculating trade estimates used in the IMPLAN'f model system. More narrowly defined sector-specific deflators are are:- ponssible further enhancement. . “.-. is Bridge tables devised as part of this research are preliminarys--and_:.:1.i ____.._.-..._. r“- "'—‘ need further review and refinement before they are widely distributed. .-if However they, or other sector allocation schemes based on BEA's detailed as pce worksheets, will be far superior to other bridge tables that are aggregated at the pce I-O category level. Many sectors unrelated to thearec spending being bridged will likely be included at the gross Irv-O category Lci level, reducing the precision of sector spending allocationand. c 167 subsequent impacts. Simply removing the obviously inappropriate sectors .../‘- does not address their influence on distribution margins. The solution is to use disaggregated, detailed pce items and aggregate these where necessary to fit survey spending categories. This issue is further discussed and illustrated in a paper by Propst and Siverts (1990). Many RPC values currently being generated by IMPLAN are highly questionable and may lead to significant distortions in economic impact estimates. This is especially true for particular industries, including many sectors affected by recreation. The extent of distortion and sectors affected differ from state to state. Current and alternative IMPLAN RPC estimates were examined for different sectors and regions. Alternative trade estimates were developed and used in the generation of IMPLAN models for different-sized regions. The influence of trade values on impact estimates was then observed. Results indicate IMPLAN's trade estimates should be revised, incorporating data from the reconciled 1977 MRIO database. New trade estimates can be calculated for the entire IMPLAN modeling system with little difficulty. This was illustrated by calculations of RPCs for the service sectors of all fifty states and for all economic sectors in the Lake States. It should be noted that, for any particular region, means probably do not exist to establish which industries will have high or low RPCs on a completely objective, secondary basis. Differences in regional economies, trade relations, and other variables prevent any possibility of establishing an absolutely precise nonsurvey means ’for determining RPCs. It will always be appropriate for impact analysts to review their estimates carefully and take advantage of opportunities to crosscheck the estimates against other data sources. Industry associations, state 168 and local government personnel, and other federal or private information sources may provide additional direction for the task of improving RPCs and subsequent impact estimates. Similar qualifications appLy to the use of standardized bridge tables and sector-specific deflators. Outapo; Recreation Ecanonig Impaata Several points originally presented in chapters one and two warrant review here. The size of estimated outdoor recreation impacts will depend on a number of factors besidaswthe_quality-of_thaginpntmdataand_ I-O model. A major influence will be the objectives of the study which, in turn, will determine the definition of outdoor recreation and what recreation activities are included in an analysis. Two types of objectives may be differentiated. One type of objective is to estimate economic impact benefits (sometimes mistakenly identified as total impacts) associated with recreationists' spending. The purpose of impact analysis under this objective is to quantify the economic effects associated with recreation, but only _asuwnarrowly defined by resident and nonresident recreationistsf expenditures. Such an. analysisb should. not be considered. as providing a comprehensive description of recreationfs contribution to a regional economy. It only ...-...... H-‘i describes economic impacts associated with a particular configuration of " economic sectbrs linked to recreationists' spending. 1 I H A. second. type of objective is t6“ gain insight into the contributions of outdoor recreation to a regionis economic development. This objective entails a more difficult, challenging undertaking for it requires more extensive analysis. Following economic base theory, which maintains that exports provide the basis of a region's economic growth, 169 the analysis may only include nonresidents' recreation expenditures. However, excluding the examination of resident spending may wbe a ,_,_..— 2 mistake. There may be significant impact differentials between sectors H-_?‘_ and ‘between subregions within the region being analyzed from the transfer of spending by resident recreationists. Adso, the issue of import substitution -- in this case, recreating more within the region rather than outside of it -- may also be important. In. any case, the objective of analyzing recreationfs economic development potential requires more of a comparative analysis or net benefit approach. Costs incurred to provide the recreation experience need to be considered and compared to benefits. These costs encompass a myriad of governmental costs. Opportunity costs to extractive industries and other affected private interests should also be considered. A truly comprehensive effort will address social costs, such as analyzing changes in the size and composition of a community and impacts relating to community identification. Resource and institutional constraints may also be important. Consideration should be given to distributional effects. These may be partly analyzed within the context of an I-O analysis if the household sector is disaggregated according to income (as it is within version 2.0 of IMPLAN). In sum, simply using an LC model to express economic impacts stemming from recreationists' spending is inadequate if the goal is to assess the economic development contribution of outdoor recreation. 1..., The use of social accounting, matrices ("SAMs") also will not provide a comprehensive picture of economic development issues. IMPLAN's SAM format will present a greater elaboration of economic flow relationships between institutions. It will offer more insights into 170 some distributional and governmental spending issues than a standard I-O table. It will remain limited, however, due to data constraints. The more extensive SAM format will use data currently incorporated in IMPLAN's I-O format, plus additional secondary data on transfers between institutions. In this author's opinion, the SAM's greater elaboration of economic relationships between institutions will be based on data that are probably less reliable than data used as the basis for IMPLAN's current nonsurvey I-O tables. For example, secondary data on transfers between institutions are more limited and less subject to crosschecking than data on sales and employment. At a minimum, the more elaborate SAM framework entails more opportunity for error. Also, extensive examination of the economic relationships depicted in SAM models, such as those between value added and final demand sectors, will not be forthcoming for some time. Furthermore, the SAM, like I-O tables generally, will only depict market transactions. Important nonmarket factors, including those that feed back upon markets such as amenity influences; on 'business location. decisions, will remain outside the f’ " models. Influences an the Size pf Lake State Qutaoor Recreapign Impacts Some of the results of this study apply only to the Lake States and particular recreation activities in the study's designated forested area of the Lake States. In particular, many of the results regarding Lake State outdoor recreation impacts may not apply to, other regionsor models with different economic structures and sector aggregations. However, influences on the size of Lake State outdoor recreatipn_impacts- may also influence impacts for other regions. An examination of these 171 influences will indicate their relative importance for consideration by recreation analysts. The results presented in chapter five indicated that the small sample size of recreation participants contributed to the greatest variability in impact measurement. Another critical issue was the specific categories of spending being considered.— Forexample, therewas a significant 7 difference found Ibetween-vtrip-‘related spending versus spending which included equipment expenditures as well. Trip-related spending was less than half of total FHW-reported spending. The influence of RPCs was not large for the Lake States. This was indicated by several statistical measures of differences between the unadjusted (UNCHLK) and adjusted (BSTLK) multipliers and small differences between these two models' summary impact estimates. Wisconsin and Minnesota had relatively high unadjusted service sector RPCs, while Michigan's were not extremely low either. Differences between multipliers based on alternative trade estimates were found for individual sector impacts. However, multiplier differences between the unaltered RPC model and the "best" RPC model were not substantial. Errors in IMPLAN's RPCs for the Lake States model involve both underestimation and overestimation. The overestimated and underestimated iRPC-s will p'arti‘allymnegate each other when multipliers are derived. Larger RPC influences were found for the State of Michigan. The evidence presented on average service sector RPCs indicates much greater RPC influences on impacts are likely to be found for other states, such as New York and New Jersey, which have RPC estimation error more uniformly underestimated or overestimated. 172 IMPLAN models of small, substate areas may also have more bias introduced to their multipliers from inaccurate RPC estimates. Small, substate economies will tend to have a greater representation of service sectors as, a proportion of the total number of sectors in their economies. Errors in IMPLAN's RPC estimates appear to be more pronounced with service sectors than commodity sectors. This is true in part because a higher proportion of service sector demands are generally met by local production than are commodity demands. This means that average service sector RPCs should be expected to be higher than average commodity sector RPCs. Therefore, zero or near-zero service sector RPCs are more divergent from true, real-world values than zero or near-zero commodity RPCs. Differences in the derivation of IMPLAN's RPCs may account for there appearing to be more dubious service sector RPCs than commodity sector RPCs. Commodity sector RPCs were derived econometrically, service sector RPCs were adopted directly from the original Jack Faucett Associates 1977 MRIO data. Aggregation of sectors also did not appear to significantly affect impact results. Differences between impact sums were generally under ten percent. The only slight exception to the lack of any pattern being observed was with regards to the aggregation scheme which lumped all ...-.....r - v“.-—-om4'- we...” .m‘u v nonessential sectors into one large "miscellaneous" or "otherjflsector. .7 —..A . ...-6 This aggregation scheme tended to overestimate impacts relative to *‘t ""W ._._.,a ._-_.—... ._ alternative aggregation schemes, even in comparison to models with fewer sectors. The results regarding aggregation should be interpreted in light of available software and hardware and user ease of incorporating different practices relating to input data or IMPLAN model parameters. Although 173 model aggregation did not appear to affect Lake State impact results dramatically, greater model size generally does not pose computer difficulties or significant additional computing time requirements. This should be interpreted as lending support to the contention that it is best for impact analysts to work at the most disaggregated level possible. The availability of computer spreadsheet software significantly' minimizes the required time to aggregate results for further interpretation and presentation to other interested parties. The large range in Lake State FHW spending estimates suggests that recreation economic impact studies should more prominently address the precision implications of sample sizes and related qualifications in... w-., H regarding the variability of impact estimates. This author was told by FHW report staff that there are seldom inquiries about or discussion of the statistical parameters appearing in FHW reports. It was their impression. that the FHW sampling statistics are almost universally ignored. This is not unique to recreation studies; most economic impacts are presented as deterministic point estimates. The excuse for multipliers and other outputs from nonsurvey models is that sampling and nonsampling e ..- errors are unknown, in part due to sudh models employing a variety of __ __.,.m-—~- :- — H .2. _~ _ 1 data sources and reconciliation procedures in their construction. For example,' the! "corrected" MRIO data employed the RAS procedure to eliminate data errors and balance the original Jack Faucett data which, in turn, was'based on reconciling many sources of secondary economic data. It is difficult, if not impossible, to say how precise the corrected MRIO data are, although it can be said that its I-O accounts have been ‘balanced and appear consistent. Also, from a regional 174 economic perspective, trade estimates derived from the corrected MRIO data appear much more reasonable than those contained in IMPLAN. It is misleading to only present point estimates when information on sampling errors is available. Sampling errors are often known or may be derived for-recreation data that are used as final demand inputs for LG models. These may be used to derive ranges of impacts based on these sampling errors. In contrast, the precision of many I-O model parameters are likely to remain unknown; however, an examination of parameter values may reveal impact estimates (or RPCs, multipliers, etc.) are likely to fall within certain ranges. For example, this study developed. floor, ceiling, and. "best estimate" multipliers based on crosschecking estimated RPCs against RPC values for other regions and alternative secondary data sources. Presentation of recreation economic impacts in the form of ranges may be important not only to reflect what is known about the estimates' precision, but also to draw attention to the reliability of such estimates. Larger FHW samples or improved sample design appear necessary to improve the precision of recreation spending estimates for substate areas. The ease of examining and incorporating more detailed sector allocations and refined RPCs also implies a sizeable savings of time and energy can be achieved if this is done at the modeling system level rather than by users. Most IMPLAN users do not have a fraction of the time this author took to research the RPC issue or develop a detailed sector allocation bridge. Many IMPLAN users may not even have the knowledge or experience to evaluate the need to perform such tasks. Generic changes in RPC estimates should take place with the IMPLAN modeling system and materials supplied in conjunction with its use. This 175 would assure some improvement in the model's use which otherwise may not take place if left up to IMPLAN model users. further Emmi: M Several points have been made throughout this thesis regarding further research needs. These may be summarized in terms of the IMPLAN modeling system and the measurement of outdoor recreation economic impacts. The foremost conclusion of this research is that adequate evidence exists to indicate IMPLAN's RPCs should be re-estimated using the corrected MRIO database. This should be done with the econometric procedures utilized in the original IMPLAN RPC estimation process, although alternative independent variables could be considered. Re- estimation of RPCs would improve the reliability of IMPLAN's impact estimates for most regional models and sectors. Efforts to re-estimate IMPLAN RPCs with alternative independent variables alone will not enhance the quality of the RPC estimates as long as a faulty database (the original MRIO data) is used to generate initial RPC dependent variable values. Another major emphasis of this research has been to refine the allocation of reported FHW recreation spending to IMPLAN's sectors. The resulting detailed bridge tables represent an improvement over past bridge tables that were based on aggregated pce I-O categories. However, the tables prepared as part of this research are £13329ng Mb“ '- further review. Also, improvements could be made to their design to m "4 . _..__,.,—-—-—— enhance their use by a wider audience, especially for non-FHW applications and I-O models other than IMPLAN. 176 There is ‘1ittle ihope for standardizing outdoor recreation definitions and measurements. This is due to the diverse nature of outdoor recreation and its association with a broad array of resources and interests. These range from tourism and leisure studies through environmental and land or water use issues. Accurate, universally applicable spending profiles of recreationists are virtually precluded by these factors. What activities are counted in outdoor recreation studies will always be somewhat arbitrary, reflecting a study's particular objectives and data availablity. This does not mean that contrasts across studies can not be made. Contrasts of comprehensive efforts at recreation economic impact assessment may be especially helpful. Comparisons of existing studies across states and regions, such as Keiner's 1985 report, are invaluable in terms of indicating how methods and data may be improved. They enable progress in identifying better secondary data sources and the treatment of particular issues. They foster standardization in the proper identification and description of what is actually measured. For example, studies focusing only on recreationists' expenditures are often referred to as being economic development studies or "total" recreation economic impact studies. Comparisons of different studies will illustrate that studies of economic development or total impacts from recreation encompass cost factors and other issues besides merely recreationists' expenditures. Public costs for recreation deserve further research to help balance Ch; SEGAI presentation of recreation economic benefits.1 A major benefit and advancement in measuring outdoor recreation economic impacts could come from identification and examination of data sources 177 pertaining to public costs for recreation. A further extension of research on public costs would be to convert them for use with IMPLAN when the modeling system is fully converted to the SAM format. Concluaipn Three major findings stem from this analysis. The first is that IMPLAN's Version 2.0 RPCs are seriously flawed and can be significantly improved by being re-estimated using a corrected 1977 MRIO database. The second is that point (deterministic) estimates of outdoor recreation economic impacts are misleading because they ignore variability implied by sampling errors. A third finding is that estimates of recreationists' spending impacts may vary by a factor of several multiples, depending on what categoriespfiofw spending “age “included. Recreation impacts that include all equipment expenditures may be more than twice the size of impacts which only include trip expenditures. Advancements in computer speed and capacities along with improvements in I-O software design can be expected to continue. Impact analysis that previously took weeks, if not months, of work may now be accomplished in minutes. For example, once familiarity with IMPLAN is achieved (which only takes a matter of a few days training or construction of a few models), the time it takes from selecting a region through printing out sets of multipliers may be under an hour. (This author completed such a process for Kalamazoo County, Michigan in under twenty minutes, despite bugs in an early version of Micro IMPLAN being used at the time.) Fast turnaround in generating impact estimates provide the opportunity for abuse, partly because the more software is "user friendly" the less care and knowledge is required to use it. Also, the 178 time devoted to the task of preparing and reviewing the impacts may decrease commensurately with the faster generation of the impact estimates. Some savings in time may be applied to more crosschecking and sensitivity analysis if impact analysts are made aware of the efficacy of engaging in these tasks. Most IMPLAN users will probably continue to take their impact results and IMPLAN materials at face value without extensive examination. It is important for the modeling system to maximize its precision while minimizing potential user error in these circumstances. Prospects for improving the IMPLAN model system identified in this thesis merely reduce the potential problems from accepting the model as is and provide its users with a more solid basis upon which to further improve impact estimates. Additional improvements, let alone the proper use of the model and correct interpretation of the results, ultimately depend upon model users and the regional science profession. APPENDICES APPENDIX A: INDUSTRY CLASSIFICATION OF THE MICRO-IMPLAN 528 SECTOR INPUT/OUTPUT TABLES* APPENDIX R. Industry Classification of the nicro-IMPLAR 320 sector Input/Output tables no. Sector Rene SEA Commodity Standard Industry Classiiieation (SIC) 1 DAIRY FARM PRQUCTS ( 1.0100) 0241 Also : part of 0191, 0239, 0291 2 POULTRY AID EGGS ( 1.0200) 0231 0232 0233 Also : part of 0191, 0219, 0239, 0291 3 RAMCM 7E0 CATTLE ( 1.0311) part of 0191, 0212, 0219, 0239, 0291 4 RANGE FED CAIILE ( 1.0312) part of 0191, 0212, 0219, 0239, 0291 3 CATTLE EEEDLOTS ( 1.0313) 0211 Also : part of 0191, 0219, 0239, 0291 6 SMEEP, LAIIS A110 GOATS ( 1.0314) 0214 Also : part of 0191, 0219, 0239, 0291 7 11003, PIGS A110 SWINE ( 1.0315) 0213 - Also : part of 0191, 0219, 0239, 0291 0 OYHER MEAI ANIMAL PRODUCTS 4 1.0316) part of 0191, 0219, 0239, 0291 9 MISCELLAMENS LIVESTOCK ( 1.0302) 0271 0272 Also : part of 0191, 0219, 0239, 0279, 0291 10 001101 4 2.0100) 0131 Also : part of 0191, 0219, 0239, 0291 11 7000 GRAIRS 1 2.0201) 0111 0112 Also 1 part of 0191, 0219, 0239, 0291 12 PEG! GRAIRS ( 2.0221) 0113 Also 8 part of 0139, 0191, 0219, 0239, 0291 13 RAY AID PASTURE ( 2.0222) part a! 0139, 0191, 0219, 0239, 0291 14 GRASS SEEDS ( 2.0203) part of 0139, 0191, 0219, 0239, 0291 13 IOSACCO 4 2.0300) 0132 Also 1 part of 0191, 0219, 0239, 0291 16 IRUITS ( 2.0401) 0171 0172 0174 0173 Also : part of 0179, 0191, 0219, 0239, 0291 17 1REE RUIS ( 2.0402) part oi 0173, 0179, 0191, 0219, 0239, 0291 10 VEGETASLES ' ( 2.0301) 0134 0161 Also : part of 0119, 0139, 0191, 0219, 0239, 029 19 SUGAR CRVS ( 2.0302) 0133 Also : part of 0191, 0219, 0239, 0291 20 MISCELLAMEIRJS CROS ( 2.0303) part of 0119, 0139, 0191, 0219, 0239, 0291 21 OIL SEARIMG CROS ( 2.0600) 0116 Also 1 part of 0119, 0139, 0173, 0219, 0239, 029 22 FOREST PRODUCTS 4 2.0701) part of 0101, 0191, 0219, 0239, 0291 23 GREENWSE AND MSERY "Q0073. ( 2.0702) 0182 0109 Also : part of 0101, 0191, 0219, 0239, 0291 24 FORESTRY PRODUCTS ( 3.0001) 0010 0020 0040 0970 23 COMMERCIAL PISMIIG 1 3.0002) 0910 26 AGRICULTURAL, FORESTRY, FISHER! SERVICES ( 4.0001) 0710 0720 0730 0760 0234 0030 0920 Also : part of 0279 27 LANDSCAPE ARO RORIICULTURAL SERVICES ( 4.0002) 0700 20 IRON ORES ( 3.0100) 1011 29 FERROALLDY WES. EXCEPT VAMADIll ( 5.0200) 1061 30 COPPER ORES ( 6.0100) 1021 31 LEAD A110 ZINC MES ( 6.0201) 1031 32 GOLD ORES 4 6.0202) 1041 33 SILVER ORES ( 6.0203) 1044 34 EAUXITE A110 DTIIER ALIRIIM “ES ( 6.0204) 1031 33 METAL MIMIMG SERVICES ( 6.0203) 1001 36 MERCURY ORES 4 6.0306) 1092 37 URARIUM-RADIUM-VAVADIUM ORES ( 6.0207) 1094 30 METAL ORES, MOT ELSUMERE CLASSIFIED ( 6.0200) 1099 39 ARIMRACITE ARD ARIRRACITE MIRIMG SERVICES ( 7.0100) 1111 . , Also : part of 1112 40 SITUMIROUS AID LIGRIIE MIRIRG, SERVICES ( 7.0200) 1211 Also : part of 1213 41 NATURAL GAS ( 0.0101) Caution : 1310 is split between 0.0101 and 0.0102 42 CRUDE PETROLEUM 4 0.0102) 1310 ' Caution : 1310 is split between 0.0101 and 0.0102 43 MATLIAL GAS LIWIDS ( 8.0200) 1321 44 DIMENSION STWE ( 9.0100) 1411 43 CRUSHED AND SROKER LIMESTONE I 9.0201) 1422 46 CRUSMED AMD SRDREM GRANITE I 9.0202) 1423 47 CRUSMED A110 ERNEM 31012, I. E. C. ( 9.0203) 1429 40 CONSTRUCTION SAND AND GRAVEL 1 9.0301) 1442 49 INDUSTRIAL SAND ( 9.0302) 1446 M-1 179 APPENDIX N . 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 323 #333 8 333832333838383288283332 Sector Rees SENTNITE FIRE CLAY FULLER'S EARTN KAOLIN AID 8ALL CLAY CLAY, CERAMIC, REFRACTMY MINERALS, N.E.C. NMETALLIC MINERALS (ENCEPT FUELS) SERVICE GYPSLH TALC, SOAPSTNE, AND ”ATE MINERALS MISC. NMETALLIC MINERALS, N.E.C. SARITE FLWRSPAR POTASN, 00:10, AND ”ATE MINERALS PNOSPNATE MK ROCK SALT SULFUR CHEMICAL, FERTILI2ER MINERAL MINING, N.E.C. NEN RESIDENTIAL STRUCTURES NEU INDUSTRIAL AND COMMERCIAL SUILDINGS NEN UTILITY STRUCTURES NEN MIGNNAYS AND STREETS NEN FARM STRUCTURES NEN MINERAL EXTRACTION FACILITIES NEH GOVERNMENT FACILITIES MAINTENANCE AND REPAIR, RESIDENTIAL MAINTENANCE AND REPAIR DTNER FACILITIES MAINTENANCE AU REPAIR OIL AD GAS KLLS 00001010 001000 01::100: 0000011100, 000001 100 00011 0000, 0.0.0 1000: 000 1000 COMPONENTS. 00001 000: SMALL ARMS WITIN 01000 00000000 000 0000000010: MEAT PACEING PLANTS :00:000: 000 01000 00000000 0001: 0001101 000s:100 PLANTS 0001101 000 000 00000::100 00000001 001100 CHEESE, 0010001 000 00000::00 000000:00 000 Ev000R0100 0100 100 00000 000 1000011 0000001: 11010 0110 000000 000 00000 :00 1000: 000000 0000101110: 000000 10011: 000 VEGETASLES 0001000100 1000 0000001: 010010:, :0000:, 000 SALAD 000::100: 100:0 00 100200 00000000 11:0 100200 10011:, 10100: 000 VEGETASLES 100200 :000101110: 11000.000 01000 00010 MILL 0000001: 000001 00000001100: 0100000 000 00000000 10000 000, 001, 000 01000 001 1000 00000000 1000:, 0.0.0 RICE MILLING 001 0000 0100100 00000, 0000, 000 0010100 0000001: 000010: 000 0000000: :0000 0001001100001 0000001: 000001010 000 00000 0000001: CMEHINO GM MALT 010000: 113C) ( 9.0400) ( 9.0500) ( 9.0600) ( 9.0700) ( 9.0800) ( 9.0900) ( 9.1000) ( 9.1100) ( 9.1200) (10.0100) (10.0200) (10.0300) (10.0400) (10.0500) (10.0600) (10.0700) (11.0100) (11.0200) (11.0300) (11.0400) (11.0500) (11.0600) (11.0700) (12.0100) (12.0200) (12.0215) (13.0100) (13.0200) (13.0300) (13.0300) (13.0600) (13.0700) (14.0101) (14.0102) (14.0103) (14.0104) (14.0200) (14.0300) (14.0400) (14.0300) (14.0600) (14.0700) (14.08“) (14.0900) (14.1000) (14.1100) (14.1200) (14.1301) (14.1302) (14.1401) (14.1402) (14.1403) (14.1301) (14.1302) (14.1600) (14.1700) (14.1001) (14.1002) (14.1900) (14.2001) (14.2002) (14.2003) (14.2101) N-2 Inductry Classification of the Micro-IMPLAN 320 sector Input/Output tables, continued. SEA Coasodity Standard Industry Classification (SIC) 1452 1453 1454 1455 1459 1481 1492 1496 1499 1472 1473 1474 1475 1476 1477 1479 1521 1522 1530 In reality 0 at. of the corresponding SIC: 1341 1342 In reality pt. of the corresponding SIC: 1623 Also : part of 1629 1611 1622 part of 108, 1112, 1213, 136, 148 1629 r lity pt. of 1629 05 e I 0 , p) 0’1309 In reality pt. of 130 3795 2061 2062 2063 181 0000110111 N. Industry Classification of the Micro-IMPLAN 320 sector lupin/Output tables, continued. No. Sector Name SEA Comdity Standard Industry Classification (SIC) 113 MALT (14.2102) 2083 114 UINES. SRANOY, AND SRANDT SPIRITS (14.2103) 2084 113 DISTILLED LIGUOR, EXCEPT SRAIDY (14.2104) 2085 116 SOTTLED 0110 CANNED 5011 00101:: (14.2200) 2086 117 FLAVORING EXTRACTS AND SYRUPS, N.E.C (14.2300) 2007 110 COTTONSEED OIL MILLS (14.2400) 2074 119 SOTSEAN OIL MILLS (14.2500) 2075 120 VEGETADLE OIL MILLS, N.E.C (14.2600) 2076 121 ANIMAL AND MARINE FATS AND OILS (14.2700) 2077 122 ROASTED COFFEE (14.2800) 2095 123 SNORTENING AND MING OILS (14.2900) 2079 124 11001010010000 ICE (14.3000) 2097 125 MACARONI AND SPAGNETTI (14.3100) 2098 126 1000 PREPARATIOIS, 0.0.0 (14.32001 2099 127 CIGARETTES (15.0101) 2110 120 CIGARS (13.0102) 2120 129 CNEUING AND WING TUACCO (13.0103) 2130 130 TOSACCO STEIRIING AID REDRYING (13.0200) 2140 131 SROADI-OVEN FASRIC MILLS A10 FINISNING (16.0100) 2210 2220 2230 2261 2262 132 NARRGI FASRIC MILLS (16.0200) 2240 133 YARN MILLS AND FINISMING OF TEXTILES NEC (16.0300) 2269 2201 2282 2283 134 TNREAO MILLS (16.0400) 2284 133 11.000 COVERINGS (17.0100) 2270 136 FELT GCXXIS, N.E.C (17.0200) 2291 137 LACE 0000: (17.0300) 2292 130 PADDING AND UPNOLSTERY FILLING (17.0400) 2293 139 PROCESSED TEXTILE NASIE (17.0300) 2294 140 COATED FAERICS, NOT RMSERIEED (17.0600) 2295 141 1100 0000 000 100010 (17.0700) 2296 , 142 0000000 0110 1111110 (17.0900) 2290 143 NONIRJVEN FASRICS (17.1001) 2297 144 TEXTILE MS, N.E.C (17.1002) 2299 145 ”(ENS NOSIERY, EXCEPT SOCKS (10.0101) 2231 146 NOSIERY, N.E.C (18.0102) 2252 147 KNIT wTERlEAR MILLS (10.0201) 2233 148 RNIT UNDERHEAR MILLS (18.0202) 2254 149 EMITTING MILLS, N.E.C (10.0203) 2239 150 KNIT FASRIC MILLS (18.0300) 2257 2238 131 APPAREL MADE FROI MCNASEO MATERIALS (18.0400) 2310 2320 2330 2540 2350 2360 2370 2380 Also : part of 3999 152 CLRTAINS AID DRAPERIES (19.0100) 2391 153 NOISEFURNISNINGS, N.E.C (19.0200) 2392 154 TEXTILE SAGS (19.0301) 2393 133 CANVAS "mutt: (19.0302) 2394 136 PLEATING AND STITCNING (19.0303) 2395 137 AUTWTIW AND APPAREL TRIIIINGS (19.0304) 2396 138 SCNIFFI MACNINE EMSROIDERIES (19.0305) 2397 159 FASRICATED TEXTILE PRQUCTS, N.E.C (19.0306) 2399 160 LOGGING CAMPS AND LOGGING COITRACTGS (20.0100) 2410 161 SAIRIILLS AND PLANING MILLS, GENERAL (20.0200) 2421 162 110000000 DIMENSIOI AND FLGIING MILLS (20.0300) 2426 163 SPECIAL PRmUCT SAIRIILLS, N.E.C (20.0400) 2429 164 MILLUORK (20.0501) 2431 165 00:01 EITCMEN CASINETS (20.0502) 2434 166 VENEER AND 01.111000 (20.0600) 2433 2436 167 STRUCTURAL 00m RIDERS, N.E.C (20.0701) 2439 160 PREFASRICATED 1.10m SUILDINGS (20.0702) 2432 169 11000 PRESERVING (20.0000) 2491 170 0000 PALLETS AND SXIDS (20.0901) 2448 171 0001 I CLEDOARD (20.0902) 2492 172 0000 PRmUCTS, N.E.C (20.0903) 2499 173 0000 CONTAINERS (21.0000) 2441 2449 174 11000 NwSENOLD FURNITLRE (22.0101) 2311 173 NWSENDLD FURNITURE, N.E.C (22.0102) 2319 176 110m TV AND RADIO CASINETS (22.0103) 2317 177 UPNDLSTERED NOJSENOLD FURNITIRE (22.0200) 2312 170 METAL NwSEMOLD FURNITIME (22.0300) 2514 179 MATTRESSES AND SEDSPRINGS (22.0400) 2313 11-3 182 00000010 N. Industry Classification of the Micro-IMPLAN 328 sector Input/Output tables, continued. 00. Sector Nane SEA Col-0dity Standard Industry Classification (SIC) 100 0000 011100 100011000 (23.0100) 252‘ 101 00101 011100 100011000 (23.0200) 2522 102 000110 30110100 100011000 (23.0300) 2531 103 0000 0001111003 000 11010003 (23.04001 2541 104 110101 0001111003 000 111110003 (23.0500) 2542 105 011003, 300003, 000 0000001 00000000 (23.0600) 2591 106 100011000 000 111110003, N.E.C (23.0700) 2599 107 PULP 01113 (24.0100) 2610 100 00000 01113, 000001 00110100 00000 (24.02001 2620 109 0000000000 01113 (24.0300) 2630 190 0011010003 (24.0400) 2642 191 30011001 00000 PRODUCTS (24.05001 2647 192 30110100 00000 000 30000 01113 (24.06021 2660 193 00000 0001100 000 0102100 (24.07011 2641 194 0003. 000001 1001110 (24.0702) 2643 195 010-001 00000 000 00000 (24.07031 2645 196 PRESSED AND TOLDED PULP 00013 (24.0704) 2646 197 3101100001 00000013 (24.0705) 2640 190 000100100 00000 00000013, 0.0.0 (24.07061 2649 199 0000000000 0001010003 000 00003 (25.0000) 2650 200 0003000003 (26.0100) 2710 201 000110110013 (26.0200) 2720 202 0000 0001130100 (26.0301) 2731 203 0000 00101100 (26.0302) 2732 204 0130011000003 0031130100 (26.0400) 2740 205 0000000101 00101100 (26.0501) 2751 2752 2754 206 111000000010 01010000100 0N0 30001033 (26.0502) 2795 207 00011010 00310033 10003 (26.0601) 2700 200 0100000003 000 100301001 0100003 (26.0602) 2702 209 GREETING CARD MLISNING (20.07110) 2770 210 000000100 000 01010 00101100 (26.0001) 2753 211 00003100100 000 0010100 0000 (26.00021 2709 212 11003011100 (26.0003) 2791 213 00010000001100 (26.0004) 2793 214 ELECTROTTPING AND STEREOTYPING (26.0005) 2794 215 1000310101 100000010, 0000010 CNENIC013 (27.01001 2010 2065 2069 In reality only 00. of 2019 216 01100000003 000 0003000110 10011112003 (27.02011 2073 2074 217 10011112003, 010100 0011 (27.0202) 2075 210 000100110001 000010013, N.E.C (27.03001 2079 219 000 000 0000 000010013 (27.0401) 2061 220 000031103 000 30010013 (27.0402) 2091 221 0001031v03 (27.0403) 2092 222 00101100 10: (27.0404) 2093 223 000000 01000 (27.0405) 2095 224 00001001 000000011003, N.E.C (27.0406) 2099 225 01031103 001001013 000 003103 (20.0100) 2021 226 310100110 000000 (20.02001 2022 227 0011010310 000-0000 113003 (20.0300) 2023 220 0000010 110003, 00110011010310 (20.04001 2024 229 00003 (29.0100) 2030 230 SOAP AND DTNER DETERGENTS (29.0201) 2841 231 00113003 000 3001101100 00003 (29.0202) 2042 232 3001000 001110 000013 (29.02031 2043 233 101101 000000011003 (29.0300) 2044 234 PAINTS AND ALLIED PRIDLICTS (30.00001 2030 235 001001000 00110100 (31.0101) 2910 236 10001001100 0113 000 0000303 (31.01021 2992 237 001001000 000 0001 00000013, N.E.C. (31.0103) 2999 230 001100 01010003 000 310003 (31.0200) 2951 239 0300011 10113 000 00011003 (31.03001 2952 240 11003 000 10000 10003 (32.01001 3010 241 000000 000 01031103 10010000 (32.0200) 3020 242 000101000 003000 (32.0301) 3030 243 1000100100 003000 00000013, N.E.C (32.03021 3060 244 MISCELLANEOUS 01031103 00000013 (32.04001 3070 245 003300 000 01031103 0030 000 3011100 (32.0500) 3040 246 1001000 1000100 000 110130100 (33.0001) 3110 I-4 183 APPENDIX N. IndUStry Classification of the Nicro-INPLAN 320 Doctor Input/Output EDDIOS. continund. ND. Sector Nun DEA cmny Stamina IMUStI-y OISSSifiCStion (SIC) 267 POOTUEAR CUT STOCR (36.0100) 3130 260 SNOES, EXCEPT RUIIER (36.0201) 3163 3166 3169 269 NOUSE SLIPPERS (36.0202) 3162 230 LEATNER GLOVES AND NITTENS (36.0301) 3130 231 LUGGAGE (36.0302) 3160 232 NONENS NANDDAGS AND PURSES (36.0303) 3171 233 PERSONAL LEATNER GOODS (36.0306) 3172 236 LEATNER GOODS, N.E.C (36.0303) 3190 233 GLASS AND GLASS PRODUCTS. Exc CONTAINERS (33.0100) 3210 3229 3230 236 GLASS CONTAINERS (33.0200) 3221 237 CENENT, NYDRAULIC (36.0100) 3260 230 SRICR AND STRUCTURAL CLAY TILE (36.0200) 3231 239 CERANIC UALL AND FLOOR TILE (36.0300) 3233 260 CLAY REPRACTORIES (36.0600) 3233 261 STRUCTURAL CLAT PRODUCTS, N.E.C (36.0300) 3239 262 VITREOUS PLUNSING PIRTURES (36.0600) 3261 263 VITREOUS CNINA POOD UTENSILS (36.0701) 3262 266 PINE EARTNENUARE FOOD UTENSILS (36.0702) 3263 263 PORCELAIN ELECTRICAL SUPPLIES (36.0000) 3266 266 POTTERY PRODUCTS, N.E.C (36.0900) 3269 267 CONCRETE ILOCK AND IRIC! (36.1000) 3271 260 CONCRETE PRwUCTS. N.E.C (36.1100) 3272 269 READY-MIXED CONCRETE (36.1200) 3273 270 LINE (36.1300) 3276 271 GYPSUI PRODUCTS (36.1600) 3273 272 CUT STONE AND STONE PRODUCTS (36.1300) 3200 273 ASRASIVE PRODUCTS (36.1600) 3291 276 ASSESTOS PRODUCTS ' (36.1700) 3292 273 GASNETS, PACKING AND SEALING DEVICES (36.1000) 3293 276 NINERALS, GROUND OR TREATED (36.1900) 3293 277. NINERAL NODL (36.2000) 3296 270 NONCLAY REPRACTORIES (36.2100) 3297 279 NONNETALLIC NINERAL PRODUCTS, N.E.C (36.2200) 3299 200 ILAST FURNACES AND STEEL HILLS (37.0101) 3312 231 ELECTROHETALLURGIOAL PRODUCTS (37.0102) 3313 202 STEEL NIRE AND RELATED PRODUCTS (37.0103) 3313 203 COLD PINISNING OP STEEL SNAPES (37.0106) 3316 206 STEEL PIPE AND TUBES (37.0105) 3317 203 IRON AND STEEL POUNDRIES (37.0200) 3320 206 IRON AND STEEL PORDINOS (37.0300) 207 NETAL NEAT TREATING (37.0601) 3390 200 PRIMARY NETAL PRODUCTS, N.E.C (37.06021 3399 209 PRINART COPPER (30.0100) 3331 290 PRINART LEAD (30.0200) 3332 291 PRINART ZINC (30.0300) 3333 292 PRINARV ALUNINUN (30.0600) 3336 ALSO port of 2019 293 PRIMARY NONFERROUS METALS, N.E.C (33.0500) 3339 296 SECONDARV NONPERROUS NETALS (30.0600) 3360 293 COPPER ROLLING AND ORANING (30.0700) 3331 296 ALUNINUN ROLLING AND DRANING (30.0000) 3333 3336 3333 297 NONPERROUS ROLLING AND DRAUING, N.E.C (30.0900) 3336 290 NONPERROUS UIRE DRAUING AND INSULATING (30.1000) 3337 299 ALUNINUN CASTINGS ' (30.1100) 3361 300 SRASS. SRONZE, AND COPPER CASTINGS (30.1200) 3362 301 NONPERROUS CASTINGS. N.E.C. (30.1300) 3369 302 NONPERROUS PORGINGS (30.1600) 3663 303 NETAL CANS (39.0100) 3611 306 NETAL SARRELS. DRUNS AND PAILS (39.0200) 3612 303 NETAL SANITARY NARE (60.0100) 3631 306 PLUNDING FIXTURE TITTINGS AND TRIN (60.0200) 3632 307 NEATING EOUIPNENT, EXCEPT ELECTRIC (60.0300) 3633 300 PASRICATED STRUCTURAL NETAL (60.0600) 3661 309 NETAL DOORS, SASN, AND TRIN (60.0300) 3662 310 PADRICATED PLATE HORN (SOILER SNOPS) (60.0600) 3663 311 SNEET NETAL NDRR (60.0700) 3666 312 ARCNITECTURAL NETAL NOR! (60.0000) 3666 313 PREPADRICATED METAL DUILDINOS (60.0901) 3660 N-S APPENDIX N. 316 313 316 317 310 319 320 321 Sector None MISCELLANEGJS KTAL “X SCNEN MACHINE PN”UCTS A” NOLTS, ETC AUTOIOTIVE STAMPINGS CAMS AND CLDSLIES METAL STAHPINGS, N.E.C. CUTLENY HAND AND EDGE T”LS, N.E.C. HAND SANS AND SAN NLADES HANDUANE, N.E.C. PLATING AND POLISHING METAL COATING AND ALLIED SENVICES MISCELLANEWS FASNICATED NINE PN”UCTS STEEL SPNINGS, EXCEPT NINE PIPE, VALVES. AND PIPE FITTINGS METAL FDIL AND LEAF FASNICATED METAL PN”UCTS. N.E.C. STEAM ENGINES AND TUNSINES INTENNAL CMDUSTI” ENGINES. N.E.C. FANM MACHINENY AND EWIPMENT LAUN AND GANDEN EWIPMENT CWSTNUCTI” MACHINENT AND ENIPFENT MINING MACHINENY, EXCEPT DIL FIELD OIL FIELD MACHINENT ELEVATMS AND WING STAINUATS COIVETNS AND CNVETING EWINNT HOISTS. CNANES, AND NAILS I NDUSTN I AL TNUCXS AND _TNACT”S MACHINE T”LS, METAL CUTTING TYPES MACHINE T”LS. METAL FWING TYPES SPECIAL DIES AND T”LS A” ACCESSGIES PMN DNIVEN HAND T”LS NOLLING MILL MACHINENY METALIUXING MACNINENT, N.E.C. FM PN”UCTS MACMINENT TEXTILE MACHINENT ”MKING MACNINENT PAPEN INDUSTNIES MACNINENY PNINTING TNADES MACHINENY SPECIAL INDUSTNY MACHINENT, N.E.C. MP3 AND CUPNESSNS NALL AND NOLLEN SEANINGS NLGIENS AND FANS INDUSTNIAL PATTENNS PNEN TNANSMISSI” ENIPKNT INDUSTNIAL FLINACES A” ”ENS GENENAL INDUSTNIAL MACMINENT, N.E.C. CANSUNETNS, PISTOIS, NINGS, VALVES MACHINERY, EXCEPT ELECTNICAL. N.E.C. ELECTNONIC COQPUTING ENIMNT CALCULATING AND ACCGNTING MACMINES SCAL'ES AND SALANCES TVPEUNITENS AND DEFICE MACMINES. N.E.C. AUTOIATIC NENCHANDISING MACHINES CONTENCIAL LAUNDNT EWIPKNT NEFNIGENATIDN AND NEATING ENIP‘NT NEASUNING AND DISPENSING MS SENVICE INDUSTNT MACHINES, N.E.C. INSTRUMENTS T0 MEASLIE ELECTNICITY TNANSFMMENS SUITCNGEAN AND SUITCNDDAND APPANATUS MOTORS AND GENENATGS INDUSTNIAL COITNDLS NELDING APPANATUS. ELECTNIC CANS” AND GNAPNITE PN”UCTS ELECTRICAL INDUSTNIAL APPANATUS. N.E.C. HWSEHOLD CflING EDUIPMENT NWSEHOLD NEFNIGENATNS AND FNEEZENS HWSEHOLD LAUNDNT EWIPMENT l 184 SEA CMIEV SEAMAN Imvy CIuSHICITIm (SIC) (60.0902) (61.01”) (61.0201) (61.0202) (61.0203) (62.01”) (62.0201) (62.0202) (62.0300) (62.0601) (62.0602) (62.03”) (62.07”) (62.0000) (62.10”) (62.11”) (63.01”) (63.0200) (66.0”1) (66.0”2) (63.01”) (63.0200) (63.0300) (66.01”) (66.02”) (66.0300) (“.06”) (67.01”) (67.0200) (67.0300) (67.0601) (67.0602) (67.0603) (60.01”) (63.0200) (60.03”) (60.0600) (60.0300) (60.0000) (69.01”) (69.0200) (69.0300) (69.06”) (69.0300) (69.06”) (69.07”) (30.0”1) (30.0002) (31.0101) (31.0102) (31.0300) (31.06”) (32.01”) (32.0200) (52.0300) (32.0600) (32:0300) (53.0100) (33.0200) (33.0300) (33.0600) (33.0300) (33.0600) (33.0700) (33.0000) (36.0100) (36.0200) (36.03”) N-b 3669 3630 Industry CleSSificetion of the Nicro-INPLAN 320 SNCEDP‘IfiDUS/OUSOUT teDleS, continued. 3379 APPENDIX N. Sector None ELECTRIC NOUSEUARES AND FANS NOUSENOLD VACUUN CLEANERS SEUING NACNINES NOUSENOLD APPLIANCES, N.E.C. ELECTRIC LANPS LIGNTING FIXTURES AND EOUIPNENT UIRING DEVICES RADIO AND TV RECEIVING SETS PNONOGRAPN RECORDS AND TAPE TELEPNONE AND TELEGRAPN APPARATUS RADIO AND TV CONNUNICATION EDUIPNENT ELECTRON TUSES SENICONDUCTDRS AND RELATED DEVICES ELECTRONIC CONPONENTS. N.E.C. STORAGE SATTERIES PRIMARY SATTERIES. DRT AND NET X-RAY APPARATUS AND TUDES ENGINE ELECTRICAL EDUIPNENT ELECTRICAL EDUIPNENT, N.E.C. TRUCX AND SUS SODIES TRUCN TRAILERS NOTOR VENICLES MOTOR VENICLE PARTS AND ACCESSORIES AIRCRAIT AIRCRAPT AND NISSILE ENGINES AND PARTS AIRCRAPT AND NISSILE EGUIPNENT, N.E.C. SNIP SUILDING AND REPAIRING ' GOAT SUILDING AND REPAIRING RAILROAD EGUIPNENT MOTORCYCLES, SICTCLES. AND PARTS TRAVEL TRAILERS AND CANPERS NOSILE NONES NOTOR NONES TRANSPORTATION EDUIPNENT. N.E.C. ENGINEERING AND SCIENTIPIC INSTRUNENTS NECNANICAL NEASURING DEVICES AUTONATIC TENPERATURE CONTROLS SURGICAL AND NEDICAL INSTRUNENTS SURGICAL APPLIANCES AND SUPPLIES DENTAL EDUIPNENT AND SUPPLIES NATCNES. CLOCRS. AND PARTS OPTICAL INSTRUNENTS AND LENSES OPNTNALNIC GOODS PNOTOGRAPNIC EOUIPNENT AND SUPPLIES JENELRT, PRECIOUS NETAL JENELERS NATERIALS AND LAPIDART NORR SILVERNARE AND PLATED HARE CDSTUNE JENELENY NUSICAL INSTRUNENTS GANES. TOYS, AND CNILDRENS VENICLES DOLLS SPORTING AND ATNLETIC GOODS, N.E.C. PENS AND NECNANICAL PENCILS LEAD PENCILS AND ART GOODS NARRING DEVICES CARSON PAPER AND INNED RISSONS ARTIPICIAL TREES AND PLDNERS SUTTONS NEEDLES, PINS. AND PASTENERS SROONS AND SRUSNES NARD SURPACE FLOOR COVERINGS SURIAL CASEETS AND VAULTS SIGNS AND ADVERTISING DISPLAVS NANUPACTURING INDUSTRIES. N.E.C. NAILNDADS AND NELATED SENVICES LOCAL, INTENUNSAN PASSENGEN TNANSIT 11(15 DEA CONIDOIty (36.0600) (36.0500) (36.0600) (36.0700) (53.0100) (35.0200) (53.0300) (36.0100) (36.0200) (36.0300) (36.0600) (57.0100) (57.0200) (57.0300) (50.0100) (30.0200) (30.0300) (30.0600) (30.0300) (59.0100) (39.0200) (59.0301) (39.0302) (60.0100) (60.0200) (60.0600) (61.0100) (61.0200) (61.0300) (61.0500) (61.0601) (61.0602) (61.0603) (61.0700) (62.0100) (62.0200) (62.0300) (62.0600) (62.0500) (62.0600) (62.0700) (63.0100) (63.0200) (63.0300) (66.0101) (66.0102) (66.0106) (66.0105) (66.0200) (66.0301) (66.0600) (66.0301) (66.0502) (66.0303) (66.0506) (66.0600) (66.0701) (66.0702) (66.0000) (66.0900) (66.1000) (66.1100) (66.1200) (63.0100) (63.0200) .67 StenOer Industry Clessificstion (SIC) 3999 In reelity pt. of 3999 6010 AISD 6110 Indultry CIsssIIicstIon of the NIcro-INPLAN 320 sector Input/Output tsDIes, contInueO. 36663667 3660 3666 3672 3673 3676 3677 3670 3679 3026 3029 6060 6760 our! of 6709 6120 6130 6160 6130 6170 186 APPENDIX N. Industry Clsssificstion of the NIcro-IIOLAN 320 sector Imam/Output teDles. continued. N0. Sector NsIRe SEA Candi" Stendsrd Industry ClessiIIcetion (SIC) 660 IDTU PREIGNT TRANSPGT AID NARENGJSING (65.0300) 6210 6220 6230 Also port DI 6709 669 HATER TRANSMTATIOI (65.06”) 6610 6620 6630 6660 6630 6660 650 AIR TRANSPOTATIN (65.0500) 6510 6320 6300 631 PIPE LINES, EXCEPT NATINAL GAS (65.0600) 6610 632 TRANSPORTATIOI SERVICES (65.0701) 6710 6723 6700 In reelIty pt. 07 6700 633 ARRANGENENT OP PASSENGER TRANSPORTATION (63.0702) 6722 636 CONNJNICATIOIS, EXCEPT RADIO AD TV (66.0000) 6010 6020 6090 655 RADIO AND TV SROADCASTING (67.00”) 6030 656 ELECTRIC SERVICES (60.01”) 6910 A106 port of 693 657 GAS "NT!” AU DISTNIIUTI” (60.02”) 6920 Also port of 693 630 NATER ”9LT AND SERRAGE STSTENS (60.0301) 6960 6952 639 SANITARY SERVICES AD STEAM ”PLT (60.0302) 6933 6959 6960 6970 Also port of 693 660 RECREATIONAL RELATED MESALE TRADE (69.0101) 3061 5063 661 DTNER NNOLESALE TRADE (69.0102) 3010 3020 3030 3062 3030 3060 3070 3ND 3090 3100 662 RECREATIOIAL RELATED RETAIL TRADE (69.0201) 5331 3361 3961 3966 5967 3960 663 DTNER RETAIL TRADE (69.0202) 3200 3300 3600 3310 3320 3330 3360 3370 3390 3600 3700 3910 3920 3930 3962 3963 3966 3963 3969 3960 3900 3990 7396 0062 666 RANKING (70.0100) 60” 663 CREDIT AGENCIES (70.0200) 6100 6710 6720 6733 6790 666 SECURITY AND CW)" MRS (70.0300) 62” 667 INSURANCE CARRIERS __ (70.0600) 6300 660 INSURANCE AGENTS AID MRS (70.0300) 6600 669 (NAIERoOCCUPIED DIELLINGS (71.0100) , 670 REAL ESTATE (71.0200) 6300 6600 Also pt. of 1331 3 Excludln DR. of 6332. 671 HOTELS AND LQGING PLACES (72.0100) 7000 672 LAUNDRY, CLEANING AND SNDE REPAIR (72.0201) 7210 7230 673 PUNERAL SERVICE AID CRENATmIES (72.0202) 7260 676 MTRAIT AND PHOTWRAPNIC SMIOS (72.0203) 7220 7290 675 ELECTRICAL REPAIR SERVICES (72.0206) 7620 676 NATCN,CLDCK..IE1£LRT AIR) ELINITIIE REPAIR (72.0203) 7630 7660 677 SEAUTT AND SARSER Sm (72.0300) 7230 7260 670 MISCELLANEGIS REPAIR 3100 (73.0101) 7690 679 SERVICES TO SUILDINGS (73.0102) 7360 600 PERSOINEL SUPPLY SERVICES (73.0103) 7360 601 COIPUTER AND DATA PROCESSING SERVICES (73.0106) 7370 602 MANAGEMENT AND CONSULTING SERVICES (73.0103) 7391 7392 7397 603 DETECTIVE AND PROTECTIVE SERVICES (73.0106) 7393 606 EOJIPNENT REPAIR AND LEASING (73.0107) 7396 603 PHOTOFINISHING, DONNRCIAL PNOTDGRAPNT (73.0100) 7332 7333 7393 606 DTNER SUSINESS SERVICES (73.0109) 7320 7331 7339 7330 7399 607 ADVERTISING (73.0200) 7310 600 LEGAL SERVICES (73.0301) 0110 609 ENGINEERING, ARCNITECTIIAL SERVICES (73.0302) 0910 690 ACCGINTING, ALOITING An ”KEEPING (73.0303) 0930 0990 691 EATING AND DRINKING PLACES (76.0000) 3000 Also port of 70 692 AUTOROSILE RENTAL AND LEASING (73.0001) 7310 693 AUTOIOSILE REPAIR Aw SERVICES (73.0002) 7530 7369 696 AUTOIOSILE PARKING AND CAR NASN (73.0003) 7320 7362 693 NOTION PICTURES (76.0100) 7000 696 DANCE NALLS, STLDIOS AND SCNODLS (76.0200) 7910 697 TNEATRICAL PRmUCERS. SANDS ETC. (76.0201) 7920 690 SONLING ALLETS AND POOL NALLS (76.0202) 7930 699 COHERCIAL SPORTS ERCEPT RACING (76.0203) 7961 300 RACING AND TRACK OPERATION (76.0206) .7960 301 NENSERSNIP SMTS AND RECREATIOI CLINS (76.0205) 7997 302 AINISENENT AND RECREATION SERVICES. NEC (76.0207) 7992 7993 7996 7999 303 DOCTORS AND DENTISTS (77.0100) 0010 0020 0030 0061 306 NDSPITALS (77.0200) 0060 305 NURSING AID PROTECTIVE CARE (77.0301) 0030 306 DTNER NEDICAL AND NEALTN SERVICES (77.0302) 0760 0069 0070 0000 0090 APPENDIX N. 187 InduStry ClessiIicstion 01 the Nicr0~INPLAN 320 seetor Input/Output tlees, continued. N0. Sector None SEA CONIodity Stsndsrd Industry Clessificstion (SIC) 307 ELENENTARY AND SECONDARY SCNOOLS (77.0601) 0210 300 COLLEGES. UNIVERSITIES. SCNOOLS (77.0602) 0220 309 DTNER EDUCATIONAL SERVICES (77.0603) 0230 0260 0290 310 SUSINESS ASSOCIATIONS (77.0301) 0610 0620 311 LASOR AND CIVIC ORGANIZATIONS (77.0302) 0630 0660 312 RELIGIOUS ORGANIZATIONS (77.0303) 0660 313 OTHER NONPROPIT ORGANIZATIONS (77.0306) 0600 0630 0690 6732 0922 516 RESIDENTIAL CARE (77.0000) 0361 313 SOCIAL SERVICES, N.E.C. (77.0900) 0321 S399 0331 0331 316 U.S. POSTAL SERVICE (70.0100) 6311 317 TEDERAL ELECTRIC UTILITIES (70.0200) port 01 691 310 DTNER PEDERAL GOVERNNENT ENTERPRISES (70.0600) -- 319 LOCAL GOVERNMENT PASSENGER TRANSIT (79.0100) DIrS 61 61 320 STATE AND LOCAL ELECTRIC UTILITIES (79.0200) port 61 691 321 DTNER STATE AND LOCAL GOVT ENTERPRISES (79.0300) -- 322 NONCONPARASLE INPORTS (00.0000) -- 323 SCRAP (01.0001) -° 326 USED AND SECONDNAND GOODS (01.0002) ~- 323 GOVERNNENT INDUSTRY (02.0000) -- 526 REST 0! INS NORLD INDUSTRY (00.0000) -- 327 NOUSENOLD INDUSTRY (06.0000) 0000 320 INVENTORY VALUATION ADJUSTMENT (03.0000) .. *Source: Alward, Gregory S. et a1. (1989). Micro IMPLAN Software Manual. Appendix N. Industry Classification of the Micro- IMPLAN 528 sector Input/Output tables. APPENDIX B: COMPARISON OF IMPLAN, REMI, AND CORRECTED MRIO RPCS (Sectors in omital letters are Wed) MICHIGAN IRIO m N DANE, mun IIPLAN h "PLAN NBII IRIO 1 DAIRY EARN mm, 1 0mm 0.9609 2 Livestock, 2-9 0.668 0.5168 3 Oottai,Grain, Td)DOOO,10-15 0.6596 0.2% 6 Misc ers, 158,5 0.5716 0.0109 * 5 Forestry Prdcts, 26 0.22 0.97 0.3665 * DOOIIIercial Fishim, 5 0.893 0.5 0.1975 ** 7 Ira) (res Hinim, 5-29 0.01% 0.602 8 NcnferrOLs (Res, 30-38 0.6262 0.765 9 GIL HINIIN), 5-60 0.‘ 0 0.0000 10 (ROE P61101801, 61 0.0629 0&2 11 Neural Gas, 62-63 0.5666 0.155 12 Stan, Clay, Hinim, 66-58 0.295 0.302 13 Chanicol Minerals, 59-65 0.0019 0.0000 16 Oastnctim, 66-76 0.9769 0.87 1.01100 20 am, 77-81 0 0.257 0.9510 21 Meat Prodnts, 82-5 0.6516 0.3913 0.6%7 22 Dairy Prdcts, 59) 0.296 0.6699 0.756 8 Cured 8. Fm PM, 91-% 0.1636 0.3696 0.1166 26 Grain Hill Pmdsts, 99-15 0.55 0.3577 0.625 25 Bakery Prodsts, 1(5-107 0.282 0.6503 0.2% 8 am & Cmfectimety, 1m-111 0.6566 0.2603 0.5% 27 W, 112-117 0.6967 0.558 0.539) 28 Other Food Prodsts, 118-18 0.6326 0.31% 0.2716 29 TdaaOOO Prodsts, 127-130 0.0007 0.1150 0.01!) 30 eric, Yarn, Thread, 131-136 mm 0.050 0.188 31 Floor Coverirm, 135-166 0.1228 0.1165 0.1711 32 Hosiery 8. Knit GOOCB, 165-6,150 0425 0.029 0.0!!) Expand, 167-8,151 0.0679 04056 0.1106 36 Other Fa). Prdcts, 169,152-9 0.659 0.6112 0.7386 5 Lomira 0 Lurber, 160-163 0.6191 0.568 0.2106 36% Prodnts, 166-7,9-173 0.5962 0.657 0.301) 37 Prefd) BIW Ham,168,613 0.- 0.573 0.536 38 Haadnld Funitu'e, 176-179 0.39.36 0.1938 0.866 5 Other Fu'nittre, 181-15 0.- 0.3%9 0.6677 60 aner Prodnts, 187-1% 0.(IXI9 0.806 0.519 61 Pm (DITAIIERS, 199 0.756 0.2926 0.5612 62 Nam, Other Printiru, 21)-216 0.3760 0.5751 0.5960 63 IIuSTRIAL OBIICALS, 215 0.5131 0.6192 0.3631 66 Aa‘icultu-al Chenicals, 216-8 0.1131 0.- 0.3115 65 Other Chanicels, 219-226 0.7% 0.222 0.526 66 Plstics E Wits, 25-28 0.655 0.527 0.6836 67 m, 229 0.1277 0.2632 0.5179 68 Counties, Clemim Prtts,80-. 0.5567 0.1531 0.1691 69 PAINTS AID ALLIE) PROM), 86 0.502 0.352 0.287 50 Petrolwn Refinirn, 85.89 0.2160 0.1761 0.118? ** S1 Rdber 8. Misc. Platics, 260-265 0.0166 0.355 0.817 52 Leather 8. Leather Putts, 266-56 0.365 0.1069 0.212 188 MIDI 01F“ LAKE IIPLAN IPP-PRIO NIS'DISIN MIWA STATES - BRIO III? (N) RIO RIO RIO 0.0151 2): 0.9965 0.55 0.9912 -0.0766 ~17X 0.5619 0.9165 0.7106 -0.3112 *3 05$ 0.6712 0.556 -0.Z376 462% 0.3210 0.91% 0.7013 0.357 55 0.7676 0.3165 0.6719 °0.1501 402); 0.0390 0.123 0.1615 -0.67% 75633 0.0000 0.262 0.56” -0.318 45% 0.6126 mm 0.5656 0.-1W 0.0000 0.0000 0.0000 -0.21m 415% 0.035 0.0000 0.1207 0.61m 77% 0.”62 0.0000 0.0770 0.3623 67% 0.535 0.555 0.6575 0.0020 1“! 0.(XXXJ 0.1!!!) 0.0000 0.0251 -5 1.0000 1.111!) 1.0000 -0.510 0.0376 0.230 0.6606 0.155 23x 0.359 0.623 0.65% -0.(X50 -1X 0.7970 0.7130 0.7550 0.0673 29% 0.5050 0.2707 0.2751 -0.2”1 -02 0.3003 0.352 0.67” -0.(I)15 4! 0.7066 0.5696 0.7%!) 0.156 61% 0.205 0.1715 0.502 -0.065 -9X 0.5310 0.7693 0.5761 0.16” 37% 0.153 0.5666 0.5” 0.018 1“ 0.0000 DJIXD D.CXX!) -0.w 46062 0.135 0.1210 0.115 4.0683 ~56 0.529 0.0360 0.516 0.0205 1W 0.7130 0.536 0.560 0.526 4!! 0.1070 0.1670 0.1616 -0.1116 45 0.35 0.635 0.625 0.ZB7 5C! 0.573 0.165 0.215 0.2162 3“ 0.655 0.659 0.6277 -0.26& 41616“ 0.597 0.653 0.56” 0.1591 6“ 0.6111 0.315 0.3660 0.3766 66% 0.6” 0.265 0.650 -0.510 595% 0.6362 0.352 0.3630 0.36 29X 0.7126 0.0707 0.67” -0.2199 69% 0.565 0.5610 0.565 0.17” 3 0.2260 0.0677 0.517 10.1”. 475% 0.1657 0.1226 0.1092 0.751 g 0.06” 0.0996 0.0652 0.1129 1% 0.507 0.506 0.6065 0.3901 45% 0.159 0.15” 0.3700 0.30" 702 0.6315 0.357 0.2710 -0.6606 -1m 0.5616 0.2667 0.615 0.52 65% 0.575 0.635 0.513 -0.2271 41% 0.569 0.506 0.26% 0.1367 5% 0.5606 0.506 0.351 Appendix B (cont'd.) 189 IRIO no. Nate, "PLAN sector tubers MI "PLAN RBII 56l3880las Prchts, 55, 56 “5681:038Claym8,57-279 0.1997 0.55 ** 55 Inn 8 Steel Forgiru,m-1,3,6,6-0 0.135 ** 56 11101 AM) 5151. WIS, E 57 Prim Naiferrols Metal,25-3(2 58 Metal m1m,&,3m'6,319729 59 Stnctu-al Metal P0565546 60 Some Ibohine Prdcts, 315-318 61 Ercims 8 Tlrbires, 330,331 62 Farm 8 Lam Equlent, 52,333 65 Castnctiai, Miniru Eqfip.,336-6 66 interials l-Iadlim Eqaip.,337-360 65 Metalworkiru Eminent, 361-366 66 Special Madiirety, 367-52 0.565 0.1011 0.%1 0.153 0.567 0.756 0.052 0.0792 0.6267 0.55 0.5716 *- 67 Other mtect. Equipmc, 353-361 0.0001 68 Office a. Oalpniru Equp., 362-365 0.1962 69 Sen/ice inchinery, 366-370 70 Electrical Eminent, 371-378 71 Hasehold Amlimces, 379-35 2 Electric Limtiru, 35-388 73 Reoeivim Sets, Reconcb, 389,393 76 Gummicatia's Eminent, 391-392 75 Electronic Camts, 53-395 76 Other Electrical Emip.,396~6m 77 Motor Vehicles 8 Parts, 601-606 78 AIRCRAFT, 1.05 79 Missiles 8 Parts, 76, 607 5 AIRCRAFT AID MISSILE EN, 65 83 Medical Eminent, 619-21,68-6 86 Other w. Prcbts, 65-665 as mums ND mama) SE, 666 5 Local Tramit, 667, 519 07 1mm FREIGHI’ MT, 668 5 MTER WTATIOI, 669 5 AIR WTATIOI, 650 5 PIPE LIES, DWI MAT GAS, 651 91 Tramportatim Services, 652-3 92 WIMTIOIS, EXEPI PIDIO, 656 93 "DID AID TV W116, 655 96 Electric Utilities,656,517,55 0.6610 0.3162 0.250 0.0109 0.1167 0.3126 0.55 0.515 0.6375 0 0.3010 0.6581 0.556 0.615 0.1366 0.530 0.365 0.5712 0.5671 0.650 0.55 0.430 0.5219 0.357 0.5192 0.1706 0.6978 0.1566 0.1710 0.0507 0.135 0.0655 0.577 0.2730 0.672 0.01 0 0.2139. 0... 0.35707 31 Other tramp. mama-12.1.1445 0.1895 0.31751 82 ScientJPhoto Emip,616-10,622,65 0.566 0.07801 0.3765 0.07757 0.560 0.19368 0.6660 0.555 0.511 0.6019 0.6227 0.655 0.25 0.575 0.66 0.215 0.215 0.35 0.6630 0.9653 0.6292 0.695 0.550 5 GAS P80151101 8 DISTRIBJTIOI, 657 0.5960 96 later quly 8 Smitaty m, 658-55156 97 lbolesale Trade, 660,661 5 EATIIB AID DRIIKIIII PLACES, 691 99-102 Other Retail Trash, 662-663 0.065 0.6% 0.5% 15 Mira, Crdit Wis, 666-666 0.552 106 Immoe, 667-668 105 Real Estate, 669-670 0.659 0.6% 0.55 0.5 0.55 0.55 0.759 0.” 0.7575 0.0655 0.573 0.515 0.666 10110 IMIO 01mm 171 RIO 0.510 0.7156 0.325 0.655 0.215 0.6750 0.550 0.730 0.6792 0.515 0.155 0.6131 0.665 0.1667 0.312 0.175 0.3010 0.350 0.6966 0.1- 0.0653 0.1191 0.536 0.250 0.6631 0.5736 0.566 0.” 0.329 0.256 0.3507 0.1073 0.5621 0.6673 1.1!!!) 0.7073 0.569 0.275 0.7M 0.5601 0.6132 0.553 0.965 1.011) 0.6276 0.0901 0.569 0.5512 0.575 0.695 '0.0313 45 -0.568 4115 -0.1090 465 -0.66% 435% 0.55 -21% -0.0797 -20x -0.m -159% -0.162 -& 0.226 32 0.650 2% -0.562 -m 0.0116 5 -0.m65 -1% 0.659 75% -o.3170-2asoax 0.5 15 0.593 55 0.0106 3% 0.2% -161% -0. 1716 -1566% 0.55 61% 0.195 62% '0.0136 -5% '0.GB6 -5 -0.055 -6% -0.5736 -0.566 0.267 90% -0.156 -97% -o.2zo1 -51% 0.60 7% -0.065 -65 -0.1101 -27% 0.- 15% -0.6® 456% -0.356 '51 0.1659 5% 0.153 65 -0.570 65 0.591 11% 0.07% 16% -0.652 -05 -0.355 -59% -0.3063 -62% -0.556 -153% -0.zaao -65 -0.6667 -56 0.0660 7% -0.0793 -16% 0.532 -16% 0.1378 0.755 0.2166 0.65 0.2613 0.651 0.6296 0.653 0.6166 0.665 0.6765 0.5917 0.565 0.557 0.513 0.2662 0.55 0.1279 0.515 0.151 0.195 0.157 0.355 0.65% 0.531 0.0000 0.0000 0.5166 0.1212 0.7066 0.593 0.075 0.0652 0.759 1.0000 1.2121 0.150 0.0000 0.593 0.6797 0.635 0.759 0.539 1.0000 0.751 0.9000 0.950 0.5701 0.757 0.7% IN RIO LKS'I’ "110 0.2697 0.765 0.115 0.656 0.150 0.651 0.5565 0.5360 0.55 0.6016 0.557 0.665 0.051 0.3652 0.615 0.315 0.1766 0.165 0.6666 0.0952 0.567 0.510 0.515 0.1759 0.051 0.3606 0.2273 0.- 0.6715 0.6522 0.6971 0.3576 1.0009 0.7999 0.563 1.2760 0.6799 1.11111 0.7111) 0.695 0.635 0.757 0.9765 1.0000 0.9955 0.9000 0.965 0.6665 0.7116 0.7!!!) 0.337 0.226 0.257 0.6976 0.2191 0.655 0.5516 0.7316 0.657 0.351 0.615 0.6151 0.656 0.2937 0.3626 0.2619 0.2766 0.2202 0.556 0.1717 0.07% 0.2097 0.315 0.537 0.6059 0.3766 0.0600 0.115 0.356 0.6591 0.3165 0.1556 0.7168 0.6501 0.9753 1.513 0.315 0.657 0.6999 0.557 0.656 0.555 0.55 1.“!!! 0.751 0.550 0.965 0.551 0.635 0.859 Appendix B (cont'd.) "110 m. 1% 1018.5 ND Lamas PUG, 471 107 Pascal, anir m, 4TZ-fl ** 103 "fee. was 8. Mariska, 478-487 ** 109 Misc. Profsiaul Svcs, 4884.5 110 Auto natal, anir, 1.92-496 111 W, 495-5Q 112 was NO oamsrs, 503 113 Hospitals, 504, 5(5 114 OTIER DEICN. ND lEALTH, 5% 115 Edntiaul Set-vim, $07-50? 116 Na'pmfit Mimics, 510-513 11706:» Social Sewing, 514, 515 118 Fed Govt. Enterprises,516,518 119OTIERSTATEND MW, 521 120 mm "HITS, 52 121 saw & SEEDDHND (mas, 523-4 12 m nousm, 55 18 111158110 nousm, 527 ueidnted was for RPCs: mid-net! W for RPCs: 0.651 0.6111 0 0 0.554 0.7021 1.0000 0.540 0.Wo5 0.560 1.0000 0.535 0.9488 0.0002 0.55 0.9503 0.9999 0.4378 0.456 m, "PLAN aectorrulhers HI 1M R911 0.118 0.530 0.- 0.516 0.501 0.6113 0.599 0.55 0.9747 0.4% 0.8459 0.512 0.5322 0.4219 190 R10 [lb-RIO DIFF/IH’X V! ”210 m “110 LKST H110 0.782 0.9000 0.7316 mam 0.87% 0.8366 0.9500 0.9000 0.9000 0.m4 0.7990 1 .(IID 0.5612 1 .0000 1 .0000 0.22 1 .0000 1 .0000 0.572 0.4930 -o.09m 46% am -472 -0.T.516 -o.ao14 -0.6Z$0 -244X -0.‘l345 49% 0.0500 5% 0.541 9% 0.” 1“ mm Z“ 0.157! W 0.0000 (R 0.4023 42% -0.511 '5X -0.55‘4187W 0.6477 67% -0.0197 ~21 -o.ooao 43 -0.1M -32x -0.%74 46% 0.N54 0.‘ 0.6654 0.645 0.$5 0.864 0.%2 0.“!!! 0.- 0.751 0.7365 1.0000 0.7279 1.0000 1.“!!! 0.3111 1.0000 1.0000 0.642 0.454 * simifies inperfect RBI! cmparism (R941 R903 are mt available for sectors left blait) ** simifies sectors for lhid’I ”PLAN's Hichim RPCs are zeno or mar zero despite siwe 9? ratios 0.7312 0.- 0.7% 0.7670 0.9000 0.8487 0.9500 mm 0.6” 0.7910 0.751 1.0000 0.7501 1.0000 1.0000 0.3105 1.0000 1.0000 0.6460 0.5030 0.735 0.8761 0.791 0.7551 0.887! 0.895 0.537 0.5m 0.7761 0.7800 0.7818 1.0000 0.6475 1.0000 1.0000 0.254 1.01” 1.0000 0.5971 0.552 APPENDIX C: ALTERNATIVE LAKE STATE RPCS AND RPCS USED AS GUIDES FOR THE ESTIMATION OF THE LAKE STATE RPCS noun MICHIGAN arcs mm "PLAN 5016 m 00!!) 153 mm U15 LIST E: K 2 M5 __ALTW E! MA! 1121.! Ell & & fl 1 Daily Fun Putts 1 0.540 0.540 0.9994 0.9996 0.540 mm 0.5 0.512 2 Palm 8 E“ 0.8132 0.4750 0.4750 0.812 0.812 0.4750 0.545 0.501 0.7184 3 Rad! Fa! Cattle 1 0.455 0.455 0.9300 0.9!!) 0.455 0.55 0.527 0.7184 4 W Fd Cattle 1 0.456 0.4% 0.9% 0.929 0.4% 0.55 0.55 0.715 5 Cattle Fedlots 0.49.5‘ 0.2748 0.2748 0.4935 0.495 0.3“ 0.753 0.2748 0.7184 6 ”Ants, I'd Guts 0.5476 0.3113 0.353 0.5476 0.5476 0.3% 0.‘ 0.1115 0.715 7 11033166 Mr: 1 0.5147 0.5147 0.569 0.569 0.5147 0.5 0.6707 0.7184 8 Other lat Millals 0.541 0.571 0.571 0.541 0.9.1 0.571 mm 0.537 0.7184 9Hisc Livestock 0.95 0.454 0.454 0.35 0.65 0.454 0.6011 0.” 0.7184 10 fatten 0 0 0 0 0 0 0 0 0.556 11 Fond Grails 0.5161 0.1% 0.1% 0.5161 0.5161 0.2491 0.546 0.1% 0.556 12 Fed Grain 1 0.556 0.556 0.7220 0.7m 0.579 0.7170 0.554 0.556 13 My 8 Pam 0.9113 0.5184 0.5184 0.7661 0.751 0.5184 0.657 0.7414 0.556 14 Grss Seat 0.5% 0.0742 0.0742 mm mm 0.0967 0.5013 0.0742 0.556 15 Tdmco 1 0 0 0.245 0.245 0 0 0.322 0.556 16 Fruits 0.4158 0.045 0.045 0.4158 0.4158 0.5418 0.045 0.3758 0.7813 17 Tm m 0 0 0 0 0 0 0 0 0.7813 18 Was 0.9391 0.6952 0.652 0.753 0.5 0.6977 0.652 0.7429 0.7813 19 an ers 1 0.55 0.0215 0.915 0.915 0.8131 0.9775 0.‘ 0.7813 5 Misc Cm 0.5937 0.3218 0.3218 0.5937 0.5937 0.4976 0.8130 0.218 0.7813 21 Oil Beoriru ers 1 0.515 0.515 0.8487 0.8487 0.552 0.514 0.515 0.7813 2 Forest Putts 0 0 0 0 0 0 0 0 0.7813 5 W 8 “say Pr 0.765 0.4446 0.4446 0.657 0.657 0.755 0.4446 0.555 0.7813 24 Forestry Putts 0.7552 0.550 0.550 0.7552 0.7559 0.7272 0.5 0.550 0.7772 0.4719 5Callllercial Fishiru 0.- 0.M7 04237 0.50.50.17.35 0.5 0.594 01257 0.1415 5‘95, For, 8 Fish SVcs 0.674 0.274 0.274 0.340? 0.345 0.274 0.2 0.452 0.456 0.7813 27 W8 Hortic SVcs 0.7546 0.524 0.524 0.3163 0.3163 0.3413 0.72 0.3% 0.524 0.7813 281rm0res 1 0.618 0.545 0.6111) 0.618 0.11117 0.57 0.615 0.11131 0.545 29 Fel‘rcalloymes 0.3622 0 0.1a!) 0.2000 0.11101 0.0mm 0.57 0 0 0.545 30 metres 1 0.0975 0.675 0.517 0.517 0.55 0.57 0.9493 0.675 0.5456 31 Lead821rc0nas 0.“ 0 0 0.0265 0.“ 0.1258 0.57 0 0 0.5456 2 fold” 0.613 0 0 0.613 0.613 0.621 0.57 0 0 0.5456 33 Silva-(mes 04m 0 0 0.- 0.- 0.0448 0.57 0 0 0.5456 34 Baxiteluomerklun 0 0 0 0 0 0 0.57 0 0 0.5456 5 Metal Hinim m 1 0.5456 0.5456 0.950 0.950 0.9956 0.57 0.9974 0.9956 0.5456 36mm 0 0 0 0 0 0 0.57 0 0 0.5456 37mm1umkdiunyamdiun 0.1115 0.005 0.015 0.0005 0.00:5 0.0005 0.57 0 0.63 0.5456 381121310113, use 0 0 0 0 0 0 0.57 0 0 0.5456 Enthrscite Hiniru 0 0 0 0 0 0 0.01 0 0 0 40 Bitunimusl. Limits 0.616 0 0 0.0218 0.618 0.0327 0.01 0.63 0.1!!)4 0 41 ”Petrol” 0.125 0 0 0.53 0.55 0.0430 0.2 0 0 0.0778 42 mat as 0.351 0 0 0.256 0.531 0.5641 0.2 0 0 0.157 43 flaunt Gs Liqai¢ 0.2378 0.663 0.663 0.2378 0.2578 0.464 0.2 0.665 0.1076 0.0778 44 Dimim Stan 0.416 0.266 0.66 0.416 0.416 0.66 0.65 0.367 0.784 0.4575 45 m 8 m Limt 1 0.42” 0.42” 0.55 0.55 0.574 0.65 0.459 0.8162 0.4575 46 Cndid 8 m Grmit 0.541 0.015 0.015 0.3541 0.541 0.015 0.65 0.8155 0.245 0.4575 47 m 8 m Stan, 0.75 0.4575 0.4575 0.7m 0.7m 0.4810 0.5 0.6m 0.584 0.4575 48 Martin“! Std 8 Gm 1 0.4575 0.4575 0.9369 0.9369 0.514 0.5 0.4“ 0.759 0.4575 191 Appendix C (cont'd.) $010. sacrum 49 Irdstrial Sad 50 Batmite 0.153 51 Fine Clay 52 Fuller's Earth 1 0 5 Kaolin 8 Ball Clay 0.613 54 Clay,0eranic,8Refracta‘ 1 55 Namllic Himls 9v 0.851 56 0mm 0.2158 57 Talc,$oqstme,8m~qhi 0.62 58 Misc Wllic Minera 0.4429 59 Barite 0 60 Flurspar 0.637 61 Potash,$o:h,8 aerate Hi 0 62 Mute Rock 0 63 Rock Salt 1 64 mlfir 0.2147 6 Omficalfiertilizer Min 0 66 New Res Stunting 1 67 New lrdstrial 8 Cam 8 1 68 New Utility Stnctwes 1 69 Na: Him 8 Streets 1 70 New Farm Status 1 71 Na: Himl Extractim 1 2 New Militaw 8 Marble: 1 73 Ibintm 8 Repair, R 0.568 74 hintm 8 quair, 0 75 hintm 8 Rain 0 76 Calplete wick! Mixile Tl humiticnptc mall ar 78 Tats 8 13k Calpcmnts 1 1 0 1 1 79 Still N16 0.0749 69mllAnnknnitim 810MM8A666$ 6MPackimlets 5m80therPremd 1 0 1 1 84 Poultty Dressim lets 0.766 6Mtry8EmProcesin mama-vanter 870w, mt8proceased 6W8quaomted 61620113”. Frances: 90Fluidllilk 1 1 1 1 1 1 mmawmo.w 5 Cunad glacialties 93 Card Fruits 8 Veg 1 1 94 Ddlyd-ated Food Putts 0.556 95 Pid:les,$anes,8$alad 0 0.9526 5 Fresh or Frozen Palm F 0.0% 97 Pm an'ts,.hioes,8v69 1 5 Frozen Specialties 1 RPC:SPFLRLK 0. 1378 0 0 0 0.63 0.215 0.6252 0.” 0.11111 00000 0.8719 0.755 0.5971 0.715 0.3181 1.61) 0.7412 0.9271 OOOOOOO 0.416 0.167 0.136 0.5” 0.122 0.546 0.5!) 0.446 0.1746 0.1476 0.1m 0.0449 0.1738 0.1915 ALTFLR 0.1378 0.613 0.215 0.62 0.” 0.0001 0.670 0.619 0.564 0.8719 0.7% 0.5971 0.715 0.3181 1.0000 0.7412 0.361!) 0.416 0.167 0.136 0.5!) 0.1392 0.546 0.56 0.446 0.1746 0.1476 0.2750 0.1320 0.0449 0.1738 0.1915 192 0.534 0.138 0.9361 0.00:1; 0.9129 0.861 0.1159 0.612 0.3000 0.GD1 0.51) 0.619 1.0000 1.01!) 1.61) 1.0000 1.0000 1.0000 1.1!!!) 0.568 1.0000 1.0000 0 0.5“!) 0.501) 0 0.501) 0 0.9181 0.9% 0.765 0.9141 0.524 0.9349 0.9558 0.915 0.538 0.636 0.62 0.3149 0.3357 0.3640 0.“! 0.2978 0.4181 0.5 0.133 0.561 0.!!!13 0.9129 0.8061 0.1159 0.632 0.4429 0.0001 0.670 0.619 1.0000 1.0000 1.0000 1.0000 0.9181 0.9141 0.975 0.“!!! 0.“ 0.” 0.9454 0.750 0.670 0.0062 0.6m 0.0000 0.0115 0.612 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.1378 0.6259 0 0 0 0 "I"? 0.875 0.846 0.0% 0 0 0 0.215 0.69767 0 0 0.0269 0 0 0 0.687 0 0 0.00062 0 0 0.6 0.“!114 0.6 0.0003 0.6 0.41 0.5 0.13 0.6 0.16 0.54 0.5 0.44 0.6 0.34 0.33 0.6 0.40 0.6 0.27 0.29 0 1 1 1 1 0.7155 0.3181 1.66 0.7900 1.0000 OOOOOOO 0.915 0.542 0.559 0.504 0.550 0.9143 0.9313 mm 0.952 0." 0.543 0.671 0.1473 0.160 0.659 0.1937 0.3574 0 0 0.601 0.0001 0 0.564 0.8719 0.9116 0.541 0.4193 0.5147 0.55 0.935 0.514 0.5 0.929 0.62 0.1746 0.209 0.293 0.- 0.167 0.1738 0.543 .6 dad—D-b-D-b-l-D—Dooooooo P 0.6404 0.6404 0.6404 0.6404 0.6404 0.45% 0.45% 0.45% 0.45% 0.7558 0.7558 0.7558 0.7558 0.7558 0.2751 0.2751 0.2751 0.2751 0.2751 0.2751 0.2751 0.2751 Appendix C (cont'd.) $06. sacrum: 99Flar80therGI-ainflil 1 16 Gaul Win 1 101 Blathd 8. PW Fla: 1 16 Oommtfiother Pet Food 0.695 16 PM Postman: 0.56 106 Rice Hillim 0.266 16 6t Can Hillim 0.103 16 Bread,Cd(e,8Related Prd 0.661 107 Odds 6 Graders 0.661 16 au- 0.7163 109 Caiiectiaiety rehts 0.7666 110 Oncolate 8. Cocoa Prd:t 0.556 111 Mm am 1 112 mlt Lian 1 113 mlt 1 116 Uiresfirrd/fi 8116/ 5 0.136 115 Distilled Lianptc Br 0.3077 116 Bottled 8. Canad Soft D 1 117 Elmira Extracts 8. 5y 0.6261 118 Cottcreeed Oil Hills 0.1992 119 Sam Oil Mills 0.62 120 Vegetwle Oil Mills,nec 1 121 Animl 8. Him Fats 8. 1 122 Roasted Coffee 0.2636 123 Sid-tame 8. Cookim Oi 0.7016 126 mattered Ice 0.6789 15 Mi 8. Wi 1 126 Food Preparatiasmec 0.7516 127 Cigarettes 0.632 128 Cimrs 0.63 129 Oieuim 8. Smkiru 160:: 0.0382 130 166000 Stunnim 8. Rear 1 131 Brando/en Furic Mills 0.2712 132 Nam Fdric Mills 0.669 133 Yam Hills 8. Finishiru 0.1777 136 Thread Hills 0.677 135 Floor emeritus 0.666 136 Felt 0006mm 0.306 137 Lace Goods 0.2921 138 Put-lira 8. mnlstery Fi 0.7307 139 Processed Textile haste 0.6678 160 Cmtd Fdrics,mt rib 0.3167 161 Tine 00rd 8. Ftric 0.0136 162 Oaths: 8. twin 0.32 163 11am Purim 0.3619 166 'I’extile 000115.110: 0.952 165 this: 003137.010 0001: 0.0266 166 Hosietymac 0.2132 167 Knit Mm Mills 0 168 Knit ltdenear Mills 0 RPC:EPFLRLK 0.6071 0.- 0.1000 Odin 0.2500 0 0 0.376 0.2627 0.676 0.56 0.0000 0.126 0.536 0.1158 0.‘ 0.0m? 0.4300 0.116 0.1200 0.36 0.0730 0.36 0.6” 0.156 0.36 0.0012 0.1130 0.036 0.066 0.016 0.0116 01152 04156 0.62 0.656 0.66 0.0063 0.662 0.659 0.037 0.151. 0.0053 0 0 ALTFLR 0.6071 0.630 0.1000 0.156 0.2000 0.376 0.2627 0.676 0.56 0.276 0.1311 0.536 0.1158 0.‘ 0.687 0.636 0.116 0.376 0.66 0.36 0.030 0.36 0.6!) 0.156 0.36 0.0002 0330 0.66 0.5“!) 0.0054 0.016 0.0116 0.0052 0.0054 0." 0.6 0.66 0.“ 0.662 0.659 0.0737 0.66 0.053 0 0 193 0.65 0.7076 0.616 0.5632 0.560 0.2256 0.1072 0.661 0.661 0.5“!) 0.5000 0.1.000 0.6!!) 0.7666 0.756 0.136 0.3077 0.6697 0.6261 0.1992 0.5% 0.1.000 0.812 0.2636 0.7016 0.676 0.762 0.7516 0.1!!! 0.630 0.036 0.511!) 0.0162 0.0615 0.66 01162 0.666 0.1581 0.0779 0.660 0.1565 0.529 0.662 0.1787 0.66 0.637 0.666 0.1000 0 0 0.625 0.7076 0.616 0.566 0.5” 0.229. 0.1072 0.661 0.661 0.7163 0.7666 0.556 0.55 0.7666 0.876 0.136 0.313 0.6697 0.6261 0.162 0.620 0.5572 0.812 0.2636 0.7016 0.695 0.762 0.7516 0.11132 0.GBO 0.66 0.66 0.0162 0.0615 0.61) 01152 0.666 0.1581 0.0779 0.326 0.1565 0.529 0.662 0.1787 0.2!“ 0.0937 0.666 0.216 0 0 N1: 119 0.6691 0.621 0.666 0.1516 0.268 01188 0.1860 0.666 0.691 0.660 0.6617 0.1613 0.6 0.666 0.1158 0.2127 0.616 0.5111 0.58 0.236 0.167 0.163 0.793 0.1763 0.3627 0.663 0.816 0.5797 0.616 0116 0.670 0.953 01156 0.- 0.0116 0.665 0.036 0.159 0.1312 0.53 0.1671 0.8!) 0.6119 0.510 0.165 0.656 0.0200 0.016 0 0 0.43 0.44 0.10 0.15 0.26 0.00 0.00 0.37 0.30 0.10 0.29 0.00 0.12 0.53 0.13 0.09 0.43 0.1.3 0.11 0.00 0.00 0.02 0.36 0.15 0.35 0.02 0.15 0.33 0.00 0.00 0.00 0.00 0.03 0.01 0.20 0.00 0.01 0.09 0.00 0.00 0.03 0.55 0.00 0.1.2 0.10 0.16 0.00 0.01 0.0!. . 0 101119 0.5310 0.36 0.666 0.667 0.4300 0.366 0.1136 0.5975 0.2627 0.816 0.662 0.3090 0.1711 0.5319 0.7” 0.“ 0.687 0.5350 0.555 0.3162 0.666 0.6079 0.666 0.562 0.8166 0.668 0.835 0.361 0.0003 0.“!11 0.0766 0.9583 0.0155 0.63 0.0195 04156 0.656 0.62 0.0012 0.656 0.1719 0.“ 0.0126 0.076 0.037 0.0369 01166 0.- 0 0 Hill? 0.6071 0.66 0.602 0.6161 0.602 0.611 0.66 0.636 0.6763 0.676 0.7978 0.7599 0.1860 0.6867 0.8366 0.116 0.629 0.5106 0.363 0.0663 0.0669 0.2% 0.8638 0.0730 0.7% 0.367 0.667 0.816 04112 0.668 0.669 0.953 0.016 0.0187 0.0171 0.666 0.671 0.252 0.613 0.” 0.1121 0.131 0.677 0.659 0.950 0.1676 0.661 0.7965 0 0 LK R10 0.676 0.676 0.676 0.676 0.676 0.676 0.676 0.766 0.766 0.56 0.56 0.56 0.56 0.5761 0.5761 0.5761 0.5761 0.5761 0.5761 0.36 0.36 0.360 0.1616 0.1616 Appendix C (cont'd.) EC 10. 5015 ME 169 Knittim Hills,m 0 150 Knit Ffll‘ic Hills 0.625 151 ml fran W 0.6671 152 mm 8. quaeries 0.356 13 Whitman: 0.1669 156 Textile Bap 0.6166 155 Cum m 1 156 Platiro 8. Stitchim 0.157 157 Auto & annel triumim 1 158 Schiffi machine Humid 0.576 159 Fabricated textile PM: 0.575 160 Lowiru Chips 8. Cam-ac 0.6958 161 Sunills 8. leim Mill 0.3156 162 W 01min 8. Flo 0.6639 163 Saacial Prdn Sam’lls, 0.6062 166 Hillinrk 165 lbw Kitcha'l 051m 1 1 166M8-Softide8 0.619 167 Strumral Hood Halters 168 Prefbricated wood Nil 169 lbad Presewiru 170 Hand Pallets at! Skicb 171 Particlebaand 172 Head Prdcts, mt: 1 1 0.31m A 1 1 1 173Nldl.kldflaes&€mta 0.688 176 Narmlstrd 1d Hshld 0.5613 175 Wield Fu'nimJec 0.3264 176 Hi wimflvfiafl'm H 0.65 177 mulstrd H! mld R: 0.2297 178 Metal Hshld FunitLre 0.5963 179 Hattrmes rd Bethprin 0.6657 15 Hand Office Furniture 181 Metal Office Fu'nim 1 1 , 15 PLblic mildim Fu'nim 0.572 183 Hand Partiticm & Fixtu 186 Metal Partition 8. Fixt 15 Blim,m,mm 1 1 1 15 Fu'niure 6. Fixtures, n 187 Pulp Hi Us 188 Pmer Mills,exc Bld‘u P 1 15 W Mills 1 15 Emelqm 1 191 Smitan/ Parr Prodzts 1 1 1 1 0.2766 192 mildim Pmer 8. Blcb B 193 Par Coatirc ad Glazi 196 Osman: Textile an 0.55 195 Die-cut Paar I'd aner 1 196 Pressed 8. Molcbd Pulp 6 0.5861 197 Statia'eryleets 8. Re 0.779 1% Med aner 8. Pmerbr 0.605 RPC:SPFLfl.K 0 0.5” 0.51) 0.015 0.0900 0.577 0.065 0.576 0.215 0.215 0.215 0.065 0.4000 0.3900 0.5 0.38” 0.0259 0.1569 1 0.6277 0.6277 0.6277 0.265 0.165 0.5» 0.065 0.075 0.165 0.325 0.163 0.2900 0.55 0.3200 0.1579 0.175 0.!!!)7 0.516 0.11116 0.51) 0.537 011116 0.11113 0.” 0.512 0.0003 ALYFLR 0.5!) 0.5!) 0.5” 0.015 0.51) 0.5” 0.065 0.576 0.215 0.215 0.215 0.065 0.6“” 0.3900 0.53 0.3300 0.” 0.1569 0.6277 0.62" 0.6277 0.265 0.165 0.51) . 0.365 0.065 0.075 0.165 0.325 0.163 0.2900 0.55 0.32” 0.1579 0.175 0.55 0.29” 0.51) 0.015 0.55 0.55 0.175 0.55 0.55 0.517 0.55 0.5 196 0.625 0.0730 0.356 0.1669 0.6166 0.9169 0.157 0.3100 0.576 0.575 0.6936 0.3156 0.511!) 0.592 0.5967 0.518 0.615 0.551 0.3000 0.353 0.577 0.516 0.N5 0.611!) 0.5177 0.511 0.6635 0.597 0.565 0.5736 0.6111) 0.71!” 0.6879 0.576 0.561 0.6!!!) 0.675 0.155 0.6% 0.4000 0.4000 0.55 0.655 0.555 0.55 0.5110 0.55 0.55 0.3000 0.615 0.0730 0.356 0.1669 0.6166 0.915 0.157 0.9% 0.576 0.579 0.6936 0.3156 mum1 0.55 0.3967 0.518 0.615 0.8651 0.0551 0.353 0.857 0.516 0.775 0.7066 0.5177 0.511 0.665 0.597 0.565 0.5736 0." 0.% 0.6879 0.576 0.6661 0.657 0.675 0.55 0.1127 0.- 0.511 0.5 0.0129 0.576 0.511 0.515 0.517 0.512 0.GB6 0.516 0.065 0.2768 0.1!!!) 0.557 0.516 0.3072 0.955 0.065 0.658 0.375 0.5639 0.155 0.7766 0.- 0.568 0.55 0.0269 0.1569 0.9037 0.865 0.765 0.257 0.693 0.1627 0.6965 0.1561 0.665 0.353 0.9267 0.9716 0.7162 0.8731 0.775 0.7019 0.759 0.00115 0.11137 0.” 0.5% 0.513 0.576 0.55 0.513 0.” 0.511 0.512 0.01113 0.5 0.5 0.5 0.01 0.01 0.5 0.5 0.62 0.5 0.5 0.56 0.5 0.50 0.06 0.60 0.5 0.11 0.38 0.5 0.5 0.3 0.50 0.61 0.26 0.26 0.5 0.5 0.06 0.07 0.16 0.5 0.57 0.29 0.23 0.5 0.18 0.17 0.5 0.29 0.19 0.01 0.2759 0.0763 0.665 0.3% 0.9577 0.5979 0.577 0.566 0.510 0.6ZB 0.80% 0.2152 0.551 0.065 0.915 0.55 0.58 0.560 0.375 0.6776 0.W3 0.fl78 0.765 0.751 0.165 0.2679 0.6106 0.596 0.6669 0.5691 0.6110 0. 163 0.6732 0.8106 0.750 0.1579 0.556 0.012 0.512 0.516 0.- 0.5 0.536 0.5 0.17 0.5 0.5 0.5 0.5 0.5 0.399 0.065 0.019 0.538 0.0001 0.513 0.562 11111? 0.951 0.1366 0.62% 0.510 0.518 0.9366 0.568 0.065 0.1163 0.735 0.3m 0.555 0.6663 0.6768 0.9191 0.561 0.752 0.59 0.0910 0.785 0.9063 0.8668 0.7878 0.7693 0.55 0.579 0.659 0.5665 0.6738 0.7338 0.6755 0.615 0.7015 0.7652 0.56% 0.- 0.7697 0.516 0.11” 0.11137 0.515 0.“ 0.GB7 0.530 0.562 0.512 0.1156 0.519 0.0111 LK "210 0.6% 0.335 0.1616 0.6” 0.6% 0.63 0.6m 0.6% 0.6a 0.6a 0.6” 0.215 0.215 0.215 0.215 0.6277 0.6277 0.6277 0.6277 0.565 0.6277 0.6277 0.6277 0.6277 0.6277 0.3660 0.3660 0.3660 0.3660 0.3660 0.3660 0.658 0.658 0.658 0.6% 0.658 0.4053 0.658 0.3630 0.3630 0.3630 0.3630 0.3630 0.3630 0.3630 0.3630 0.3630 0.3630 0.3630 0.3630 Appendix C (cont'd.) mmmtmiOJSW 2m “slappers 0.2 21 Periodicals 0.27 32 Bank fiblidiim 0.4491 2'5 Book Printiro 1 204 Misc fiblidiirq 0.597 2:15 Columnist Printim 1 ans Lithorqiiic Plstmakin 0.9748 27 linifold fisirss fonts 0.75% an Glam 8 Looseleaf 1 29 Greetiru Cad fiblishin 0.09” 210 EIv-aviru 8. Plate Print 0.6349 211 Midirq 8. Related ll 0.4866 212 Typesettiru 0.9143 213 Photouv‘avim 0.6172 214 Electrotypiru 8. Stereot 0.291 215 lrdstriel immigor 0.5178 216 Nitrogsns 8. Pliesdiati 0.1954 217 Fertilizers, Hixim ml 0 218 Agricultlnl Chanicals, 0.2225 219 mu 8 mod Ounicals 0.454 20 Achesives 8. Sealants 0.678 21 Explosives 0.4472 222 Printim irit 0.693 28 Certain Bled: 0.136 24 menial Pmtiasm 1 225 Plstics hterials 8. Re 0.4437 2% synthetic fiber 0.501 227 Cellulasic W Fib 0.562 228 awic Fibers,mnellu 0.2229 229 0:13 0.6611 230 Soup 8. Other Detergents 1 251 Polishes 8. Sanitation G 1 82 arface Active Aamts 0.6564 ZS Toilet Prqau'atia's 0.52 ' 34 Points I Allis! Prbts 0.52 E Petrolun Refinim 0.559 256 Lubricatiru Oils 1. time 0.4978 87 Petrolam 8. Dual Putts mm 88 Pavim Mixtures 8. Block 0.6471 239 Asthelt Felts 8 Goetira 0.4699 240 Tires 8. lmer Tunes 0.259 241 fiber 1. Plastics Footu 1 242 Reclaim m 0.3232 243 Marinated fiber Prd:t 0.338 244 Misc Platics Putts 0.279 245 fiber 8. Plastics Hose 0.814 246 Leather Tmiru I. Finis 1 247 Footwear on Steel: 0.122 248 fines, en: fiber 0.652 rum: 0.212 0.5146 0.1848 0.1140 0.4436 0.429 0.521 0.452 0.4697 0.462 0.2” 0.593 0.247 0.1.200 0.152 0 0.217 0.0050 0 0.012 0 0.252 0 0.52 0 0.252 0.232 0.172 0 0 0.0497 0.2718 0.032 0.125 0.1744 0.213 0.1000 0.042 0.23 0 0 0.1188 0.244 0.- 0.212 0 0.18!) ALTFLR 0.212 0.5146 0.1848 0.1140 0.4436 0.4909 0.521 0.452 0.427 0.462 0.22 0.595 0.3247 0.1.2:!) 0.152 0.4000 0.217 0.4150 0.012 0.2000 0.0652 0.2000 0.52 0.0652 0.22 0.1700 0.131) 0.121) 0.0497 0.2718 0.22 0.2700 0.272 0.155 0.1744 0.213 0.2000 0.1% 0.042 0.072 0.52 0.- 0.13!) 0.52 0.6m 0.212 0.1510 195 0.6% 0.562 0.270 0.4491 0.529 0.5791 0.551 0.5747 0.581 0.5317 0.277 0.5819 0.456 0.5% 0.5561 0.542 0.5179 0.172 0.2174 0.454 0.6785 0.3” 0.6930 0.132 0.246 0.4437 0.501 0.562 0.229 0.2!) 0.52 0.6!!!) 0.6564 0.53!) 0.5% 0.560 0.4978 0.60” 0.5000 0.3“!) 0.217 0.2!) 0.10” 0.101) 0.52 0.063 0.7577 0.122 0.427 0.527 0.5672 0.571 0.44% 0.529 0.5791 0.551 0.5747 0.581 0.517 0.277 0.5819 0.4870 0.55% 0.5561 0.542 0.5179 0.172 0.2174 0.454 0.672 0.442 0.6930 0.132 0.246 0.4437 0.501 0.562 0.229 0.1138 0.9587 0.55 0.564 0.5200 0.5% 0.231 0.5185 0.&7 0.6471 0.4699 0.217 01187 0.- 0.244 04150 0.- 0.7577 0.1. 0.427 III: II? 0.756 0.5214 0.1848 0.1140 0.5445 0.5688 0.521 0.5% 0.5143 0.4857 0.210 0.293 0.4714 0.5272 0.4% 0.4438 0.5131 0.281 0.1579 0.212 0.548 0.25 0.67% 0.60 0.8927 0.5% 0.151 0.12.52 0.658 0.1278 0.9341 0.7091 0.” 0.224 0.25 0.125 0.523 0.“ 0.542 0.127 0.219 0.1136 0.215 0.0059 0.0052 0.277 0.5730 0.274 0.421 0.21 0.2 0.2 0.24 0.56 0.59 0.62 0.45 0.54 0.48 0.2 0.27 0.5 0.42 0.15 0.2 0.34 0.01 0.11 0.01 0.2 0.16 0.2 0.5 0.2 0.29 0.2 0.17 0.2 0.2 0.21 0.29 0.03 0.2 0.2 0.2 0.19 0.2- 0.2 0.10 0.04 0.07 0.2 0.2 0.13 0.36 0.2 0.52 0.2 0.18 0.591 0.5146 0.558 0.52 0.4436 0.4909 0.5710 0.5241 0.4697 0.5124 0.11% 0.622 0.3247 0.5246 0.488 0.5778 0.4559 0.1491 0.1036 0.4217 0.510 0.922 0.9111 0.8347 0.575 0.28 0.2334 0.156 0.0497 0.615 0.699 0.9160 0.- 0.135 0.427 0.556 0.527 0.521 0.275 0.018 0.0001 01150 0.258 0.244 0.0.54 0.215 0.0000 0.4070 Hill? 0.%2 0.55 0.550 0.550 0.5776 0.521 0.5381 0.532 0.584 0.462 0.112 0.442 0.4646 0.4678 0.562 0.432 0.542 0.142 0 0.356 0.215 0.7140 0.1323 0.5158 0.546 0.57 0.4047 0.3901 0.3% 0.0788 0.092 0.870 0.8748 0.755 0.22 0.242 0.1744 0.4318 0.3218 0.244 0.218 0.“ 0.214 0.‘ 0.247 0412 0.- 0.7832 0.4036 0.442 LK RIO 0.672 0.562 0.562 0.562 0.562 0.562 0.562 0.562 0.562 0.562 0.562 0.562 0.562 0.562 0.562 0.562 0.217 0.182 0.122 0.122 0.252 0.0652 0.0552 0.252 0.252 0.252 0.4045 0.4045 0.4045 0.4045 0.372 0.2718 0.2718 0.2718 0.2718 0.615 0.213 0.213 0.213 0.213 0.213 0.24% 0.24% 0.302 0.24% 0.242 0.24% 0.321 0.321 0.321 Appendix C (cont'd.) SEC 10. m M RPC: 81’ 249 Muse Slippers 0.158 50 Leather Glam 8 Mittm 1 51 League 0.1418 52 m Ham 8 fines 0.297 53 Personal Leather Goods 0.257 54 Leather 0005,73: 0.1697 55 films 8 Glass Prtttspl 0.25 56 Glass Cmteimrs 0.1759 57 CaImtJMhulic 0.8475 58 Brick 8 Strumral Clay 0.1692 59 Ceranic tall 8 Floor Ti 0.543 50 Clay Refractories 0.0717 51 Stnctwal Clay Putts, 0.234 262 Vitreus Plurbiru Fixtu 1 53 Vitreaa (him Food Ute 0.0115 54 Fire armature Food U 0.52 55 Porcelain Electrical 9; 1 56 Pottery Pnbtsmec 0.505 57 l‘mcrete Block 8 Brick 1 58 Oatrete Prcbtsmec 1 269 Ready-Mixed Cacrete 0.9456 270 Line 0.4363 271 Gypsum Putts 0.0915 272 wt Stale 8 Stme Putt 1 273 Abrasive Prdcts 1 274 Asbestos Prctts 0.1409 275 Gsketsfiaekim 8 Seeli 0.5456 276 Mineralsfimm or Trea 0.945 277 Mineral wool 0.4341 278 Marley Refractories 0.5424 279 Metallic Mineral Prd 0.3922 22 Blast Fu‘raoes 8 Steel 0.4575 281 Electraretalllrgical Pr 0.292 a Steel wine 8 Related Pr 1 215 Gold Finishiru of Steel 0 284 Steel Pipe 8 Tunes 0 E lrm8 Steel Feud-ies 0.9391 25 im 8 Steel Forgims 1 287 Metal Meat Treatim 1 as Primary Metal Putts, 71 0.4974 209 Prim Comer 0.3448 290 Prim Lend 0.6091 291 Primary Zinc 0.1128 292 Prim Alunirun 0.3613 293 Prim Naiferrus Meta 0.1404 294 W Wet-rats Me 0 295 Owner Rolliru 8 Drain 0.6097 296 Alunirun Rolliro 8 Draw 0.3113 297 Mm Rolliro 8 Dr 0.52 22 Mmferras Hire Dmirq 0.3661 FLRLK 0.272 0.2!) 0.012 0.” 0.2007 0.127 0.0003 0.0001 0.2” 0.201 0.25 0.2 0.0000 0.0000 0.0559 0.11505 0.01:.7 0.0009 0.0003 0.0105 0.0.17 0001.1 0.0219 0.0005 0.0009 0.0090 0.0790 0.0107 0.0252 0.1200 0.217 0.0751 0.124 ALTFLR 0.272 0.2!) 0.012 0.3130 mm 0.387 0.127 0.1.200 0.162 0.012 0.042 0.2” 0.5000 011134 0.52 0.472 0.1250 0.5000 0.5“!) 0.472 0.2” 0.28 0.22 0.1639 0.241 0.1310 0.472 0.2200 0.272 0.212 0.25 0.246 0.1200 0.362 0.3!!) 0.124 0.22 0.121 0.1374 0.259 0.127 0.236 0.25 0.0903 0.2!) 0.050 196 0.158 0.55% 0.1418 0.297 0.452 0.1697 0.2535 0.1759 0.7000 0.152 0.214 0.213 0.1” 0.1484 0.12104 0.52 0.452 0.2000 0.4!!!) 0.452 0.3!!) 0.52 0.28 0.2000 0.52 0.241 0.3111) 0.52 0.52 0.452 0.2000 0.3!!) 0.246 0.3000 0 0 0.452 0.52 0.2150 0.1642 0.121 0.1374 0.239 0.127 0.297 0 0.1843 0.1451 0.1417 0.1188 0.158 0.55% 0.1470 0.297 0.6299 0.1734 0.543 0.1779 0.0158 02150 0.214 0.213 0.0455 0.1484 041114 0.22 0.7214 0.0200 0.0732 0.281 0.214 0.- 0.22 0.219 0.13 0.241 0.1310 0.224 0.223 0.297 0.382 omas 0.246 0.4584 0 0 0.2 0.2% 0.2150 0.1642 0.121 0.1374 0.25 0.127 0.297 0.1843 0.1451 0.1417 0.112 MI: 11? 0.204 0.442 0.507 0.012 0.1874 0.12% 0.271 0.127 0.0414 0.112 0.218 0.212 0.07.9 0.234 0.22 0.0987 0.0992 0.0167 0.059 0.22 0.0147 0.0101 0.562 0.015 0.1136 0.!!20 0.1445 0.29 0.112 0.122 0.529 0.1174 0.045 0.5093 0 0 0.246 0.52 0.521 0.2137 0.1464 0.1768 0.234 0.1236 0.042 0.575 0.3739 0.1769 0.1530 0.2 0.27 0.2 0.04 0.2 0.10 0.43 0.14 0.70 0.16 0.01 0.04 0.2 "I"? 0.0175 0.542 0.124 0.‘ 0.4307 0.0% 0.591 0.452 0.0003 0.2 0.- 0.!!01 0.248 0.2 0.00002 0.2 0.00001 0.2 0.47 0.30 0.50 0.53 0.65 0.5 0.55 0.2 0.37 0.2 0.31 0.58 0.37 0.48 0.21 0.5 0.5 0.12 0.2 0.13 0.36 0.38 0.18 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.11 0.2 0.10 0.219 0.0000 0.0008 0.212 0.®9 0.- 0.!!!)9 0.25 0.151 0.55 0.2! 0.259 0.1155 0.1875 0530 0.278 0.0407 0.2% 0.321 0.217 0.0751 0.124 0.0736 0.258 0.1043 0.278 0.249 0.236 0.25 0.0905 0.22 0.” "1119 0.5013 0.527 0.4134 0.22 0.6974 0.332 0.2%1 0.1110 0.0013 0.2 0.11101 0.- 0.2 0.4749 0.” 0.2 0.842 0.0436 0.072 0.04” 0.273 0.22 0.0003 0.25 0.0417 0.0117 0.125 0.217 0.0462 0.204 0.072 0.211 0.®2 0.3819 0 0 0.218 0.211 0.212 0.1574 0.®7 0.1441 0.0476 0.1340 0.1262 0 0.244 0.09m 0.1693 0.1369 LK ”110 0.321 0.321 0.321 0.331 0.321 0.321 0.337 0.237 0.7224 0.724 0.724 0.724 0.724 0.7224 0.724 0.724 0.724 0.724 0.724 0.724 0.7.21. 0.7224 0.724 0.7224 0.224 0.7224 0.7224 0.7224 0.724 0.724 0.724 0.227 0.&7 0.452 0.2067 0.2067 0.4976 0.2357 0.2367 0.2057 0.2191 0.2191 0.2191 0.2191 0.2191 0.2191 0.2191 0.2191 0.2191 0.2191 Appendix C (cont'd.) $8 10. 812 ME RPC: 91’ 299 Almirun Castim 1 32 Brasfim'meJquer 0a 1 301 m Castirwgnc 1 32 Marie-res Forgirw 0.975 23 Metal m 0.712 304 Ibtal Barrels,0nns,8 P 0.445 32 Metal Smitary lhre 1 32 Plu1biru Fixtu'e Fittin 1 307 Meatim Swim,” El 0.9971 32 qurieated Metal 1 309 Metal Doors,Sd1,8 Trim 1 310 Ftrieated Plate Hark ( 0.726 311 sieet Metal work 1 312 Ardiitectlral Metal Her 1 313 Pnefdarioatd Metal Bld 0.4396 314 Misc Metal Hark 0.4092 315 Screu itchine Prchts 8 1 316 Momtive Stmpiros 1 317 Coats 8 Cleans 0.4092 318 Metal Starpimsmac 1 319 Cutlety 0.1929 320 Had 8 Ecbe Tools,nac 1 321 Had Saws 8 Sad Blah: 0.332 35 "mn,mc 1 E Platiru 8 Polishirc 1 324 Metal Coatim 8 Allied 0.765 35 Misc Farieated Hire Pr 0.8777 35 Steel Sal-1195,96 wire 0.3291 327 Pipe,Valves,8 Pipe Fitt 0.7096 35 Metal Foil 8 Leaf 0.558 329 Ftrieated Metal Putts 0.7872 330 Stean Erainec 8 Tu'bire 0.259 331 lnterml Covhstim Era 1 332 Fem Machinely 8 Eqfip 1 333 L501 8 @1131 Emip 1 334 Caetnstim Machinery 1 35 Miniru Ibdiineryfixc oi 0.924 56 Oil Field Ibdiinew 0.1047 337 Elevators 8 Moviru Stai 0.6572 58 Cam 8 Caweyor Eq 1 59 Hoists,Crm,8 Mcmrai 1 340 Irdstrial Tncks 8 Tra 1 341 Inchim Tools,Metal wt 1 342 Mine Tools,Metal For 0.5584 343 wial 0m 8 Tools 8 1 344 Power Driven Had Tools 0.3514 345 Rollin Mill winery 0.1848 346 Metaluorkiru ilk-dainty, 1 347 Food Prebts Machinery 1 348 Textile minty 0.0752 FLRLK 0.172 0.22 0.132 0.0781 0.0411 0.1000 0.66 0.276 0.1419 0.15 0.122 0.142 0.2129 0.1747 0.072 0.132 0.232 0 0.250 0.52 0.012 0.201) 0 0.3247 0.412 0.52 0.52 0.22 0 0 0 0.22 0.2942 0.3900 0.3901 0.1010 0.224 0 0.242 0.4216 0.242 0.3768 0.412 0.52 0.5% 0.22 0.22 0.3900 0.2937 0.012 ALTFLR 0.172 0.22 0.1300 0.0781 0.0411 0.1000 0.®6 0.276 0.1419 0.19 0.1092 0.142 0.2129 0.1747 0.072 0.132 0.582 0.5110 0.250 0.3!) 0.012 0.2000 0.172 0.3247 0.412 0.52 0.583 0.22 0.52 0.52 0.3900 0.22 0.222 0.3900 0.3901 0.1010 0.0901 0.23 0.242 0.4216 0.242 0.3768 0.412 0.52 0.525 0.22 0.22 0.3900 0.2937 0.012 197 0.2461 0.2733 0.2270 0.3000 0.3792 0.3119 0.5000 0.3569 0.2273 0.55 0.3257 0.3456 0.3107 0.1596 0.1782 0.2510 0.3275 0.5313 0.1662 0.3009. 0.1000 0.1102 0.3021 0.5270 0.5119 0.5248 0.1135 0.3291 0.1171 0.1110 0.1053 0.2359 0.7000 0.6000 0.7000 0.3500 0.1500 0.0373 0.1110 0.6457 0.6301 0.6166 0.5511 0.1000 0.6000 0.3511 0.1848 0.7178 0.6000 0.0752 0.2461 0.275 0.270 0.356 0.3792 0.3119 0.772 0.569 0.2278 0.52 0.37 0.3456 0.3407 0.3% 0.172 0.510 0.323 0.5813 0.1662 0.332 0.126 0.412 0.312 0.5270 0.5149 0.5248 0.415 0.3292 0.4174 0.4140 0.423 0.59 0.853 0.” 0.” 0.53 0.329 0.23 0.4140 0.6457 0.201 0.6166 0.5541 0.552 0.7777 0.514 0.1848 0.7178 0.7572 0.0752 MT: "P 0.274 0.2912 0.2734 0.2 0.0411 0.517 0.221 0.2791 0.1419 0.16 mm 0.1537 0.849 mm 0.127 0.127 0.4191 0.759 0.214 0.3441 0.184 0.3417 0.1301 0.524 0.5453 0.5117 0.4439 0.3136 0.433 0.2970 0.91 0.353 0.853 0.818 0.244 0.1010 0.224 0.- 0.272 0.6950 0.5% 0.652 0.572 0.6165 0.8197 0.142 0.1681 0.783 0.4471 0.072 0.17 0.2 0.13 0.34 0.2 0.10 0.24 0.18 0.8 0.24 0.5 0.8 0.2 0.36 0.07 0.13 0.37 0.56 0.5 0.5 0.01 0.5 0.2 0.46 0.41 0.5 0.29 0.2 0.2 0.2 0.38 0.2 0.58 0.5 0.40 0.39 0.52 0.2 0.24 0.52 0.24 0.42 0.42 0.5 0.5 0.2 0.2 0.5 0.5 0.01 0.212 0.1687 0.121 0.0781 0.568 0.83 0.5 0.276 0.1439 0.195 0.182 0.142 0.2129 0.1747 0.0747 0.142 0.232 0.250 0.3845 0.207 0.1391 0.2927 0.3247 0.458 0.4713 0.- 0.2448 0 0.427 0 0.2479 0.222 0.3909 0.622 0.1784 0.184 0.‘ 0.5“!) 0.4216 0.785 0.512 0.412 0.4134 0.55 0.24% 0.1m 0.5619 0.7568 0.0779 Hill? 0.315 0.52 0.2487 0.1900 0.4794 0.1437 0.240 0.511 0.2118 0.355 0.121 0.412 0.2% 0.529 0.1939 0.1951 0.2738 0.1101 0.1931 0.462 0.574 0.33 0.2931 0.5373 0.132 0.49” 0.3759 0.3158 0.3955 0.365 0.3715 0.520 0.8573 0.0061 0.250 0.362 0.4313 0.456 0.423 0.5183 0.3768 0.5271 0.3319 0.6457 0.522 0.592 0.343 0.7576 0.2 LK HUD 0.2191 0.2191 0.2191 0.2191 0.452 0.452 0.5516 0.5516 0.5516 0.5516 0.5516 0.5516 0.5516 0.5516 0.5516 0.5516 0.316 0.316 0.316 0.316 0.455 0.452 0.4520 0.4520 0.452 0.452 0.4520 0.4520 0.4520 0.4520 0.455 0.4887 0.4887 0.3901 0.321 0.412 0.4183 0.412 0.6151 0.6151 0.6151 0.6151 0.626 0.626 0.626 0.626 0.626 0.626 0.2937 0.2937 Appendix C (cont'd.) SEC 10. 5010 ME PC: 91’ 369 Wits Ibohiruy 1 350 Pmer irdstris Min 1 $1 Printirc Truhs lbchine 0.662 352 anciel lrdstry Ibohin 0.36 35 Hm & 0:11pm 0.7971 56 Bell 8 Roller Beerim 0.2627 55 Blowers 8. Fa: 0.8878 356 Mistrial Potter: 1 357 Paar Tmuniseim Eqfi 1 358 Mistrial Hm I. 0 1 359 Metal 1rdstriel inch 0.565 360 More,l>istas,kin 0.8% 361 Miran/pic Electrice 0.95& 362 Electrcnic Camim Eq 1 363 Celwlatiro 8. Acountin 0.5568 366 Scales 8. Belarus 0.5677 365 Typeuriters 8. Office lb 0.1635 366 Momtic Hawaisiru 0.9168 367 Oamiel Lamb] 6qu 0.6767 368 Refrigeration & Heetiru 0.566 369 Heearim 8. Dimim 0.5966 370 Service lrdstry Ibdiin 1 371 lmtrurents to Hanan 0.6899 372 Tmfomers 1 373 slitdigeer 8. Suitdboer 0.6127 376 Motors 8 Gemmtors 1 375 [mistrial Cmtnols 1 376 Heldim Amman, Elec 1 3T! Cerium 8 Graphite Prcht 0.%18 378 Electrical [mistrial A 0.6836 379 Wald Cooking Eqsip 3m Wield Refrigerators 381 Mld may Emip 382 Electric unsure: 8. F 33 Handwold Vecum Cleme 386 Selim Mines 0.2176 3% Roadbld Again-neg: 1 3% Electric Lulu 0.1687 387 Liwtiro Fixtires I'd E 0.607 388 Uiriro Devices 0.8903 359 Radio 8. TV Receiviru Se 0457 3% W Recount 8 Ta 0.2756 391 Teled'me 8 Teleg-qii A 0.0952 32 Mo Tv Cummicatim 0.675 393 Electrm Tm 0.0771 336 Senicadntors & Relate 0.1329 395 Electrmic Cows-Item 0.6155 396 Storm Batteries 0.675 397 Primary Batteries, Dry 0.7522 3% x-rq W 8. Tibet 1 ddddd rum: 0.1500 0.1600 0.1300 0.2937 0.111113 0.00003 0.0001 0.11112 0.1!!! 0.0001 0.01111 0.111131 0.0001 0.16m 0.11101 0.00001 omen 0mm 0.0500 0.27% 0.1900 0.27% 0.1511) 0.%m 0.06“) 0.1310 0.0900 0.2000 0.1!!) 0.0!!) 0.“ 0.0155 0125!. 0.m78 0.0116 0.“!!! 0.%18 0.0000 0.1152 0.1156 0.1186 0.07% 0.00m 0.06m 0.111111 0.00001 0.60 0.1200 0.Q75 0.07m ALTFLR 0.1500 0.1400 0.1300 0.293? 0.311) 0.3100 0.2500 0.36% 0.3m 0.2700 0.1200 0.36% 0.3400 0.16m 0.26m 0.26m 0mm 0.0200 0.0500 0.27% 0.1900 0.27% 0.1!!!) 015% 0.061!) 0.1200 0.0900 0.31” 0.1300 0.0300 0.%5 0.0155 0.66 0.m78 0.0116 0.0175 0.1518 0.0107 0.(IB2 0.1156 0.1186 0.07% 0.0952 0.06% 0.0771 0.139 0.60 0.111) 0.Q75 0.0700 198 0.7500 0.5% 0.66“) 0.7500 0.6110 0.2000 0.1!!) 0.61!!! 0.3000 0.3000 0.2000 0.5500 0.7500 0.3000 0.4000 0.60!) 0.165 0.6“” 0.3000 0.5500 0.6“!) 0.7111) 0.36% 0.3909 0.30 0.6730 0.5512 0.’ 0.3977 0.2110 0.1%6 0.6“” 0.581 0.Wo3 0.339 0.0175 0.3000 0.0107 0.1!!!) 0.1500 0.67 0.2756 0.G52 0.311!) 0.0771 0.129 0.3500 0.67% 0.7.2 0.569 0.937? 0.851 0.4420 0.92115 0.11131 0.0001 0.11111 0.0001 0.1112 0.1112 0.1112 0.0mm 0.0”1 0.9518 0.5568 0.565 0.1636 0.9116 0.6767 0.“ 0.5956 0.999 0.3660 0.3909 0.30 0.6730 0.5512 0.’ 0.37? 0.2110 0. 1%6 0.551 0.381 0.0963 0.25 0.0175 0.137 0.0107 0.1075 0.01m 0.” 0.2756 0.0952 0.67% 0.0771 0. 129 0.5016 0.6792 0.72 0.669 0.656 0.6613 0.25% 0.9199 0.00m 0.“ 0.00m 0.Gl)1 0.00:3 0.0113 0.0002 0.0001 0.0001 0.1736 0.6665 0.” 0.07% 0.1860 0.311? 0.6156 0.665 0.9999 0.2993 0.153 0.2673 0.3621 0.35% 0.5%6 0.331 0.1739 0.16m 0.6” 0.“ 0.N78 0.1366 0.N16 0.1103 0.11152 0.0153 0.11% 0.”? 0.3119 0.0267 0.5516 0.%10 0.152 0.399 0.2090 omen QM 0.15 0.16 0.13 0.3? 0.35 IN"? 0.%21 0.1668 0.6656 0.9200 0.0001 0.3 0.0003 0.5 0.” 0.3 0.0612 0.25 0.27 0.39 0.0006 0.11111 0 0.52 0.00001 0.76 0.16 0.00 0.00 0.03 0.12 0.15 0.66 0.19 0.38 0.% 0.% 0.06 0.12 0.09 0.20 0.13 0.03 0.16 0.10 0.09 0.01 0.13 0.00 0.3 0.00 0.Q 0.07 0.11 0.12 0.1!) 0.06 0.00 0.1!) 0.3 0.12 0.06 0.07 0.0001 0.9518 0.6677 0.”?6 0.3085 0.9795 0.780 0.975 0.923? 0.9998 0.6360 0.2172 0.572 0.6567 0.298 0.3106 0.663? 0.%61 now. 0.19% 0.066 0.01 13 0.3161 0.0667 0.1518 0.0100 0.052 moms 0.1m. 0.1973 0.2910 0.3939 0.1325 0.1365 0.6616 0.9328 0.Q75 0.%1? 11111? 0.92% 0.531 077% 0.9m. 0.” 0.0001 03131 0.0002 0.009 0.00m 0.0001 0.0001 0.0002 0.193 0.4250 0.25113 0.1503 0.97% ma 0.9726 0.2121 0.W99 0.2921 0.5557 0.6326 0.5517 0.7156 0.556 0.336 0.35 0.“ 0.0155 0.168 0.1768 0.0116 0.0015 0.1703 0.0163 0.0063 0.‘ 0.1225 0.337 0.w7 0.529 0.0309 0.%87 0.3.3 0.” 0.5755 0.” LK 10110 0.293? 0.293? 0.293? 0.293? 0.3626 0.36% 0.3626 0.3626 0.3626 0.3626 0.368. 0.3626 0.3626 0.2619 0.2619 0.2619 0.2619 0.27% 0.27% 0.27% 0.27% 0.27% 0.2232 0.2282 0.2282 0.223 0.2252 0.22& 0.22m 0.2% 0.586 0.536 0.580 0.5286 0.5221. 0.536 0.586 0.171? 0.1717 0.1717 0.07% 0.07% 0.397 0.25? 0.3115 0.3133 0.3133 0.&3? 0.35? 0.”? Appendix C (cont'd.) 50 10. $015 ME RPC: 91’ 399 Emine Electrical Equip 0.318 65 Electrical Emipmac 0.7665 601 1nd: 8 as 00010:: 0.5m 65 Tank Trailem 0.9022 603 Motor Vehicles 1 606 Motor Vehicle Parts 8 A 1 65 Aim-aft 0 65 Aircraft 8 Missile Emi 0.9669 607 Aircraft 8 Missile Emi 0 65 Ship wildim 8 aniri 1 609 float wildiro 8 Repiri 1 610 Railroad Emip 0.333 611 Motomyclsfiicyclsfi 0.6912 612 1ravel Trailers 8 Cape 0.5366 613 Mdaile Mam 0.665 616 Motor Mares 0.539 615 inundation EmipJe 1 616 Emimriro 8 Scimtifi 1 617 Madmical Measlriru De 0.568 618 Monatic Imperatu‘e C 0.965 619 mical 8 mdical lrst 0.7396 620 amical Applirces 8 S 0.7669 621 De1tal Eqfip 8 smlies 0.6993 63 lhtdiesfilocksfi Parts 0.6507 63 Cptical Imtrumnts 8 L 0.6271 626 Mthalmic 00008 0.5711 63 Photoy-qific Emip 8 Su 0.3397 626 Joelty,Precias Metal 0.377 627 Jaelers Materials 8 La 0.0792 628 Si lvenare 8 Plated lhr 0.565 629 Dostme Jewelry 0.3665 630 itsical lmtrmmts 0.759 631 GatesJoysA Oiildwe‘s 0.5651 632 Dolls 0.3951 633 aartiru 8 Athletic Good 0.9697 636 Pas 8 Med-mical Pati 0.6265 635 LedPacils8ArtGood0.383 636 Miro Devices 1 637 Cerium Pmer 8 Irked Ri 0.3583 638 Artificial Trees 8 Flou 0.6128 639 ”tan 0.1663 660 Maedles,Pim, 8 Fasten 0.2599 661 0110013 8 m 1 662 Hand arface Floor Dove 0.0709 663 arial Cskets 8 Vaults 0.655 666 Sign 8 Achertisim Dis 1 665 Wactu‘im irdstrie 0.751 666 Railroads 8 Related a»: 0.9921 667 Local,1nteruh1 Paesm 0.539 668 Motor Freidit Tru'eport 1 rum: 0.2200 0.0100 0.11.1.9 0.1001 0.0091 0.1145 0 0.00001 0 0.1540 0.1991 0.0011. 0.0100 0.1350 0.0005 0.0002 0300 0.1005 0.0000 0.00001 0.0000 0.0700 0.0000 0.0100 0.0500 0.015 0.035 0.111111 0.00001 0.111131 0.0176 0.566 0.1010 0.55 0.0900 0.591 0.521 0.- 0.1015 0.0753 0.3!!” 0.2900 0.3911 ALTFLR 0.2200 0.015 0.1669 0.131 0mm 0.1165 0.350 0.560 0.151 0.516 0.065 0.156 0.00115 0.0102 0.31) 0.153 0.0200 0.66!) 0.55 0.075 0.51) 0.0160 0.035 0.015 0.0200 0.563 0.0166 0.518 0.0176 0.566 0.1676 0.1010 0.1630 0.509 0.51) 0.591 0.512 0.0152 0.566 0.0921 0.013 0.- 0.1015 0.0763 0.3000 0.2900 0.3911 199 0.318 0.35 0.1!!) 0.151 0.511!) 0.655 0.3000 0.363 0.355 0.2168 0.1812 0.359 0.1000 0.3263 0.6115 0.366 0.232 0.5% 0.555 0.0000 0.1.000 0.1% 0.1.000 0.355 0.113 0.157 0.563 0.0166 0.518 0.139 0.351 0.1676 0.1715 0.1630 0.033 0.336 0.0768 0.512 0.0152 0.566 0.365 0.0138 0.58 0.1939 0.125 0.0000 0.7M 0.9000 0.2023 0.3731 0.1669 0.151 0.3316 0.35 0.635 0.363 0.315 0.2168 0.1812 0.339 0.518 0.263 0.6615 0.366 0.2222 0.6070 0.73% 0.7677 0.6993 0.1& 0.6271 0.5711 0.113 0.157 0.563 0.0166 0.518 0.125 0.351 0.1676 0.1715 0.1630 0.053 0.386 0.0768 0.512 0.0152 0.566 0.3705 0.0138 0.5% 0.1939 0.133 0.W78 0.0995 1.0000 0.361 0.3762 0.16% 0.33 0.655 0.633 0.“ 0.151 0.351 0.516 0.065 0.3559 0.0000 0.3661 0.381 0.1003 0.131 0.0799 0.5% 0.3566 0.6651 0.113 0.6677 0.237 0.037 0.515 0.0188 0.0166 0.- 0.131 0.566 0.0m1 0.15% 0.539 0131 0.1065 0.599 0.552 0.562 0.0167 0.“! 0.013 0.575 0.1015 0.0763 0.6660 0.55% 0.311 0.3 0.01 0.61 0.3 0.93 0.5 0.625 0.3001. 0.1591 0.2136 0.0091 0.1165 0 0.02219 0.3 0.31 0.3 0.58 0.11 0.06 0.61 0.11 0.3 0.3 0.17 0.5 0.5 0.5 0.07 0.5 0.5 0.03 0.5 0.01 0.5 0.5 0.5 0.5 0.23 0.3 0.5 0.35 0.5 0.03 0.5 0.07 0.5 0.5 0.5 0.3 0.5 0.5 0.29 0.3 0.30 0.3 0.66 0.11% 0.560 0.310 0.186 0.065 0.263 0.0005 0.566 0.536 0.2169 0.36% 0.7681 0.8159 0.9113 0.365 0.015 0.757 0.6635 0.3005 0.2261 0.1150 0.560 0.151 0.0176 0.315 0.356 0.1010 0.578 0.0009 0.136 0.591 0.562 0.01% 0.579 0.0921 0.1180 0.- 0.163 0.1606 0.832 0.511 0.392 0.315 0.3011 0.2613 0.236 0.35 0.2768 0.2115 0.1991 0.135 0.336 0.1356 0.0133 0.01m 0.313 0.363 0.313 0.675 0.813 0.51? 0.736 0.321 0.5613 0.5526 0.516 0.- 0.113 0.0160 0.®5 0.0901 0.578 0.- 0.1313 0.350 0.0671 0.2168 0.560 0.0090 0.015 0.1219 0.373 0.0166 0.0761 0.2166 0.1127 0386 1.11!” 1.0000 LK 10610 0.35? 0.35? 0.659 0.659 0.659 0.659 0.37% 0.115 0.1113 0.356 0.356 0.356 0.356 0.3356 0.565 0.356 0.356 0.6591 0.6591 0.6591 0.31% 0.31% 0.3165 0.6591 0.315 0.3166 0.6591 0.1556 0.1556 0.1556 0.1556 0.1556 0.1556 0.1556 0.1556 0.1556 0.1556 0.1556 0.1556 0.1556 0.1556 0.1556 0.1556 0.1556 0.1556 0.1556 0.1556 0.7168 0.501 0.973 200 Appendix C (cont'd.) seem. mm 11mi» mu ALTFLR 03m: 00m Mun? 119111011me uumo 669W TWim 0.6292 0.15” 0.165 0.623 0.- 0.6019 0.16 0.5927 0.7617 1.0000 650Air immortatkn 0.252 0.1900 0.1900 0.0000 0.7661 0.6% 0.19 0.975 0m 0.315 651?“: Lira,” Neural 0.53 0.3200 0.3 0.55 0.558 0.66% 0.3 0.7613 0.- 0.637 652 Freid'it PM 8 0! 0.7392 0.6630 0.6630 0.7390 0.915 0.6630 0.91 0.756 0.6516 0.0999 653m of Faeroe 1 0.6630 0.665 0.9000 1.0000 0.663 0.“ 0.330 1.51) 0.” 656 Mimics,” Rafi 0.765 0.665 0.665 0.55 0.7913 0.623 0.66 0.7516 0.%2 0.537 65 R8310 8 TV Witt 0.8791 0.636 0.636 0.755 0.8792 0.693 0.87 0.99m 0.9127 0.636 656 Elatric chs 0.8711 0.531 0.531 0.8711 0.” 0.531 0.97 0.8152 0.575 0.5 657 Gs m1m8bistri 1 0.551 0.551 1.0000 1.0000 0.551 0.5 0.581 0.815 0.53 658M101“ amylx W 0.912 0.1300 0.- 0.913 0.9150 0.6157 0.5 0.592 1.0000 1.0000 659 Smituy SVcs, Stean8 0.%6 0.6157 0.6157 0.%6 1.115 0.6157 0.97 0.5 1.0000 1.!!!” 660 nautical Relatdlh 0.363 0.6955 0.655 0.363 0.7621 0.695 0.78 0.9903 0.597 0.731 6610ther1hOlsaleYruh 0RD? 0.6955 0.6955 0.35 0.0:05 0.0103 0.78 0.015 0.1159 0.7321 6628mtia‘al Relate! Re 0.963 0.597 0.5597 0.755 0.399 0.319 0.75 0.312 0.2167 0.963 650ther Retail Trah 0.955 0.5597 0.5597 0.955 0.95% 0.597 0.75 0.955 0.9778 0.%3 666 Miro 0.507 0.551 0.551 0.507 0.562 0.6065 0.5 0.531 0.351 0.551 65 Mt mica 0.83 0.5831 0.551 0.873 0.8757 0.6065 0.% 0.9931 0.715 0.551 “Surity8Camw1ty8r 0.6%.? 0.6% 0.6% 0.632 0.636 0.5368 0.76 0.531 0.6% 0.5831 6671mm Carriers 0.376 0.3900 0.3900 0.375 0.381 0.6” 0.39 0.55 1.0000 0.55 6681mm 8 811* 0.9209 0.6” 0.6” 0.9209 0.9209 0.6” 0.3 0.9950 0.6% 0.0390 “W5 Dialliru 1 0.6115 0.0000 1.0000 1.0000 0.6% 0.62 0.9159 0.9166 0.639 670 Rel Estate 0.837 0.6% 0.6115 0.“? 0.9.25 0.55 0.62 0.9159 0.763 0.659 671 Hotels8chbira Places 0.6813 0.115 0.115 0.6813 0.7138 0.632 0.11 0.713 0.8“ 0.73% 63 Lardyfilmiro 8 fine 0.56 0.6112 0.6112 0.56 0.936 0.6112 0.81 0.57 0.9202 0.8761 63 Farrel M8Crumtori 1 0.6112 0.6112 1.0000 1.0000 0.6112 0.92 0.57 1.0000 0.8761 676 Photo Studim8MiscPe 1 0.6112 0.6112 1.0000 1.0000 0.6112 0.93 0.%7 1.1!!!) 0.8761 675 Elxtrical Remir 91133 0.0537 0.6112 0.6112 0.537 0.361 0.6112 0.5 0.753 0.35 0.8761 676 Wtchfilockflaeltyfiu‘ 1 0.515 0.515 1.0000 1.0000 0.6112 0.51 0.37 1.0000 0.8761 677801878 W 91:13 1 0.615 0.615 1.0!” 1.“!!! 0.6112 0.61 0.%7 1.0000 0.8761 678 Misc anirgiqx 0.673 0 0.673 0.6773 0.327 0 0.92 0.555 0.3% 0.301 6795!: tomildirm 0.593 0 0.53 0.593 0.6% 0 0.67 0.7778 0.665 0.301 65Persa~d amym 0.577 0 0.577 0.377 0.935 0 0.93 0.550 1.0000 0.301 681 Capitalibatai’mOJZB 0 0.725 0.733 0.753 0 0.3 0.713 0.8601 0.301 63W8d2multim 1 0 0.35 0.811) 1.0000 0 0.5 0.9756 0.5278 0.301 650etective8Pnotective 0.7106 0 0.7106 0.7106 0.7681 0 0.5 0.5561 0.361 0.301 6868;11pile'1tal8 Lmsira 0.559 0 0.559 0.559 0.639 0 0.87 0.635 0.5137 0.301 65 Pintofinishim,0mnwc10.53 0 0.35 0.9582 1.0000 0 0.5 0.556 0.8186 0.301 6%Other 8131mm was 0.9905 0 0.35 0.” 1.0000 0 0.3 0.” 0.755 0.751 687kherfisim 0.87% 0 0.7500 0.87% 0.55 0 0.5 0.%6 1.0000 0.751 688 L091 9108 0.%76 0 0.735 0.%76 0.9001 0 0.3 0.” 0.570 0.751 matinerierchitectlr 0.923 0 0.731) 0.55 0.955 0 0.97 0.5270 0.520 0.751 65 Aoountim,hditim,880.865 0 0.35 0.865 1.0000 0 0.95 0.9379 1.0000 0.751 691 Eatim80r1rkim lee 1 0651 0.fl21 0.31) 1.!!!” 0.631 0.71 0.%3 1.0000 0.550 63 Mo Raital 8Lasim 1 0.356 0.356 0.911) 1.01!) 0.356 0.3 0.9912 0.55 0.577 693 Auto anir8$Vcs 0.%5 0.356 0.356 0.” 0.352 0.356 0.5 0.512 0.516 0.577 696 Auto Parkiru 8 Car 1&1 0.9376 0.356 0.356 0.9000 1.0000 0.356 0.67 0.%0 1.1!!!) 0.8877 695 lbtim Pictu'es 0.613 0.665 0.665 0.613 0.6669 0.338 0.66 0.615 0.651 0.“ 6% Dane Hollofitufios 8 5 0.737 0.615 0.615 0.757 0.737 0.755 0.61 0.7675 0.330 0.- 697 Theatriml PMJa 0.6018 0.655 0.655 0.6018 0.6617 0.5261 0.65 0.515 0.525 0.“ 658011100 Alleys 8 Pool 11 0.57 0.7W7 0.757 0.570 0.8370 0.757 0.97 0.8177 0.91% 0.- 201 Appendix C (cont'd.) seem. sacrum RPC: SIP rum: AL‘I’FLR BSTLK unm "1:09 1181110111? mm Latino 4” W181M,“ R 0.8766 0.- 0.- 0.8766 0.956 0.53 0.5 0.8741 0.566 0.‘ son Rmim8 1rd: auntie 0.812 0.4“!) 0.46m 0.812 0851 0.845 0.46 0.4679 0.653 0.833 501 Wipm 8 Rs 0.85 0.8290 0.8290 0.%5 0.” 0.%5 0.5 0.8751 0.9m 0.8383 5mm 8 Ra: Scam 0.616 0.3!) 0.5300 0.6!5 0.6!» 0.w15 0.3 0.6167 0.659 0.8583 SGSDxtorsluDatists 1 0.537 0.9437 1.0000 1.0000 1.0000 0.97 0.” 1.1!!!) 0.9437 504 Hmpitals 0.9963 0.8436 0.8436 0.9963 0.9953 1.1111) 0.5 0.8436 1.0000 0.”?7 55 MimlnPersall One 1 0.377 0.597! 1.0000 1.00110 0.5117 0.5 0.9912 1.0000 0.8977 SCEOtherMical & Held! 1 0.T761 0.7761 1.0000 1.0000 0.%46 0.5 0.525 0.9743 0.7761 507 Elcmmly85wa'rhry 1 0.71.00 0.7400 1.0000 1.1!!!) 1.1!!!) 0.74 0.524 1.1!!!) 0.7840 513 Collewsfihiversitiesfi 1 0.3100 0.3100 0.3100 1.0000 1.0000 0.31 0.524 1.0!” 0.7840 ”Otherfidntianl We: 1 0.46m 0.4611) 1.0000 1.0000 1.0000 0.46 0.584 1.0000 0.7840 510 Mimss Asociatias 0.&21 0.7395 0.755 0.3221 0.87.” 0.510 0.74 0.813 0.7.95 0.7818 S11Ldnr8Civicowiizet 0.516 0.7818 0.7818 0.9916 0.516 0.587 0.94 0.5% 0.9599 0.7818 512 Religias mimics 1 0.7818 0.7818 1.0000 1.1!!!) 0.587 0.5 0.92% 1.0!!) 0.7818 S130therflarmrship (Tami 0.55 0.5100 0.5100 0.595 0.524 0.” 0.51 0.8713 0.%18 0.7818 514 Resiauial Cane 1 0.5m 0.5m 1.0000 1.0000 1.0000 0.91 0.5729 1.0000 1.0000 515 Social Sumac 1 0.5m 0.5729 1.0000 1.1!!!) 1.0000 0.91 0.5729 0.9914 1.1111) 516 U.S. Postal Service 1 0.6475 0.6475 1.0000 1.1110 1.111!) 0.578 0.9a» 0.6475 517 Ferrel Electric Utilit 0 0 0 0 0 0 0 0 0.555 518 Other Fechral Gov Enter 0.551 0.51& 0.519 0.5921 0.5921 0.6285 0.513 0.6099 0.6475 519 Local Gov Pseuuer Tra 0 0 0 0 0 0 0 0 0.6501 530 State 8. Local Electric 0 0 0 0 0 0 0 0 0.%55 521 Other State 8. Local Gav 0.599 0.7704 0.7704 0.599 0.8766 0.945 0.519 0.fl04 1.0000 522 We Inports 0.11.04 0.1104 0.11114 0.0004 0.0004 0.01132 0.1115 0.1115 1.0000 58 Scrq: 0.54 0.254 0.254 0.540 0.540 0.549 0.%5 0.750 0.254 524 lbed 8. Secaflwd Goods 1 0.254 0.254 1.0000 1.0000 1.0000 0.9185 1.1!!!) 0.254 55 Guermmt lrdstty 1 0.5730 0.5730 1.0000 1.01!) 0.9803 0.9506 0.5730 1.0000 5% Rest of the world mm 0 0 0 0 0 0 0.9414 0 1.0000 527 mm Irdstry 1 1.0000 1.0000 1.1!!!) 1.0000 1.0000 0.9975 0 1.0000 5% Inventory Valtatim Adj 1 0 0 0 0 0 0 0 1.01!) AVERAE RPC M58501”: 0.6467 0.1699 0.%6 0.4435 0.4362 0.301 0.x 0.382 0.%6 0.46% SIMON!) WHICH: 0.3572 0.2240 0.232 0.3150 0.357 0.3219 0.8 0.3621 0.58 0.24% APPENDIX D: IMPLAN AGGREGATION AND SECTORIZATION SCHEMES The four aggregation and sectorization schemes used for Lake State models are listed below. Their names and the number of sectors they contain (in parentheses) are as follows: UNAGMN (308), AGMN (158), A630 (31), A616 (16). The aggregation schemes are presented first. They appear in the exact format by which they were used as aggregation templates with 1MPLAM; therefore, only aggregated sectors are shown. two lines for each aggregated sector. On the first line, the first number is the lowest IMPLAN sector number of all the 1MPLAN sectors that are to be aggregated together to form a new aggregated sector. The next number on the same line indicates how many other sectors are to be aggregated to form the new sector. The name for the new aggregated aggregated sector completes the data on the first line. The next line below these data lists the remaining iMPLAM sector numbers for the other sectors which comprise the new aggregated sector. As an example, the first two lines for the UMAGMN models aggregation scheme indicate sector 3 is being aggregated with six other sectors (4 through 9) to form a new sector called "Meat Animals 8 Misc. Livestock." The first line for the AGMN aggregation scheme indicates 326 sectors are being aggregated with sector 1 to form a sector called “All Other.“ The sectorization schemes list all model sectors which result from the aggregation schemes. Aggregated sectors are indicated by the abbreviation "AGG;" unaggregated sectors are indicated by all capital letters. Appendix A lists all 528 IMPLAN sectors. Aggregation Scheme for UNAGMN models 3 6Meat Animals 8 Misc. Livestock 4 5 6 7 8 9 11 3Food,Feed Grains 8 Grass Seed 12 13 14 16 1Fruits & Tree Nuts 17 18 2Vegs, Sugar & Misc. Crops 19 20 22 1Forest, Grnhs & Nursery Prdcts 23 24 1Agri, Forstry, 8 Fish Prdcts 25 26 1Agri, For, & Fish Svcs 27 28 11ron 8 Ferroalloy Ore Mining 29 31 7Nonferrous Ore Mining,exc copper 32 33 34 35 36 37 38 39 icoal Mining 40 44 50im,Crshd,Constr Stone 8 1nd Sand 45 46 47 48 49 SO 15Misc. Nonmetallic Minerals, nec 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 ' 66 6New Construction 67 68 69 70 71 72 73 2Maintenance & Repair Construction 74 75 76 4Militay Ordnance 77 78 406 407 97 1Frozen Fruits, Juices,& Vegs 98 127 3Tobacco 128 129 130 160 13wood Prdcts,exc furnitureSpaper 161 162 163 164 165 166 167 168 169 170 171 172 173 180 40ffice Furniture 181 182 183 184 225 3Plastics 8 Synthetic Materials 226 227 228 235 2Petro Refining 8 Misc. Petrc Prdct 236 237 257 22Stone & Clay Prdcts 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 8Primary iron 8 Steel Manufacturing 281 282 283 284 285 286 287 288 289 13Primary Nonferrous Metals Manuf 290 291 292 293 294 295 296 297 298 299 300 301 302 303 1Metal Cans 304 305 9Heating,Plumbing,&Fab.Metal Prdcts 306 307 308 309 310 311 312 313 314 322 70ther Fabricated Metal Prdcts 323 324 325 326 327 328 329 330 1Engines & Turbines 202 The format involves 2([3 Appendix D (cont'd.) Aggregation Scheme for UHAGMH models (cont'd.) 331 334 2Construction 8 Mining Machinery 464 28anking 8 Other Finance institut 335 336 465 466 337 3Material Handling Mach 8 Equip 469 1Real Estate 338 339 340 470 353 86en indstries Machinery 8 Equip 472 40ther Personal 8 Repair Svcs 354 355 356 357 358 359 360 361 473 474 475 476 362 30ffice Computing8Acctng Machines 478 80ther Business Svcs 363 364 365 479 480 481 482 483 484 485 486 366 48ervice industry Machinery 488 ZMisc. Professional Svcs 367 368 369 370 489 490 389 3Radio,Tv & Communications Equip 492 2Auto Repair 8 Svcs 390 391 392 493 494 434 3Marking Devices 496 6Amusement 8 Rec Svcs, nec 435 436 437 497 498 499 500 501 502 452 1Freight Forwrdrs & Other Transp 505 10ther Medical 8 Health Svcs 453 506 458 1water Supply 8 Sanitary Svcs 507 2Educational Svcs 459 508 509 460 10ther wholesale Trade 510 3Honprofit Organizations 461 511 512 513 462 10ther Retail Trade 523 1Scrap, Used 8 Secondhand Goods 463 524 Aggregation Scheme for AGMH models 1326ALL OTHER 350 351 352 353 354 355 356 357 358 359 3 4 5 6 7 8 9 10 15 28 360 361 362 363 364 365 366 367 368 369 29 3O 31 32 33 34 35 36 37 38 370 371 373 375 376 377 379 380 381 383 39 40 41 42 43 44 45 46 47 48 384 385 388 389 390 391 392 393 394 395 49 50 51 52 53 54 55 56 57 58 397 398 399 401 402 403 404 405 406 407 60 61 62 63 64 65 66 67 68 69 408 409 410 411 412 413 416 418 421 424 70 71 72 73 74 75 76 77 78 79 428 430 434 435 436 437 442 443 444 452 81 113 118 120 121 127 128 129 130 135 453 455 456 457 458 459 464 465 466 467 136 138 139 140 141 143 144 147 148 149 469 470 487 488 489 490 503 504 505 506 154 155 160 161 162 163 164 165 166 167 516 517 518 519 520 522 168 169 170 171 172 173 176 180 181 182 11 3Food,Feed Crainslcrass Seeds 183 184 185 187 188 189 190 192 193 194 12 13 14 195 196 197 199 203 206 207 208 209 210 16 1Fruits 8 Tree Huts 211 212 213 214 216 217 218 219 220 221 17 222 223 225 226 227 228 231 232 234 238 18 2Veggies,Sugar & Misc. Crops 239 240 241 242 246 247 248 249 256 257 19 20 258 259 260 261 262 263 264 265 266 267 22 1Forest,Grnhs 8 Nursery Prdcts 268 269 270 271 272 273 274 275 276 277 23 278 279 280 281 283 284 285 286 287 289 24 1Forestry 8 Fishery Prdcts 290 291 292 293 294 295 296 297 298 299 25 300 301 302 303 304 305 306 307 308 309 26 1Agri, Forestry 8 Fish Svcs 310 311 312 313 314 316 317 318 322 323 27 324 325 326 327 328 329 330 331 334 335 97 1Froz Fruits, Juices 8 Vegs 336 337 338 339 340 342 343 345 347 348 98 204 Appendix D (cont'd.) Aggregation Scheme for AGMH models (cont'd.) 235 2Petro Refining 6 Misc Petro Prdcts 492 2Auto Repair 6 Services 236 237 493 494 460 10ther wholesale Trade 496 6Amusement 6 Rec Svcs, nec 461 497 498 499 500 501 502 462 10ther Retail Trade 507 2Educational Svcs 463 508 509 472 4Pers 6 Repair Svcs,exc auto6beauty 510 3Honprofit Organizations 473 474 475 476 511 512 513 478 80ther Business Services 523 1Scap, Used 6 Secondhand Goods 479 480 481 482 483 484 485 486 524 Aggregation Scheme for A630 models 1 22Farm Products 200 14Printing 6 Publishing 2 3 4 5 6 7 8 9 10 11 201 202 203 204 205 206 207 208 209 210 12 13 14 15 16 17 18 19 20 21 211 212 213 214 22 23 215 19Chemicals 24 3Agri, For 6 Fish Prdcts 216 217 218 219 220 221 222 223 224 225 25 26 27 226 227 228 229 230 231 232 233 234 28 37Mining 235 4Petroleum Production 29 30 31 32 33 34 35 36 37 38 236 237 238 239 39 40 41 42 43 44 45 46 47 48 240 14Rubber,Leather6Misc Plastics 49 50 S1 52 S3 54 55 56 57 58 241 242 243 244 245 246 247 248 249 250 59 60 61 62 63 64 65 251 252 253 254 66 9Construction 255 24Stone,Clay66lass Prdcts 67 68 69 70 71 72 73 74 75 256 257 258 259 260 261 262 263 264 265 76 110ther Transport Vehicles 266 267 268 269 270 271 272 273 274 275 78 405 406 407 408 409 410 411 412 414 276 277 278 279 415 280 21Primary Metals Manufacturing 77 31Fabricated Metals Manufacturing 281 282 283 284 285 286 287 288 289 290 79 80 81 302 303 304 305 306 307 308 291 292 293 294 295 296 297 298 299 300 309 310 311 312 313 314 315 316 317 318 301 319 320 321 322 323 324 325 326 327 328 330 40Machinery 6 Equipment 329 331 332 333 334 335 336 337 338 339 340 82 48Food 6 Kindred Prdcts 341 342 343 344 345 346 347 348 349 350 83 84 85 86 87 88 89 90 91 92 351 352 353 354 355 356 357 358 359 360 93 94 95 96 97 98 99 100 101 102 361 362 363 364 365 366 367 368 369 370 103 104 105 106 107 108 109 110 111 112 371 30lnstruments6Misc Manufacturing 113 114 115 116 117 118 119 120 121 122 416 417 418 419 420 421 422 423 424 425 123 124 125 126 127 128 129 130 426 427 428 429 430 431 432 433 434 435 131 28Textiles 6 Apparel 436 437 438 439 440 441 442 443 444 445 132 133 134 135 136 137 138 139 140 141 372 285lectric6Electronic Equipment 142 143 144 145 146 147 148 149 150 151 . 373 374 375 376 377 378 379 380 381 382 152 153 154 155 156 157 158 159 383 384 385 386 387 388 389 390 391 392 160 40Forest Products 393 394 395 396 397 398 399 400 161 162 163 164 165 166 167 168 169 170 401 3Motor Vehicles 171 172 173 174 175 176 177 178 179 180 402 403 404 181 182 183 184 185 186 187 188 189 190 446 7Transportation Services 191 192 193 194 195 196 197 198 199 413 447 448 449 450 451 452 453 Appendix D (cont'd.) 454 455 205 Aggregation Scheme for A630 models (cont'd.) 1Communications 456 3Public Utilities 457 458 459 460 461 462 463 1Hholesale Trade 1Retail Trade 464 6Finance,insur,Real Estate 465 466 467 468 469 470 1 26Farm Products 2 12 3 13 4 14 22 23 24 28 37Mining 29 30 31 39 4O 41 49 50 51 S9 60 61 66 9Construction 67 68 69 70 71 762860ther Manufacturing 77 78 79 80 81 165 175 185 195 205 215 225 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 166 176 186 196 206 216 226 241 251 261 271 281 291 301 311 321 331 341 351 361 371 381 391 167 177 187 197 207 217 227 242 252 262 272 282 292 302 312 322 332 342 352 362 372 382 392 5 15 25 32 42 52 62 168 178 188 198 208 218 228 243 253 263 273 283 293 303 313 323 333 343 353 363 373 383 393 6 16 26 33 43 53 63 169 179 189 199 209 219 229 244 254 264 274 284 294 304 314 324 334 344 354 364 374 384 394 Aggregation 7 8 9 17 18 19 27 35 45 55 65 36 46 56 34 44 54 64 72 74 162 172 182 192 202 212 222 232 247 257 267 277 287 297 307 317 327 337 347 357 367 377 387 397 161 171 181 191 201 211 221 231 246 256 266 276 286 296 306 316 326 336 346 356 366 376 386 396 160 170 180 190 200 210 220 230 245 255 265 275 285 295 305 315 325 335 345 355 365 375 385 395 472 35Misc. Services, nec 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 492 493 494 495 503 504 505 506 507 508 509 510 511 512 513 514 515 496 6Amusement 6 Rec Services 497 498 499 500 501 502 516 5Federal 6 State Gov. 517 518 519 520 521 522 60ther 60v,Hshld, world industry 523 524 525 526 527 528 Scheme for A616 models 10 20 37 47 57 75 163 173 183 193 203 213 223 233 248 258 268 278 288 298 308 318 328 338 348 358 368 378 388 398 11 21 38 48 58 164 174 184 194 204 214 224 234 249 259 269 279 289 299 309 319 329 339 349 359 369 379 389 399 407 417 427 437 403 404 405 406 413 414 415 416 423 424 425 426 433 434 435 436 443 444 445 Kindred Prdcts 86 87 88 89 96 97 98 99 103 104 105 106 107 108 109 113 114 115 116 117 118 119 123 124 125 126 127 128 129 400 401 402 410 411 412 420 421 422 430 431 432 440 441 442 82 48Food 6 83 84 85 93 94 95 90 100 110 120 130 131 28Textiles 6 Apparel 138 148 158 139 149 159 132 133 134 135 136 137 142 143 144 145 146 147 152 153 154 155 156 157 235 4Petroleum Production 236 237 238 239 446 7Transportation Services 447 448 449 450 451 452 453 454 480ther Services 460 462 455 456 457 458 459 464 465 466 469 470 472 473 474 475 476 477 480 481 482 483 484 485 486 487 490 492 493 494 495 503 504 505 508 509 510 511 512 513 514 515 1wholesale Trade 461 1Retail Trade 463 496 6Amusement 6 Rec Services 497 498 499 500 501 502 516 5Federal 6 State Gov. 517 518 519 520 521 408 418 428 438 91 101 111 121 140 150 467 478 488 506 409 419 429 439 92 102 112 122 141 151 468 479 489 507 522 60ther 66v,Hshld, world industry 523 524 525 526 527 528 Appendix D (cont'd.) auto—s 16 18 21 22 24 26 28 30 40 41 42 43 44 50 66 76 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 99 100 101 102 103 104 105 106 107 206 UNAGMN 308 Sectorization Scheme DAIRY FARM PRODUCTS POULTRY AND EGGS A66 Meat Animals 6 Misc. Livestock A66 Food,Feed Grains 6 Grass Seed TOBACCO FRUITS AGG Vegs, Sugar 6 Misc. Crops OIL BEARING CROPS A66 Forest, 6rnhs 6 Nursery Prdcts A66 Agri, Forstry, 6 Fish Prdcts A66 Agri, For, 6 Fish Svcs iRON ORES A66 Nonferrous Ore Mining,exc copper BITUMINOUS AND LIGNITE MINING, SERVI NATURAL GAS CRUDE PETROLEUM NATURAL GAS LIOUIDS AGG Dim,Crshd,Constr Stone 6 Ind Sand AGG Misc. Nonmetallic Minerals, nec AGG New Construction AGG Maintenance 6 Repair Construction AGG Militay Ordnance SMALL ARMS SMALL ARMS AMMUNITION OTHER ORDNANCE AND ACCESSORIES MEAT PACKING PLANTS SAUSAGES AND OTHER PREPARED MEATS POULTRY DRESSING PLANTS POULTRY AND E66 PROCESSING CREAMERY BUTTER CHEESE, NATURAL AND PROCESSED CONDENSED AND EVAPORATED MILK ICE CREAM AND FROZEN DESSERTS FLUID MILK CANNED AND CURED SEA FOODS CANNED SPECIALTIES CANNED FRUITS AND VEGETABLES DEHYDRATED FOOD PRODUCTS PICKLES, SAUCES, AND SALAD DRESSINGS FRESH OR FROZEN PACKAGED FISH A66 Frozen Fruits, Juices,6 Vegs FLOUR AND OTHER GRAIN MILL PRODUCTS CEREAL PREPARATIONS BLENDED AND PREPARED FLOUR DOG, CAT, AND OTHER PET FOOD PREPARED FEEDS, N.E.C RICE MILLING NET CORN MILLING BREAD, CAKE, AND RELATED PRODUCTS COOKIES AND CRACKERS 108 109 110 111 112 113 114 115 116 117 119 120 122 123 124 125 126 128 131 132 133 135 142 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 174 175 176 177 178 179 180 185 186 187 188 SUGAR CONFECTIONERY PRODUCTS CHOCOLATE AND COCOA PRODUCTS CHENING GUM MALT LIOUORS MALT NINES, BRANDY, AND BRANDY SPIRITS DISTILLED LIQUOR, EXCEPT BRANDY BOTTLED AND CANNED SOFT DRINKS FLAVORING EXTRACTS AND SYRUPS, N.E.C. SOYBEAN OIL MILLS A66 Other Fats, Oils, 6 Oil Hills ROASTED COFFEE SHORTENING AND COOKING OILS MANUFACTURED ICE MACARONI AND SPAGHETTI FOOD PREPARATIONS, N.E.C AGG Tobacco BROADUOVEN FABRIC MILLS AND FINISHING NARROU FABRIC MILLS YARN MILLS AND FINISHING OF TEXTILES A66 Other Misc. Textl Gds6Flr Cvrngs CORDAGE AND TNINE NOMENS HOSIERY, EXCEPT SOCKS HOSIERY, N.E.C KNIT OUTERNEAR MILLS KNIT UNDERNEAR MILLS KNITTING MILLS, N.E.C KNIT FABRIC MILLS APPAREL MADE FROM PURCHASED MATERIAL CURTAINS AND DRAPERIES HOUSEFURNISHINGS, N.E.C TEXTILE BAGS CANVAS PRODUCTS PLEATING AND STITCHING AUTOMOTIVE AND APPAREL TRIMMINGS SCHIFFI MACHINE EMBROIDERIES FABRICATED TEXTILE PRODUCTS, N.E.C A66 wood Prdcts,exc furniture6paper HOOD HOUSEHOLD FURNITURE HOUSEHOLD FURNITURE, N.E.C HOOD TV AND RADIO CABINETS UPNOLSTERED HOUSEHOLD FURNITURE METAL HOUSEHOLD FURNITURE MATTRESSES AND BEDSPRINGS A66 Office Furniture BLINDS, SHADES, AND DRAPERY HARDNARE FURNITURE AND FIXTURES, N.E.C PULP MILLS PAPER MILLS, EXCEPT BUILDING PAPER Appendix D (cont'd.) 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 229 230 231 232 233 234 235 238 239 240 241 242 243 207 UNAGMN 308 Sectorization Scheme (cont'd.) PAPERBOARD MILLS ENVELOPES SANITARY PAPER PRODUCTS BUILDING PAPER AND BOARD MILLS PAPER COATING AND GLAZING BAGS, EXCEPT TEXTILE DIE-CUT PAPER AND BOARD PRESSED AND MOLDED PULP GOODS STATIONERY PRODUCTS CONVERTED PAPER PRODUCTS, N.E.C PAPERBOARD CONTAINERS AND BOXES NENSPAPERS PERIODICALS BOOK PUBLISHING BOOK PRINTING MISCELLANEOUS PUBLISHING COMMERCIAL PRINTING LITHOGRAPHIC PLATEMAKING AND SERVICE MANIFOLD BUSINESS FORMS BLANKBOOKS AND LOOSELEAF BINDERS GREETING CARD PUBLISHING ENGRAVING AND PLATE PRINTING BOOKBINDING AND RELATED NORK TYPESETTING PHOTOENGRAVING ELECTROTYPING AND STEREOTYPING INDUSTRIAL INORGANIC, ORGANIC CHEMIC NITROGENOUS AND PHOSPHATIC FERTILIZE FERTILIZERS, MIXING ONLY AGRICULTURAL CHEMICALS, N.E.C GUM AND NOOD CHEMICALS ADHESIVES AND SEALANTS EXPLOSIVES PRINTING INK CARBON BLACK CHEMICAL PREPARATIONS, N.E.C AGG Plastics 6 Synthetic Materials DRUGS SOAP AND OTHER DETERGENTS POLISHES AND SANITATION GOODS SURFACE ACTIVE AGENTS TOILET PREPARATIONS PAINTS AND ALLIED PRODUCTS AGG Petro Refining 6 Misc. Petro Prd PAVING MIXTURES AND BLOCKS ASPHALT FELTS AND COATINGS TIRES AND INNER TUBES RUBBER AND PLASTICS FOOTNEAR RECLAIMED RUBBER FABRICATED RUBBER PRODUCTS, N.E.C 244 245 246 247 248 249 250 251 252 253 254 255 256 257 280 289 303 305 315 316 317 318 319 320 321 322 330 332 333 334 337 341 342 343 344 345 346 347 348 349 350 351 352 353 362 366 . 371 372 373 374 MISCELLANEOUS PLASTICS PRODUCTS RUBBER AND PLASTICS HOSE AND BELTING LEATHER TANNING AND FINISHING FOOTNEAR CUT STOCK SHOES, EXCEPT RUBBER HOUSE SLIPPERS LEATHER GLOVES AND MITTENS LUGGAGE NOMENS HANDBAGS AND PURSES PERSONAL LEATHER GOODS LEATHER GOODS, N.E.C GLASS AND GLASS PRODUCTS, EXC CONTAI GLASS CONTAINERS AGG Stone 6 Clay Prdcts A66 Primary iron 6 Steel Manufacturng A66 Primary Nonferrous Metals Manuf A66 Metal Cans A66 Heating,Plumbing,6Fab.Metal Prdc SCREN MACHINE PRODUCTS AND BOLTS, ET AUTOMOTIVE STAMPINGS CRONNS AND CLOSURES METAL STAMPINGS, N.E.C. CUTLERY HAND AND EDGE TOOLS, N.E.C. HAND SANS AND SAN BLADES A66 Other Fabricated Metal Prdcts A66 Engines 6 Turbines FARM MACHINERY AND EQUIPMENT LANN AND GARDEN EQUIPMENT AGG Construction 6 Mining Machinery A66 Material Handling Mach 6 Equip MACHINE TOOLS, METAL CUTTING TYPES MACHINE TOOLS, METAL FORMING TYPES SPECIAL DIES AND TOOLS AND ACCESSORI PONER DRIVEN HAND TOOLS ROLLING MILL MACHINERY METALNORKING MACHINERY, N.E.C. FOOD PRODUCTS MACHINERY TEXTILE MACHINERY NOODNORKING MACHINERY PAPER INDUSTRIES MACHINERY PRINTING TRADES MACHINERY SPECIAL INDUSTRY MACHINERY, N.E.C. AGG Gen Indstries Machinery 6 Equip A66 Office Computing6Acctng Machines A66 Service industry Machinery INSTRUMENTS TO MEASURE ELECTRICITY TRANSFORMERS SNITCHGEAR AND SNITCHBOARD APPARATUS MOTORS AND GENERATORS Appendix D (cont'd.) 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 393 394 395 396 397 398 399 400 401 402 403 404 405 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 208 UNAGMN 308 Sectorization Scheme (cont.'d) INDUSTRIAL CONTROLS NELDING APPARATUS, ELECTRIC CARBON AND GRAPHITE PRODUCTS ELECTRICAL INDUSTRIAL APPARATUS, N.E. HOUSEHOLD COOKING EQUIPMENT HOUSEHOLD REFRIGERATORS AND FREEZERS HOUSEHOLD LAUNDRY EQUIPMENT ELECTRIC HOUSENARES AND FANS HOUSEHOLD VACUUM CLEANERS SENING MACHINES HOUSEHOLD APPLIANCES, N.E.C. ELECTRIC LAMPS LIGHTING FIXTURES AND EQUIPMENT NIRING DEVICES AGG Radio,Tv 6 Communications Equip ELECTRON TUBES SEMICONDUCTORS AND RELATED DEVICES ELECTRONIC COMPONENTS, N.E.C. STORAGE BATTERIES PRIMARY BATTERIES, DRY AND NET X-RAY APPARATUS AND TUBES ENGINE ELECTRICAL EQUIPMENT ELECTRICAL EQUIPMENT, N.E.C. TRUCK AND BUS BODIES TRUCK TRAILERS MOTOR VEHICLES MOTOR VEHICLE PARTS AND ACCESSORIES AIRCRAFT SHIP BUILDING AND REPAIRING BOAT BUILDING AND REPAIRING RAILROAD EQUIPMENT MOTORCYCLES, BICYCLES, AND PARTS TRAVEL TRAILERS AND CAMPERS MOBILE HOMES MOTOR HOMES TRANSPORTATION EQUIPMENT, N.E.C. ENGINEERING AND SCIENTIFIC INSTRUMEN MECHANICAL MEASURING DEVICES AUTOMATIC TEMPERATURE CONTROLS SURGICAL AND MEDICAL iNSTRUMENTS SURGICAL APPLIANCES AND SUPPLIES DENTAL EQUIPMENT AND SUPPLIES NATCHES, CLOCKS, AND PARTS OPTICAL INSTRUMENTS AND LENSES OPHTHALMIC GOODS ' PHOTOGRAPHIC EQUIPMENT AND SUPPLIES JENELRY, PRECIOUS METAL JENELERS MATERIALS AND LAPIDARY NORK SILVERNARE AND PLATED NARE COSTUME JENELERY 430 431 432 433 434 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 454 455 456 457 458 460 462 464 467 468 469 471 472 477 478 487 488 491 492 495 496 503 504 505 507 510 514 515 516 517 C. MUSICAL INSTRUMENTS GAMES, TOYS, AND CNILDRENS VEHICLES DOLLS SPORTING AND ATHLETIC GOODS, N.E.C. AGG Marking Devices ARTIFICIAL TREES AND FLONERS BUTTONS NEEDLES, PINS, AND FASTENERS BROOMS AND BRUSHES HARD SURFACE FLOOR COVERINGS BURIAL GASKETS AND VAULTS SIGNS AND ADVERTISING DISPLAYS MANUFACTURING INDUSTRIES, N.E.C. RAILROADS AND RELATED SERVICES LOCAL, INTERURBAN PASSENGER TRANSIT MOTOR FREIGHT TRANSPORT AND NAREHOUS NATER TRANSPORTATION AIR TRANSPORTATION PIPE LINES, EXCEPT NATURAL GAS AGG Freight Forwrdrs 6 Other Transp COMMUNICATIONS, EXCEPT RADIO AND TV RADIO AND TV BROADCASTING ELECTRIC SERVICES GAS PRODUCTION AND DISTRIBUTION AGG Nater Supply 6 Sanitary Svcs AGG Other Nholesale Trade AGG Other Retail Trade AGG Banking 6 Other Finance Institut iNSURANCE CARRIERS INSURANCE AGENTS AND BROKERS AGG Real Estate HOTELS AND LODGING PLACES AGG Other Personal 6 Repair Svcs BEAUTY AND BARBER SHOPS AGG Other Business Svcs ADVERTISING AGG Misc. Professional Svcs EATING AND DRINKING PLACES AGG Auto Repair 6 Svcs MOTION PICTURES AGG Amusement 6 Rec Svcs, nec DOCTORS AND DENTISTS HOSPITALS AGG Other Medical 6 Health Svcs AGG Educational Svcs AGG Nonprofit Organizations RESIDENTIAL CARE SOCIAL SERVICES, N.E.C. U.S. POSTAL SERVICE FEDERAL ELECTRIC UTILITIES Appendix D (cont'd.) 518 519 520 521 N-I 16 18 21 22 24 26 80 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 99 100 101 102 103 104 105 106 107 108 109 110 111 112 114 115 209 UNAGMN 308 Sectorization Scheme (cont.'d) OTHER FEDERAL GOVERNMENT ENTERPRISES LOCAL GOVERNMENT PASSENGER TRANSIT STATE AND LOCAL ELECTRIC UTILITIES OTHER STATE AND LOCAL GOVT ENTERPRIS 525 526 527 528 GOVERNMENT INDUSTRY REST OF THE NORLD INDUSTRY HOUSEHOLD INDUSTRY INVENTORY VALUATION ADJUSTMENT AGMN 158 Sectorization Scheme AGG ALL OTHER POULTRY AND EGGS AGG Food,Feed Grains6Grass Seeds FRUITS AGG Veggies,Sugar 6 Misc. Crops OIL BEARING CROPS AGG Forest,Grnhs 6 Nursery Prdcts AGG Forestry 6 Fishery Prdcts AGG Agri, Forestry 6 Fish Svcs SMALL ARMS AMMUNITION MEAT PACKING PLANTS SAUSAGES AND OTHER PREPARED MEATS POULTRY DRESSING PLANTS POULTRY AND EGG PROCESSING CREAMERY BUTTER CHEESE, NATURAL AND PROCESSED CONDENSED AND EVAPORATED MILK ICE CREAM AND FROZEN DESSERTS FLUID MILK CANNED AND CURED SEA FOODS CANNED SPECIALTIES CANNED FRUITS AND VEGETABLES DEHYDRATED FOOD PRODUCTS PICKLES, SAUCES, AND SALAD DRESSINGS FRESH OR FROZEN PACKAGED FISH AGG Froz Fruits, Juices 6 Vegs FLOUR AND OTHER GRAIN MILL PRODUCTS CEREAL PREPARATIONS BLENDED AND PREPARED FLOUR DOG, CAT, AND OTHER PET FOOD PREPARED FEEDS, N.E.C RICE MILLING NET CORN MILLING BREAD, CAKE, AND RELATED PRODUCTS COOKIES AND CRACKERS SUGAR CONFECTIONERY PRODUCTS CHOCOLATE AND COCOA PRODUCTS CHENING GUM MALT LIQUORS NINES, BRANDY, AND BRANDY SPIRITS DISTILLED LIQUOR, EXCEPT BRANDY 116 117 119 122 123 124 125 126 131 132 133 137 142 145 146 150 151 152 153 156 157 158 159 174 175 177 178 179 186 191 198 200 201 202 204 205 215 224 229 230 233 235 BOTTLED AND CANNED SOFT DRINKS FLAVORING EXTRACTS AND SYRUPS, N.E.C SOYBEAN OIL MILLS ROASTED COFFEE SHORTENING AND COOKING OILS MANUFACTURED ICE MACARONI AND SPAGHETTI FOOD PREPARATIONS, N.E.C BROADNOVEN FABRIC MILLS AND FINISHIN NARRON FABRIC MILLS YARN MILLS AND FINISHING OF TEXTILES LACE GOODS CORDAGE AND TNiNE NOMENS HOSIERY, EXCEPT SOCKS HOSIERY, N.E.C KNIT FABRIC MILLS APPAREL MADE FROM PURCHASED MATERIAL CURTAINS AND DRAPERIES HOUSEFURNISHINGS, N.E.C PLEATING AND STITCHING AUTOMOTIVE AND APPAREL TRIMNINGS SCHIFFI MACHINE EMBROIDERIES FABRICATED TEXTILE PRODUCTS, N.E.C NOOD HOUSEHOLD FURNITURE HOUSEHOLD FURNITURE, N.E.C UPHOLSTERED HOUSEHOLD FURNITURE METAL HOUSEHOLD FURNITURE MATTRESSES AND BEDSPRINGS FURNITURE AND FIXTURES, N.E.C SANITARY PAPER PRODUCTS CONVERTED PAPER PRODUCTS, N.E.C NENSPAPERS PERIODICALS BOOK PUBLISHING MISCELLANEOUS PUBLISHING COMMERCIAL PRINTING INDUSTRIAL INORGANIC, ORGANIC CHEMIC CHEMICAL PREPARATIONS, N.E.C DRUGS SOAP AND OTHER DETERGENTS TOILET PREPARATIONS AGG Petro Refining 6 Misc Petro Prdc Appendix D (cont'd.) 243 244 245 250 251 252 253 254 255 282 288 315 319 320 321 332 333 341 344 346 349 372 374 378 382 386 387 396 400 414 415 417 419 420 422 423 425 426 210 AGMN 158 Sectorization Scheme (cont.'d) FABRICATED RUBBER PRODUCTS, N.E.C MISCELLANEOUS PLASTICS PRODUCTS RUBBER AND PLASTICS HOSE AND BELTING LEATHER GLOVES AND MITTENS LUGGAGE NOMENS HANDBAGS AND PURSES PERSONAL LEATHER GOODS LEATHER GOODS, N.E.C GLASS AND GLASS PRODUCTS, EXC CONTAI STEEL NIRE AND RELATED PRODUCTS PRIMARY METAL PRODUCTS, N.E.C SCREN MACHINE PRODUCTS AND BOLTS, ET CUTLERY HAND AND EDGE TOOLS, N.E.C. HAND SANS AND SAN BLADES FARM MACHINERY AND EQUIPMENT LANN AND GARDEN EQUIPMENT MACHINE TOOLS, METAL CUTTING TYPES PONER DRIVEN HAND TOOLS METALNORKING MACHINERY, N.E.C. NOODNORKING MACHINERY TRANSFORMERS MOTORS AND GENERATORS ELECTRICAL INDUSTRIAL APPARATUS, N.E ELECTRIC HOUSENARES AND FANS ELECTRIC LAMPS LIGHTING FIXTURES AND EQUIPMENT STORAGE BATTERIES ELECTRICAL EQUIPMENT, N.E.C. MOTOR HOMES TRANSPORTATION EQUIPMENT, N.E.C. MECHANICAL MEASURING DEVICES SURGICAL AND MEDICAL INSTRUMENTS SURGICAL APPLIANCES AND SUPPLIES NATCHES, CLOCKS, AND PARTS OPTICAL INSTRUMENTS AND LENSES PHOTOGRAPHIC EQUIPMENT AND SUPPLIES JENELRY, PRECIOUS METAL 427 429 431 432 433 438 439 440 441 445 446 447 448 449 450 451 454 460 462 468 471 472 477 478 491 492 495 496 507 510 514 515 521 525 526 527 528 JENELERS MATERIALS AND LAPIDARY NORK COSTUME JENELERY GAMES, TOYS, AND CNILDRENS VEHICLES DOLLS SPORTING AND ATHLETIC GOODS, N.E.C. ARTIFICIAL TREES AND FLONERS BUTTONS NEEDLES, PINS, AND FASTENERS BROOMS AND BRUSHES MANUFACTURING INDUSTRIES, N.E.C. RAILROADS AND RELATED SERVICES LOCAL, iNTERURBAN PASSENGER TRANSIT MOTOR FREIGHT TRANSPORT AND NAREHOUS NATER TRANSPORTATION AIR TRANSPORTATION PIPE LINES, EXCEPT NATURAL GAS COMMUNICATIONS, EXCEPT RADIO AND TV AGG Other Nholesale Trade AGG Other Retail Trade INSURANCE AGENTS AND BROKERS HOTELS AND LODGING PLACES AGG Pers 6 Repair Svcs,exc auto6beau BEAUTY AND BARBER SHOPS AGG Other Business Services EATING AND DRINKING PLACES AGG Auto Repair 6 Services MOTION PICTURES AGG Amusement 6 Rec Svcs, nec AGG Educational Svcs AGG Nonprofit Organizations RESIDENTIAL CARE SOCIAL SERVICES, N.E.C. OTHER STATE AND LOCAL GOVT ENTERPRIS GOVERNMENT INDUSTRY REST OF THE NORLD INDUSTRY HOUSEHOLD INDUSTRY INVENTORY VALUATION ADJUSTMENT Appendix D (cont'd.) 24 28 66 76 77 82 131 160 200 215 235 240 255 280 330 21]. A630 Sectorization Scheme AGG Farm Products AGG Agri, For 6 Fish Prdcts AGG Mining AGG Construction AGG Other Transport Vehicles AGG Fabricated Metals Manufacturing AGG Food 6 Kindred Prdcts AGG Textiles 6 Apparel AGG Forest Products AGG Printing 6 Publishing AGG Chemicals A66 Petroleum Production AGG Rubber,Leather6Misc Plastics AGG Stone,Clay6Glass Prdcts AGG Primary Metals Manufacturing AGG Machinery 6 Equipment 1 AGG 28 AGG 66 AGG 77 AGG 82 AGG 131 AGG 235 AGG 446 AGG . 371 372 401 446 454 456 460 462 464 471 472 491 496 516 525 AGG AGG AGG AGG AGG AGG AGG AGG AGG Instruments6Misc Manufacturing Electric6Electronic Equipment Motor Vehicles Transportation Services Communications Public Utilities Nholesale Trade Retail Trade Finance,Insur,Real Estate HOTELS AND LODGING PLACES AGG Misc. Services, nec EATING AND DRINKING PLACES AGG AGG AGG AG 16 Sectorization Scheme Farm Products Mining Construction Other Manufacturing Food 6 Kindred Prdcts Textiles 6 Apparel Petroleum Production Transportation Services Amusement 6 Rec Services Federal 6 State Gov. Other Gov,Hshld, Norld Industry 454 460 462 471 491 496 516 525 A66 Other Services AGG Nholesale Trade AGG Retail Trade HOTELS AND LODGING PLACES EATING AND DRINKING PLACES AGG Amusement 6 Rec Services AGG Federal 6 State Gov. AGG Other Gov,Hshld, Norld In APPENDIX E: LAKE STATE OUTDOOR RECREATION IMPACTS The following tables present estimated outdoor recreation economic impacts for the Lake States FHW study area. The estimates are stated in terms of millions of 1982 dollars for output and personal income, and numbers of jobs_for employment. They are presented according to type of recreation spending, aggregation scheme, trade estimate, and multiplier type (I or 111) they are associated with The five sets of trade estimates used in deriving impacts and which appear as column headings include: SDP - trade estimates based on the supply-demand pooling trade estimation technique; these represent ceiling values for IMPLAN RPCs and generate the largest multipliers FLRLK - RPCs are based on the minimum RPC values for the three Lake states, including questionable low RPCs ALTFLK - RPCs are the same as for FLRLK except for those FLRLK RPCs which appeared highly contradictory to a combination of SDP, MRIO, REMI, and Census estimates; ALTFLK RPCs represent a more accurate, or reasonable set of minimum RPC values. BSTLK - RPCs based on other RPCs, secondary data, and this author's judgment . UNCHLK - unchanged IMPLAN (version 2.0) RPC estimates Model aggregation and type of recreation spending are indicated in row headings. The aggregations are identified by the following prefix abbreviations: disaggregated (502 sector) model - ”DIS" minimally aggregated (308 sector) model - ”UNAG” all nonrecreation aggregation (159 sector) model - "AGLK" highly aggregated (31 sector) model - ”LK30" most aggregated (16 sector) model - ”LK16” Four types of recreation spending categories are indicated by abbreviations used as suffixes in row headings as follows: nonresident spending only - "NONRES" trip spending only - "TRIP” total spending less special equipment - "~SPEC" total spending - ”TOTAL” Two sets of low and high estimated impacts are presented. One set includes recreation spending encompassing transfers from one set of Lake State residents to another set of Lake State residents. The other set does not include such intraregional transfers, but does include payments from nonresidents to resident households. An exception to this is that “NONRES” rows are different. Nonresident outdoor recreation economic impacts stemming from the 1987 Governors' Conference on Fbrestry study are presented with the “LON w/res pces” impacts; total impacts stemming from the 1987 study are presented with the "LON nonres pces." FHW low nonresident impacts are presented with the "HIGH w/res pces” impacts and FHN high nonresident impacts are listed with the "HIGH nonres pces.“ No impacts based on the 1987 study are listed with the disaggregated model because 1987 spending categories could not be objectively disaggregated. 212 UNCHLK TYPE III UNCHLK TYPE III UNCHLK TYPE I UNCHLK TYPE I (1982 million 3) TYPE III BSTLK TYPE III BSTLK LON, w/res pces TYPE I BSTLK TYPE III TYPE III TYPE I 213 FLRLK ALTFLK ALTFLK BSTLK TYPE I TYPE III FLRLK ALTFLK ALTFLK TYPE III TYPE I FLRLK TYPE I FLRLK TYPE I SDPLK TYPE III 450.8 524.5 819.2 SDPLK TYPE III LAKE STATE OUTDOOR RECREATION ECONOMIC IMPACTS: 661.9 1,026.1 SDPLK TYPE I TYPE I 304.0 504.2 681.2 272.8 319.0 517.5 2 850 401 I P 8 1 1 2 PERSONAL SDPLK INCOME OUTPUT Appendix E (cont'd.) DISNONRES DISTRIP DIS'SPEC DISTOTAL AGLKNONRES AGLKTRIP LK16NONRES LK16TRIP UNAGNONRES UNAGTRIP AGLK-SPEC AGLKTOTAL LKIb-SPEC LK16TOTAL DISNONRES DISTRIP DIS-SPEC DISTOTAL UNAG-SPEC UNAGTOTAL AGLKNONRES AGLKTRIP AGLK-SPEC AGLKTOTAL 55:3 2.5m alml 35 s E CL mwnn RSO mmmm .LLLL 297.7 498.3 651.9 268.9 LK16NONRES LK16TRIP LK16-SPEC LK16TOTAL 214 Appendix E (cont'd.) LAKE STATE unooos RECREATIUI EMIC IFPACTS: L111, w/rss pces (cont'd.) (jobs) EMPLMIENT SDPLK SDPLK FLRLK FLRLK ALTFLK ALTFLK BSTLK BSTLK WCHLK WCHLK TYPE I TYPE III TYPE I TYPE III TYPE I TYPE III TYPE I TYPE III TYPE I TYPE III DISNGIRES DISTRIP 22,627 33,194 18,615 22,954 19,347 24,172 21,495 30,393 20,782 29,154 DIS-SPEC 31,704 46,511 5,688 31,676 26,940 33,659 29,927 42,317 28,737 40,315 DISTOTAL 37,615 55,183 28,9” 35,748 30,604 38,237 34,536 48,833 32,766 45,967 WAGMRES 19,227 28,191 15,765 19,56 16,470 20,382 18,254 25,654 17,975 5,397 lllAGTRIP 22,292 32,684 18,094 22,102 18,798 23,262 21,030 29,556 20,727 29,286 WAG-SPEC 33,237 48,732 26,969 32,943 28,174 34,865 31,284 43,968 30,733 43,424 WAGTOTAL 40,889 59,951 31,894 38,958 33,446 41,389 37,547 52,770 36,706 51,862 AGLKIKNIRES 19,575 29,044 15,734 19,049 16,409 20,114 18,357 25,794 18,065 25,380 AGLKTRIP 22,613 33,551 18,074 21,883 18, 746 22,978 21,124 29,683 20,826 29,260 AGLK-SPEC 33,189 49,244 26,432 32,001 27,566 3,7” 30,881 43,393 30,294 42,561 AGLKTOTAL 40, 069 59,452 31 ,369 37,979 32, 799 40,205 36,890 51,837 36, 076 50,684 LOOIOIRES 19,393 28,262 15,614 18,921 16,373 20,152 18,310 25,544 17,54 5,217 LK30TRIP 22,958 3,457 18,3” 22,296 19,176 23,601 21,606 30,142 21,172 29,688 LK30-SPEC 33,547 48,888 26,800 32,478 28,58 34,546 31,455 43,883 30,646 42,974 LK30TOTAL 40,907 59,613 31,658 38,365 33,339 41,034 37,609 52,468 36,499 51,180 LK1610NRES 19,50 28,129 15,497 18,757 16,m8 19,573 18,101 25,305 17,775 24,970 LK16TRIP 22,152 32,369 17,584 21,283 18,099 22,129 20,743 29,000 20,317 28,540 LK16'SPEC 32,263 47,143 25,270 30,586 26,139 31,960 29,861 41,747 29,014 40,758 LK16TOTAL 39,180 57,250 30,163 36,508 31,313 38,286 35,818 50,074 34,623 48,638 UNCHLK TYPE III UNCHLK TYPE III TYPE I UNCHLK TYPE I (1982 million 8) UNCHLK BSTLK TYPE III BSTLK TYPE III BSTLK TYPE I LON, nonres pces ALTFLK BSTLK TYPE III TYPE I TYPE III ALTFLK ALTFLK 2115 mu: ALTFLK me i we 1 TYPE III FLRLK TYPE III FLRLK TYPE III TYPE I FLRLK TYPE I SDPLK SDPLK TYPE III LAKE STATE OUTDOOR RECREATION ECONOMIC IMPACTS: SDPLK TYPE I TYPE I INCOME Appendix E (cont'd.) OUTPUT DISNONRES DISTRIP DIS-SPEC DISTOTAL AGLKNONRES PERSONAL SDPLK AGLKTRIP LK16NONRES LK30-SPEC LK16TRIP LK16'SPEC LK16TOTAL LK30TOTAL LK30NONRES LK30TRIP UNAG'SPEC UNAGTOTAL AGLK°SPEC AGLKTOTAL DISNONRES DISTRIP UNAGTRIP 08‘ see mmm 7 9 9 302. 489 666 DIS-SPEC DISTOTAL 0150 O I O O 3‘68 1 4 4 6 794 265. 434. 568. 419.7 .7 658.0 843.7 814.2 1,244.6 591.4 275.9 8 6 6 8 888 3. 491. 668. UNAGNONRES UNAGTRIP UNAG'SPEC UNAGTOTAL ‘5005 707.0 878.0 816.4 1,350.9 274.1 440.9 558.7 318. 508. 652. 931.4 311.3 501.6 681.1 932. LK30NONRES LK30TRIP AGLKNONRES AGLKTRIP AGLK-SPEC AGLKTOTAL LK30°SPEC LK30TOTAL LK16NONRES LK16TRIP LK16'SPEC LK16TOTAL 216 Appendix E (cont'd.) LAKE STATE (moons RECREATIGI EWIC IMPACTS: LN, nonres p608 (cont'd.) (jobs) EMPLOYENT SDPLK SDPLK FLRLK FLRLK ALTFLK ALTFLK BSTLK BSTLK lliCHLK LNCH LK TYPE I TYPE ”1 TYPE I TYPE III TYPE I TYPE III TYPE I TYPE III TYPE I TYPE ”I DISNGIRES DISTRIP 22,552 5,085 18,555 22,880 19,283 24,093 21,424 30,294 20,713 29,058 DIS-SPEC 30,891 45,318 5,031 30,866 26,249 32,796 29,160 41,51 27,994 39,272 DISTOTAL 36,802 53,991 28,35 34,938 29,913 37,374 5,768 47,748 32,023 44,924 WWRES 65,957 96,706 54,577 66,665 57, 020 70,561 62,738 88,172 62,163 87,831 WAGTRIP 22,217 32,574 18,034 22,029 18,735 5,185 20,959 29,456 20,658 29,188 WAG-SPEC 32,416 47,527 26,319 32,148 27,489 34,017 30,515 42,887 29,976 42,354 lRlAGTOTAL 40,067 58,746 31,243 38,163 32,761 40,542 36,777 51,688 35,949 50,792 AGLKIKNIRES 67,315 99,875 54,487 65,967 56,812 69,639 63,220 88,55 62,606 87,957 AGLKTRIP 22,561 5,473 18,032 21,51 18,701 22,924 21,075 29,614 20,778 29,192 AGLK-SPEC 32,621 48,45 5,972 31,445 27,088 5,203 30,348 42,644 29,774 41,50 AGLKTOTAL 39,501 58,609 30,910 37,45 32,321 39,618 36,357 51,088 35,556 49,954 LOOMRES 66,988 97,620 54,672 66,55 57,303 70,529 63,476 88,554 62,570 87,738 LK30TRIP 22,894 5,363 18,346 22,53 19,121 5,534 21,545 30,057 21,112 29,604 LK30-SPEC 32,842 47,861 26,55 31,794 27,469 5,809 30,792 42,958 29,992 42,056 LK30TOTAL 40,202 58,585 31,093 37,680 32,741 40,297 36,946 51,543 35,844 50,262 mamas 65,996 96,45 53,808 65,127 55,562 67,935 62,246 87,020 61,311 86,127 LK16TRIP 22,078 32,261 17,527 21,214 18,040 22,057 20,675 28,904 20,50 28,446 LK16-SPEC 31,547 46,097 24,696 29,891 25,542 31,230 29,191 40,809 28,353 39,829 LK16TOTAL 38,549 56,328 29,649 35,885 30,777 37,630 35,224 49,244 34,037 47,814 UNCHLK TYPE III TYPE I (1982 million 5) BSTLK UNCHLK TYPE III HIGH, w/rss pces TYPE III TYPE I ALTFLK ALTFLK BSTLK TYPE I 217 TYPE III FLRLK FLRLK TYPE I TYPE III LAKE STATE OUTDOOR RECREATION ECONOMIC IMPACTS: SDPLK SDPLK TYPE I OUTPUT Appendix E (cont'd.) DISNONRES DISTRIP DIS'SPEC DISTOTAL UNAGNONRES UNAGTRIP AGLKNONRES AGLKTRIP LK30TOTAL UNAG'SPEC UNAGTOTAL LK30NONRES LK30TRIP LK30-SPEC AGLK-SPEC AGLKTOTAL 5 1 1 7 2 55 20 .34 . 5 3 1 2 a 6 LK16NONRES LK16TRIP LK16-SPEC LK16TOTAL UNCHLK TYPE III BSTLK UNCHLK TYPE I TYPE III ALTFLK BSTLK TYPE III TYPE I ALTFLK TYPE III TYPE I FLRLK FLRLK TYPE I SDPLK TYPE III TYPE I PERSONAL SDPLK INCOME UNAGNONRES UNAGTRIP AGLKNONRES AGLKTRIP LK16NONRES LK30-SPEC LK16TRIP LK16-SPEC LK16TOTAL LK30TOTAL 'LK30NONRE8 DISNONRES LK30TRIP DISTRIP DIS-SPEC DISTOTAL UNAG-SPEC UNAGTOTAL AGLK‘SPEC AGLKTOTAL 218 Appendix E (cont'd.) LAKE STATE ouroooa RECREATION ECOIOIIC IIPACTS: HIGH, N7". peas (cmt'dd (jobs) EMPLOYENT SDPLK EPLK FLRLK FLRLK ALTFLK ALTFLK BSTLK BSTLK WCHLK lRiCHLK TYPE I TYPE III TYPE I TYPE III TYPE I TYPE III TYPE 1 TYPE ”I TYPE i TYPE III DISNOIRES 4,813 7,061 3,968 4,893 4,126 5,155 4,578 6,473 4,430 6,215 DISTRIP 49,369 72,427 40,45 49,848 42,080 52,576 46,818 66,200 45,266 63,502 DIS-SPEC 76,261 111,880 61,401 75,714 64,584 80,692 71,848 101,591 68,988 96,781 DISTOTAL 101,044 148,238 75,317 92,873 80,59 99,963 91,151 128,886 85,843 120,428 lllAGIDNRES 4,759 6,978 3,872 4,730 4,025 4,981 4,497 6,320 4,442 6,276 WAGTRIP 48,54 71,629 39,478 48,222 41,51 50,837 46,019 64,676 45,515 64,309 WAG-SPEC 79,811 117,018 64,326 78,573 67,399 5,406 74,980 105,379 73,749 104,201 lRlAGTOTAL 112,288 164,635 5,448 104,374 89,969 111,335 101,621 142,822 99,213 140,179 AGLKWIRES 4,895 7,263 3,918 4,744 4,54 4,982 4,575 6,429 4,520 6,351 AGLKTRIP 49,57 73,053 39,144 47,392 40,652 49,51 45,930 64,540 45,461 63,870 AGLK-SPEC 77,185 114,520 61,35 74,55 64,104 78,578 71,843 100,952 70,617 99,211 AGLKTOTAL 100,010 148,388 78,031 94,472 81,691 15,135 91,777 128,963 ' 89,819 126,189 LK301KNIRES 4,914 7,162 3,947 4,783 4,116 5,066 4,632 6,463 4,545 6,373 LK30TRIP 48, 725 71,05 39,538 47,915 41,227 50,742 46,040 64,50 45,57 63,460 LK30-SPEC 77,127 112,396 62,524 75,770 65,522 80,644 72,662 101,369 70,951 99,490 LK30TOTAL 108,554 158,194 5,311 100,960 88,076 108,404 98,966 138,56 95,943 134,536 LK16KNIRES 4,764 6,962 3,802 4,601 3,913 4,75 4,472 6,52 4,386 6,162 LK16TRIP 48,910 71,468 38,827 46,994 39,999 48,95 45,85 64,55 44,950 63,145 LK16-SPEC 72,763 15,322 57,930 70,115 59,961 73,314 67,648 94,574 65,831 92,477 LK16TOTAL 102,788 150,195 79,284 95,961 82,520 15,897 93,597 130,51 90,267 126,804 219 Appendix E (cont'd.) (1982 million 3) man, mum pee: LAKE STATE OUTDOOR RECREATION ECONOMIC IMPACTS: UNCHLK TYPE III BSTLK UMCMLK TYPE I FLRLK ALTFLK ALTFLK BSTLK TYPE III TYPE I TYPE III TYPE I TYPE III FLRLK TYPE I SDPLK TYPE III SDPLK TYPE I OUTPUT DISNONRES DISTRIP DIS-SPEC DISTOTAL UMACNOMRES UMACTRIP LK16MOMRES LK30TOTAL LK16TRIP LK16‘SPEC LK16TOTAL LK30-SPEC LK3OMOMRES UMAC‘SPEC UNAGTOTAL AGLKMOMRES AGLKTRIP ACLK-SPEC LK30TRIP AGLKTOTAL BSTLK UNCHLK UNCHLK TYPE III TYPE I TYPE III BSTLK TYPE I ALTFLK TYPE III ALTFLK TYPE I SDPLK FLRLK FLRLK TYPE I TYPE III TYPE I TYPE III PERSONAL SDPLK INCOME UMAGMOMRES UMAGTRIP AGLKNONRES AGLKTRIP LK16NOMRES LK30TOTAL LK16TRIP LK16-SPEC LK16TOTAL LK30-SPEC LK30NOMRES DISNONRES DISTRIP DIS-SPEC UNAG-SPEC UMACTOTAL AGLK-SPEC AGLKTOTAL LK30TRIP DISTOTAL 220 Appendix E (cont'd.) LAKE STATE 5T05R RECREATIGI ECOIKHIC IMPACTS: HIGH, nonres pee: (cont'd.) (jobs) EMPLOYED" SDPLK SDPLK FLRLK FLRLK ALTFLK ALTFLK BSTLK BSTLK WCHLK WCHLK TYPE I TYPE III TYPE I TYPE III TYPE I TYPE III TYPE I TYPE III TYPE I TYPE III DISIIGIRES 12,504 18,344 10,251 12,641 10,677 13,341 11,869 16,783 11,486 16,113 DISTRIP 49,194 72,170 40,25 49,673 41,930 52,389 46,652 65,56 45,105 63,277 DIS-SPEC 74,53 15,934 59,778 73,713 62,878 78,560 69,952 98,911 67,151 94,205 DISTOTAL 99,036 145,292 73,694 90,872 78,302 97,52 89,256 126,25 84,57 117,852 WWES 12,420 18,210 10,048 12,274 10,465 12,951 11,714 16,464 11,628 16,430 WAGTRIP 48,677 71,369 39,338 48,051 40,933 50,654 45,853 64,443 45,352 64,078 WAC-SPEC 77,783 114,044 62,719 76,610 65,707 81,312 5,079 102,708 71,879 101,559 WAGTOTAL 110,59 161,661 5,841 102,411 88,277 109,241 99,721 140,151 97,344 137,538 AGLKWRES 12,667 18,794 10,038 12,153 10,434 12,790 11,85 16,585 11,723 16,469 AGLKTRIP 49,113 72,870 39,045 47,271 40,548 49,703 45,814 64,377 45,348 63,711 AGLK-SPEC 75,774 112,428 60,194 72,877 62,918 77,15 70,522 99,096 69,327 97,399 AGLKTOTAL 98,599 146,294 76,892 93,093 80,504 5,680 90,455 127,15 88,529 124,376 LGOMRES 12,56 18,95 10,401 12,604 10,866 13,35 12,240 17,075 12,52 16,871 LK30TRIP 48,572 70,75 39,416 47,766 41,098 50,55 45,896 64,029 45,115 63,261 LK30-SPEC 75,378 109,848 61,126 74,075 64,59 78,820 71,018 99,075 69,329 97,216 LK30TOTAL 106,55 155 ,646 81,912 99,266 86,594 15,55 97,322 135, 772 94,321 132,262 LK1610IRES 12,551 18,340 10,05 12,111 10,312 12,608 11,788 16,480 11,584 16,272 LK16TRIP 48,57 71,215 38,692 46,51 39,59 48,55 45,665 63,841 44,793 62,95 LK16-SPEC 70,994 15,58 56,511 68,398 58,486 71,510 65,992 92,58 64,198 90,15 LK16TOTAL 101,024 147,617 77,867 94,246 81,047 99,096 91,944 128,541 88,637 124,514 APPENDIX F: BRIDGE OF NATIONAL SURVEY OF FISHING, HUNTING, AND WILDLIFE-ASSOCIATE RECREATION (FHW) SPENDING T0 IMPLAN ”Samua'dofttle.w Allo- Fifi who Ont-only sum of cotiai (NU It. Oily) Fm Allmim; PE Ina, H) m, SIC Orb, *la' IMAM Setor Allmtim " use: TRIP-RELATE foodlatiru fad Grooorioo u-o 1110, Food mn- Off-Fruit. Can-p; also 1-0 1500, tdaooo) ileuicnl not: 0.6 ism (1-0 113), w hols I. W) sadi- 0.6 lmim all grain to m i 471 trnportotim Mali: twins 5 an 0.5 Air travel (1” “50) Gulch-R's 0.6 Railronh (1M “46) Travel 0.01 an. (new “47) Yrs-b 0.5 Print. Winn Ml.) Ntl PG: 0.92 G- ! oil (1-0 8140) 0.62 Auto mir (1-0 8130,- lure mrrouly, PE 1358 S 139, I” I493) 0.3 1 other an I. Diana-I WM rut-l (PE 175-1724, W 55(7)) 5 Fill 0.14 trprlod out: with fcu (1M 55) 0.12 ' Pod: trips (1” 5(2) 0.359 Mali: ltd no (Pd, St, 8 Lo: Gov, m utiplia') 0.164 Prim lad no 0.26 mint (SIC 0771, min! I” 94; or 5(2) :5 hoc, loool 0.12 Md (distrilmo w Pl: moor; no mu m Pl: sector) stuiioo/cpiniafldfi m Fimim Swim real (Pl! 172-1m, 1M 55(7)) 6 nu 0.88 trp-rltd out: Boot lu'diiru (m 55) 0.174 thick (on mm» 512) 0.078 Pain trip (1191M 55) 0.21 ablic ltd no (Fed, St, 8 Lot (in, re ultiplier) 0.131 Prim [I'd no 0.158 will (SIC 0911; Fl! 51, 175-7; nryird m a.) d boo, loool 0.6 Md (diurilmo w PE actor; no mu in PE m) tunic/miniatm boot Yul p 8 oil (1-0 8140) but minor/nine but Inorim 8 mire (PG 165, I19 “49) d hoe, loo-l 0.24 bait ice “TUB m t rifl. muniticn bout-intuit: (PIE 15078155, ”Mania-d) (SIC 579, I” a) (SIC 2097, P5 15, I” 815 I “M (PG 1481, I” m a mint) (Pl! 1441-1444, an» m - mitt!) 221 main/minim 0.76 222 Appendix F (cont'd.) FIN 990m mam Fume MUG; PG I131, I-O um, SIC (me, +15 I” m WHO! “IS WTIOI 2 edict-humip Maxim (SIC 3969, mm, H2165, Imus-ruins!) 6m Ialncqaic aidm (SIC 385, I-0 «no, PG 136-36, new 63 3 min!) Dacoyaluucalla (903969, #0935, PI: 166661671, Niacin-11!!!!) 6431p or w- cnoalcarriara (SIC 3969, PO 965, K215“, new 63 a mind) Had lodim «pip (SIC 3959, I-0 935, P5 1666 I. 1671, 1101M 633 n mind) Miro (hp 8. momenta (by sardine (SIC 579 I. 0752, PI! 1706-5 61709-10, I“ a n FIB an mind) food (SIC 2067, PI! 96-5, mm 15 a mind) minty (SIC 0762, PG 1756, "PIN! ms a mit'd) mm (SIC 3199, a: 1695, mum £56 a was) Initim knivu (SIC 3621, PI! 16%, ”FUN “19 n mind) Unpacifidothar ( “ SIC”, POMS“, MKS-mind) WM field gluon SIC 3852, I-0 965, P5 196.36, noun 625 n mird binowlara SIC m, I-O 9600, PG 1:56-36, new 68 8 WM filalav. filo (SIC 351, PG 1662, Immanmirld) adhoc “min (SIC 7395, PI! 175-27, I” 665 n mi!!!) odiard‘uotoaqfip $103851analaopt.3832&3661),Imwnmind) canyiru cnaalclothirc 1w (SIC 3161, PC! 671-77 1 158, W $1) d hoc clodiim (SIC 25, 1mm 151 a main!) bird and SIC 3367, PG 97, 1M #15 a mind 3 Whiditlll 6 cum-Wimp FISHIin rock racla lira, human. lu'aa I. (11. tackla hm- cracla, nu nit cuminra acalaa 6 him 5 other (iii aopip lbod bird “1“,” (SIC 26”, Pl! 762, I” £172 a mil-III) ad hoc Ccranic bind no: (SIC 35-7, Pl: 776-7, m m (or am a mind) sncmmcsw, [MEI—VIM SICM, PG 1667, ”“35an SIC 3969, PCE 1667, "FUN “33 n mim lino (SIC 296, Pt: 1665(aloarntivaly, Pl: 394), IM S162 a mind» hooks,etc. (SIC $69, PG 1667, IIPLNI “I! n min!) SIC ”09, P5 1667, I191!“ m 8 mid SI: 3969, a: 11.67, m am a mind cnaala (SIC m, Pl! 1667,IM#633.~11M:D dhoc nu (SIC m (I: 2399), FE ”908,660,”,16fi); IIPIM #162 (159) n mird) SIC 3969, 96160, m I68 a mind acalu (SIC 575, PIE 795-6, I” ”66 I .111”) kniwa (SIC 3621, PG 165-9, 1191.»: 319 a Hair-d) m t. (id: firth-a l actor olac. (iii civic. (SIC 3662, 1M 35) 5 an odnr fidiiroqfipv" SICW, #0931169“), manual-aim 0.2 0.12 0.06 0.5 0.11 0.33 0.6 0.2 0.07 0.01 0.5 0.5 0.6 0.6 0.5 0.5 0.9 0.1 0.7 0.3 0.2 0.8 0.1 0.9 0.52 0.68 223 Appendix F (cont'd.) Fill 990m um ruma mum; PE ITBI, [-0 am, SIC (me, +10 [M m WHO! MSIS ALLG‘ATIOI MILIMY WIRE“ F191 3. m M Eqfip 6 Wm Equip Fail lhathar Paar 81c Clothim Ruhr W apip him Fish or m Boots 7 Other F91 In: Equip “CIAL WIHBII carpirulumborcookim apip(SICw, P56“, ”Swamim dhoc ”acumen“, P5 was, “61553.11!“ “”15 h (SIC 2399. PG 1687, "FUN #159) Into". 6 char lidwtira quiuSIC 3&8, P5 mo, I” 587 a mi!!!) Md: mu, chifla m, comm-(filo I" - SIC 2393, PG 165, I” #156) SIC 3,!!! 655.6 6 669, I” 3151 n .11!“ SIC 2m, noun #151 n mird SIC 321, P5 29?, nvun 061. mid sn: 7699 (3732 m, H: 1569, 155-6, mo, noun am - lamina SIC 3163, Pl! 6”, IIIPLNI 968 binoculars, field glnaca (SIC 3832, I-0 965, PCE 1536-36, 1M 623 a mind 5 Flu m m a chic ($1: 3969, 9121559, now #635 n mind) Focusing-tiny can (SIC 7699, PCE 1573, IIOLNI I 678 n mird) mifid m ( "' SIC R9, I-O 935 S 965,091.»! “33 . mird) Fia‘flilntfincavip MSW BoatAccacaorin Boat trailarfliitch SIC 37.2, P5 1507.21 (5 1691?), I” m 8 m1!!! mu m (SIC $19, 95155, noun 331) 5 Fl“ alac. trolliru Mora (SIC 3621, Pl! 873 (6876) [M 676) other that macaoriaa ( 1" that ashlar-SIC 2392, I)? flSflifa W10 359 8 362, no as 6 6a); cup-Sic 2396; ETC.- anal: to no #635 as mind) SIC 3799, PI! 1531, I” “15 n mird Travel or tent Trailer travel or tent trails (SIC 3792, PI! 1297-1301, I” #612 n mird) 5 FIN Piddp,Vm,iiotor lion pickup, cum or m or 05in piddp (SIC 3711, P5 1291-2, 1M #65 n namirfl) ad hoc cupcra (SIC 3792, PI! 125-12%, "FUN “12 n val-gird) m (SIC 3716, PCE 118-5, IIPUII “16 n mird) Who-6mm, Pl! 135, ““12501'911") cbin (m m cattncticn ~pt.SIC 1521, 17: 1M 5) Off-Rod Vehicle off road miclaa (SIC 3799, Pl: 155 S 132, IM “15) 5 Flu ' mutilaa (SIC 3799, PI! 1529, 1M “15 3 WM) IcaChaca SIC3079,P5&3, Imm-nrgim 8 mump trawl ortnttrailu' (SIC 375, P51274301, ”Pull “12.!I'BIM) 5 Flu off rod which; (SIC 3799, no. 155 a 152, new «(15) piano. war. or m pidnp (51:: 3711, 91: 1291-2, new «as n mind) ad has cam (SIC 3792, P5 1259-1290, 1m an: a mind) m (SIC 3716, PI: 1113-5, I” #616 n mim u'macifidolhcr < .. s1cm,x-o93m&9(m,lm mun-vim 0.15 0.35 0.5 0.15 0.1 0.36 0.5 0.58 0.06 0.76 00m 0.17 0.12 0.58 0.29 0.5 0.26 0.13 0.17 0.97 0.03 0.5 0.51 0.37 0.18 0.5 0.16 0.07 224 Appendix F (cont'd.) FIN 9900‘ um ”Tia WTIOI; PE "a, I-O um, SIC ans, *1“ I” m mum BASIS mum 9 mnwpmtmp