Dataset User Guide - The Centre for the Study of African Economies
Contents
1. since their survey excludes furniture producers which are mainly small scale enterprises In these cases we have used the price index for the ISIC category which is closest to the missing category e g we have used the wood products ISIC 3319 price index to deflate the outputs of furniture firms in our sample This is obviously not an ideal solution and we hope in the future to develop firm specific price deflators using internal price data from our survey Some data on prices of firm outputs and material inputs were collected in all four rounds of the survey In the three RPED surveys product prices can only be derived from data on quantities of products produced and the total value of output or sales In the Wave 4 survey firms were explicitly asked for unit sale prices and input prices It is intended that this data will be used in future analysis of firm growth and productivity to construct a set of alternative producer price series for comparison with the NBS price indices Other studies have also emphasised the importance of allowing for differential changes in firms output and input prices when constructing real VAD series Capital stock deflator We do not have a reliable measure of changes in the domestic prices of firm s plant and machinery and other capital goods A considerable proportion of these capital goods are imported and hence their shilling value depends partly on changes in nominal exchange rate The capital stock deflato2. For analytical purposes where possible these main sectors are further disaggregated into a total of ten subsectors Sector ISIC Code Variable Name Food products ISIC 3110 3129 exc 3117 FOODX Bakeries ISIC 3117 BAKE Beverages ISIC 3130 3135 BEVS Textiles ISIC 3210 3219 TEXTX Garments ISIC 3220 GARMENT Footwear ISIC 3240 SHOES Wood Products ISIC 3310 3319 WOODX Furniture ISIC 3320 FURN Fabricated Metal ISIC 3810 3819 METALX Machinery ISIC 3820 3850 MACHINE Size dummies These are based upon average number of employees including both full and part time workers but excluding seasonal workers over the three waves of the survey MICRO Dummy for micro firms 1 5 employees inclusive SMALL Dummy for small firms 6 29 employees inclusive MEDIUM Dummy for medium firms 30 99 employees inclusive LARGE Dummy for large firms 100 or more employees Ownership dummies STATE Dummy for 100 state owned enterprises SSTATE Dummy for firms with some degree of state ownership less than 100 SFOR Dummy for firms with some degree of foreign ownership less than 100 TZOWN Dummy for private firms with 100 domestic ownership ANYFOR Dummy for firms with any degree of foreign ownership including 100 ANYST Dummy for firms with any degree of state ownership including 100 Alternative Ownership Dummies PRIVDOM Dummy for Private domestic owners only PRIVFOR Dummy for Pr
3. from highest level of formal education completed IMPEDUC Imputed education number of years taken to complete highest level of education calculated from worker age and year left school assuming entry at age 6 TENURE Number of years employed by current firm WAGE Wage received per time period ALLOW Allowances received per time period EARN Total earnings wages amp allowances per time period PCALLOW Ratio of allowances to total earnings Occupational Categories The worker questionnaires contain a total of 19 occupational categories into which workers were classified OCCUPAT These have been organised into a total of 8 combined categories for analytical purposes with composite groups in brackets MGMT Dummy 1 for management employed manager owner manager ADMIN Dummy for senior administrative staff engineer scientist accountant technician Note all should have university degree or professional qualification CLERIC Dummy for clerical staff skilled office unskilled office SALES Dummy for sales staff specialised sales sales assistant SUPER Dummy for production supervisors TECH Dummy for technical staff electricians plumbers welders maintenance PROD Dummy for production workers machine operator labourer helper skilled production worker service worker APPREN Dummy for trainees apprentices 14 Tanzania RPED Dataset User Guide CSAE University of Oxford 7 Price Deflators Used
4. interests of maintaining confidentiality the information that would identify the specific firm interviewed has been excluded name of firm address of firm name of person interviewed Please contact CSAE if you require further information in this regard A great deal of work has already been undertaken to manipulate the raw data into a more usable format and derive a number of the more important continuous and categorical variables in which economic researchers may be interested This work is described in more detail in Section 5 A selection of these constructed variables are contained in the SAS program and data files included with this dataset in conjunction with a descriptive listing of some of the main variables in Section 6 below Tanzania RPED Dataset User Guide CSAE University of Oxford 4 Questionnaire Structure and Coding The dataset presented here has been extracted from a detailed questionnaire conducted with the owners senior managers and for relevant sections workers of the sampled manufacturing firms The questionnaire was designed by a team from the World Bank Over the three waves of the survey the structure of the questionnaire and the range of questions included has evolved in the light of field experience and in response to emerging research issues Scanned copies of each of the questionnaires are included with the material made available here The overall questionnaire has been divided into a number of section
5. products 100 124 168 231 253 226 213 242 3118 Sugar refineries 100 108 166 215 237 249 238 243 3119 Confectionary 100 117 134 147 153 163 190 207 3121 Food products amp animal feed 100 139 138 223 221 220 250 283 3122 Food products amp animal feed 100 139 138 223 221 220 250 283 3131 Distilled spirits wine amp beer 100 110 140 156 179 188 207 212 3132 Distilled spirits wine amp beer 100 110 140 156 179 188 207 212 3133 Distilled spirits wine amp beer 100 110 140 156 179 188 207 212 3134 Soft drinks 100 129 179 210 327 401 399 391 3140 Tobacco amp cigarettes 100 117 179 210 260 261 263 264 3211 Spinning amp weaving 100 101 114 186 209 217 220 221 3212 Made up textiles 100 132 187 266 292 346 360 357 3213 Knitting mills 100 104 119 191 227 202 180 197 3214 Carpets amp rugs 100 104 119 191 227 236 241 246 3215 Cordage rope amp twine 100 126 180 240 299 328 351 371 3219 Other textiles 100 104 119 191 227 236 241 246 3220 Garments 100 104 119 191 227 236 241 246 3233 Leather products 100 104 119 191 227 236 241 246 3240 Footwear exc rubber amp plastic 100 104 119 191 221 268 269 280 3311 Sawmills 100 147 156 194 233 241 256 249 3312 Wood products 100 147 156 194 233 241 256 249 3319 Other wood products 100 147 156 194 233 241 256 249 3320 Furniture amp fittings 100 147 156 194 259 311 295 295 3511 Industrial Chemicals 100 101 155 184 198 217 247 339 3513 Plastics amp Foam 100 101 155 184 198 217 247 339 3811 Cutlery tools a
6. to be able to use any of the SAS programs described in the Data Analysis section below A SAS conversion program convspss sas is included Conversion of data files from SPSS into SAS or other statistical program formats could also be undertaken using the Stat Transfer software package There are one set of the data files one firm level and one worker level for each of the 3 waves which are the original files as received from the Helsinki School of Economics in January 1999 A significant amount of work was subsequently undertaken to clean some of the main variables contained in these files and also to incorporate new information for firms which were re interviewed in Wave 4 of the survey in late 1999 Another set of the files those with the new suffix contains the cleaned data which has been used in analytical work at CSAE Note that there are still sections of the data which may require further cleaning and consistency checks before use Hard copies of the individual firm questionnaires are stored at CSAE and may be consulted on the basis of a written request and permission being granted The coding of variables contained in these files follows the structure of the individual wave questionnaires These are described in more detail in the next section Variable names follow the page numbers though not exactly and question numbers in the questionnaire Hence variable V0301 refers to page 3 question 1 and variable V2610a refers to page 26 qu
7. 1994 accounting data smaller firms without written accounts report 1995 figures The three RPED surveys collected one year of data plus historical data for some of the key variables Wave 1 was undertaken in August 1993 data relates to 1992 Wave 2 was undertaken in October 1994 data relates to 1993 Wave 3 was undertaken in January 1996 data is from 1995 for some firms 1994 for others The time inconsistency of the data needs to be taken into account when constructing constant price series by identifying the year to which the data given by each firm refers and deflating by the respective year s price index Total sample size was 217 firms in Wave 1 213 firms for Wave 2 and 152 firms for Wave 3 Unfortunately no replacement firms were selected in Wave 3 which accounts for the fall in sample size Missing data from the relevant production section of the questionnaire and subsequent investigation of the data consistency has led us to also exclude a number of these firms from our analytical work The original sampling frame was the 1989 Industrial Census for formal sector firms produced by the Tanzanian National Bureau of Statistics which nominally included all manufacturing establishments with 10 employees or more A sample of informal sector firms were Tanzania RPED Dataset User Guide CSAE University of Oxford also chosen randomly by the interview teams in the locations in which the survey was conducted The firms co
8. 385 52 4 95 459 98 5 07 311 05 4 78 1995 574 76 339 30 457 03 5 87 527 67 5 82 386 39 5 94 1996 579 27 406 10 492 69 6 32 544 64 6 00 440 73 6 77 1997 612 12 471 40 541 76 6 95 583 98 6 44 499 54 7 68 1998 664 67 531 70 598 19 7 68 638 08 7 03 558 29 8 58 1999 739 25 573 60 656 43 8 43 706 12 7 78 606 73 9 32 Source Author Calculations Table 4 Tanzania Nominal and Real Exchange Rates 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Nominal Exchange Rate Tsh US 32 7 64 26 99 29 143 38 195 06 219 16 297 71 405 27 509 63 574 76 579 27 612 12 664 67 739 25 Index 1992 100 11 0 21 6 33 4 48 2 65 5 73 6 100 0 136 1 171 2 193 1 194 6 205 6 223 3 248 3 Real Exchange Rate a Tsh US 193 06 219 69 222 17 195 66 218 35 238 13 229 86 209 71 177 59 159 20 149 51 154 14 Index 1992 100 88 4 100 6 101 8 89 6 100 0 109 1 105 3 96 0 81 3 72 9 68 5 70 6 Notes a RER Nominal exchange rate US Export Price Index Domestic CPI 18 Tanzania RPED Dataset User Guide CSAE University of Oxford Selected Bibliography Blanc Xavier 1997 Industrial Change under Structural Adjustment Tanzania 1993 1996 Wave 3 Report Helsinki School of Economics February 1997 Harding Alan and Francis Teal 1999 Tanzanian Manufacturing Sector Performance in a Comparative Perspective 1992 95 CSAE Report University of Oxford Octob
9. In order to construct the constant price series for gross output OUTPUT and value added VAD we have experimented with the use several alternative price deflators Consumer Price Index We initially used the consumer price index CPI for mainland Tanzania as our price deflator The trend path of this index is shown in Table 1 Price inflation was an important factor throughout the survey period peaking at 33 1 in 1994 hence our results are potentially sensitive to price changes It is believed that changes in prices faced by domestic producers for their inputs and outputs may differ considerably from levels of consumer price inflation due to increased competition in most product markets and a number of additional price distortions facing domestic producers including indirect taxation and tariffs on their imported inputs Table 1 National Consumer Price Index Mainland Tanzania CPI change Food change Non Food change 1992 100 1992 100 1992 100 1988 36 0 1989 47 0 30 3 1990 63 8 35 8 1991 82 1 28 7 1992 100 0 21 8 100 0 100 0 1993 125 3 25 3 120 1 20 1 133 8 33 8 1994 166 7 33 1 167 1 39 1 165 8 23 9 1995 216 4 29 8 216 7 29 7 208 9 26 0 1996 259 0 19 7 260 9 20 4 254 8 22 0 1997 300 6 16 1 306 5 17 5 288 0 13 0 1998 339 1 12 8 351 6 14 7 311 3 8 1 1999 365 8 7 9 382 5 8 8 328 7 5 6 Source IMF International Financial Statistics Bank of Tanzania Economic Bulletin 2000 Q1 To take account of possible int
10. Regional Program on Enterprise Development RPED Tanzania Manufacturing Sector Survey Waves I III 1993 1995 Dataset User Guide Centre for the Study of African Economies University of Oxford January 2002 This data was originally collected by a team from the Helsinki School of Economics in conjunction with the Tanzanian Confederation of Industries and the University of Dar es Salaam The surveys were undertaken as part of the Regional Program on Enterprise Development RPED organised by the Africa Technical Department of the World Bank We are grateful to Xavier Blanc from Helsinki for making the data available to CSAE researchers in preparation for Wave 4 of the Tanzanian Manufacturing Enterprise Survey undertaken in late 1999 see separate dataset and documentation also on CSAE website This dataset forms part of an ongoing CSAE research project into manufacturing sector performance in Tanzania and Ghana funded by the Department for International Development DFID Address for correspondence Centre for the Study of African Economies Department of Economics University of Oxford Manor Road Oxford OX1 3UL Please contact Alan Harding alan harding economics ox ac uk or Dr Francis Teal francis teal economics ox ac uk for further information Tanzania RPED Dataset User Guide CSAE University of Oxford Contents Ackn wledgemems mnn oevckcc evan tase chavcsas Web a ew avelox dass Nav aad Ned Case aE 2 We MOUTON a
11. Where there were significant doubts about the veracity or consistency of data for specific firms these firms were excluded from the analysis in the relevant SAS program 10 Tanzania RPED Dataset User Guide CSAE University of Oxford characteristics of each worker e g sex age educational standard attained and previous work experience In Wave 2 the worker s name was recorded and then each worker was asked whether they had been interviewed previously If they said they had then the time invariant questions were skipped However without the worker s name in Wave 1 it is often difficult to match these Wave 2 workers with their crucial human capital characteristics recorded in Wave 1 where up to 10 workers are recorded per firm Matching can only be done using job categories and wages received but these are often inconsistent and common to several of the workers The worker ID s across waves are not consistent hence these cannot be used either i e the worker in firm x with WID 2 in wave 2 is not necessarily the same worker with WID 2 in wave 1 In Wave 3 workers were asked their names and whether they had been interviewed in the previous round Observations on Wave 2 3 only workers are relatively easy to match since their names are recorded in both years But observations on Wave 1 2 3 workers suffer from the same matching problem as above In order to try to sort out these problems and minimise the number of usable obser
12. al wave files e g industrial sector INDU formal or informal firm SECT isic por file containing 4 digit ISIC International Standard Industrial Classifications code for each firm Questionnaires wlquest File containing a scanned version of the Wave 1 questionnaires firm and worker showing the coding structure utilised in the data entry programme w2quest File containing a scanned version of the Wave 2 questionnaires showing the coding structure utilised in the data entry programme w3quest File containing a scanned version of the Wave 3 questionnaire showing the coding structure utilised in the data entry programme SAS Programs amp Datasets SAS program convspss sas which undertakes conversion of SPSS portable files raw data into SAS data file format sd2 so that they can be used with the SAS programs listed here A set of 12 additional SAS program files sas which use the raw data files above and create a series of both firm level and worker level variables see table below Each program file generates one or more permanent SAS datasets sd2 containing the main variables created A selection of these are included here see table below Tanzania RPED Dataset User Guide CSAE University of Oxford SAS Program Description Datasets Notes Created genfirm1 sas organises general firm variables wlchars sd2 some firms excluded due to unreliable from Wa
13. aw materials purchased by an amount equivalent to the change in stocks of raw materials Indirect Costs MISC Indirect costs of production including rent utilities transport security maintenance and other overheads but excluding labour or capital costs Value Added VAD Calculated by taking firm output less raw material inputs and indirect costs Labour costs LABCOST Cost of labour including wages WAGES and allowances ALLOW Calculated Profits PRCALC Calculated by subtracting labour costs from value added This is a measure of gross profits before tax and other charges including interest payments and depreciation Reported profits PROFIT Gross profit before tax and capital charges as reported in the firm s accounts where available Profit Rate PRATE Measured as the ratio between calculated firm profits and the replacement value of the firm s capital stock PRCALC CAP Capital Stock CAP This is the replacement value of the firm s total stock of plant and equipment excluding land and buildings which is the preferred measure used in all derived variables using capital Another variable CAPSALE measures the capital stock s estimated resale value Total Employment EMP Total number of firm employees including the owner manager including both full and part time workers but excluding seasonal workers Capacity Utilisation CUTIL Measured as actual output potential output 100 Potential output was det
14. cne aie otek aige tear el Pun airs ehd a E tain die a a a 3 2 List of Files and their Content 0 0 0 0 eecceeescecesececeeseeeeseeeeeseneeeesneeesseeeeeaeeees 5 34 Raw Data Piles ico cascsstasaiesstianinardveiaeedlarunrdredatedraraaard uaardamaumavaws 7 4 Questionnaire Structure amp Codit lt ssscccesscvecsstasetesseleveasssccvsnav ence aecsveriacecssaeed 9 De Basic Data Analysis itno neo i aaa ei A EE E A A Ea EA 10 Additional Note on Worker Data 6 Variable DeMons enna ienn E E N a a 12 a Continuous Variables used in Production Function b Categorical Variables Firm Level Characteristics c Worker Level Variables and Characteristics de Price Detlatorss UW Sed cessitsrt ceoted aa cote a AT AE T amuses 15 Bibliography casas estiewastuseied Ges ashaas Aleut asin ash ae ashe aaa 19 Acknowledgements This data was originally collected by a team from the Helsinki School of Economics in conjunction with the Tanzanian Confederation of Industries and the University of Dar es Salaam The surveys were undertaken as part of the Regional Program on Enterprise Development RPED organised by the Africa Technical Department of the World Bank We are particularly grateful to Xavier Blanc from Helsinki for making the data available for CSAE researchers in January 1999 in preparation for Wave 4 of the Tanzanian Manufacturing Enterprise Survey undertaken in late 1999 see separate dataset and documentation also on the CSAE website This
15. data has been cleaned and analysed as part of an ongoing CSAE research project into manufacturing sector performance in Tanzania and Ghana funded by the Department for International Development DFID Tanzania RPED Dataset User Guide CSAE University of Oxford 1 Introduction The objective of this user guide is to provide potential users with some basic information and explanations about the dataset which will hopefully facilitate their ability to use it for their own research or other purposes The relevant material is a comprehensive panel data set on a sample of firms within the Tanzanian manufacturing sector This data was collected over the period 1993 to 1995 in a series of three annual surveys referred to here as Waves I III as part of the Regional Program on Enterprise Development RPED organised by the World Bank and funded by bilateral donor governments The data was collected by a team from the Helsinki School of Economics in collaboration with the Tanzanian Confederation of Industries and the University of Dar es Salaam The timing and results of the three surveys were as follows Wave of Timing Firm Data For No Firms No Workers Survey Wave 1 Autumn 1993 1992 217 firms 1086 Aug Oct Wave 2 Autumn 1994 1993 213 firms 653 Oct Dec 40 replacements Wave 3 Early 1996 1994 1995 152 firms 342 Feb Mar no replacements 1996 Large firms with written accounts generally report
16. er 1999 Harding Alan and Francis Teal 2000 Firm Survival Growth and Productivity Tanzanian Manufacturing in the 1990 s mimeo CSAE University of Oxford October 2000 Helsinki School of Economics 1995 Dynamics of Enterprise Development in Tanzania Final Report on the Round 2 Survey Data Helsinki July 1995 Ndulu Benno and Joseph Semboja 1992 Trade and Industrialisation in Tanzania A Review of Experience and Issues in Trade Policy Industrialization and Development New Perspectives G K Helleiner ed Oxford Clarendon Press pp 515 53 Wangwe Samuel F Musonda and J Kweka 1997 Policies for Manufacturing Competitiveness The Case of Tanzania Working Paper Economic and Social Research Foundation ESRF Dar es Salaam 19
17. er sectoral inflation differentials we also undertook analysis using the food and non food components of the CPI index for firms in the relevant sectors It can be observed that food prices have risen faster in Tanzania since 1992 than prices of a basket of non food products Available producer price data shows that the rate of increase of producer prices has been below the CPI changes for this period Hence the use of the CPI as a price deflator will have introduced an artificial downward bias into our calculations of real output and value added for the later years Producer Price Deflators More recently we have used 45 producer price series at the 4 digit ISIC level as a set of deflators for firms real output and value added This price data was obtained from the National Bureau of Statistics NBS in Dar es Salaam and is based upon price indices taken from returns to their Quarterly Survey of Industrial Production QSIP 15 Tanzania RPED Dataset User Guide CSAE University of Oxford This producer price index was last published in 1996 but has now been updated to June 1999 These indices are presented in Table 2 below There are firms in our survey which fall within 4 digit ISIC product groups for which there is no price series available in the NBS indices presumably because there are no firms in their sample producing these products One example of this is the lack of a price index for furniture ISIC 3320 in the NBS data
18. ermined by asking firms how much additional output they could produce with no additional investment in plant and equipment Note the original data was not collected on a consistent basis across all three waves Export Dummy EXPDUM Dummy for whether firm exports Additional information is available for some waves as to the proportion of output exported to African countries EXPAF and the proportion outside of Africa EXPNAF Care needs to be taken in using these variables as the questions asked differ between waves Derived Variables Capital Labour Ratio CAPEMP Labour Productivity measure VADEMP Return on Capital Employed VADCAP Constant price series Output OUTCP Value added VADCP Capital stock CAPCP 12 Tanzania RPED Dataset User Guide CSAE University of Oxford b Categorical Variables Firm Level Characteristics Location dummies DSM Dummy for firms based in Dar es Salaam MOROG Dummy for firms based in Morogoro TANGA Dummy for firms based in Tanga ARUSH Dummy for firms based in Arusha Moshi MWANZ Dummy for firms based in Mwanza IRINGA Dummy for firms based in Iringa Njombe CAPCITY Dummy for firms based in capital i e Dar es Salaam Sector dummies 4 main sectors and disaggregated The survey covers four main manufacturing sectors food and beverages FOOD textiles and garments TEXT wood processing and furniture WOOD and fabricated metal and machinery METAL
19. estion 10a Note that not all of the variables have been manually coded in the questionnaires since fully coded versions were not available to us However it is relatively simple to track down variables by finding the question you are interested in and using the variable name structure outlined above to search for the relevant variable in the data file A significant proportion of the thousands of variables have been labeled in the data files which further assists in this process Firms are identified by a firm ID number firm which remains constant across the waves of the survey Firm numbers run in series based upon firm location Dar es Salaam 1 199 Arusha Moshi 201 299 Iringa Njombe 301 399 Morogoro 401 499 Mwanza 501 599 Tanga 601 699 In the worker files individual workers are identified by an ID number called either wnb in the SPSS data files or renamed WID in the SAS program files ranging between 10 which together with the firm ID number can be used to identify them and manipulate this data Note that the wnb WID is missing in some of the uncleaned worker files Tanzania RPED Dataset User Guide CSAE University of Oxford In each firm level data file there are variables which identify each firm s sector and location There is an additional file called isic por which gives 4 digit ISIC codes for all of the firms interviewed in the first 4 waves of the Tanzania survey However in the
20. her sector locati dn waer ship refinements to some of these variables legal status unionisation export status and firm age humancap sas creates average human capital humcap sd2 also creates owner manager education variables by firm including firm3 sd2 and experience variables in cd firm3 worker age tenure education and previous work experience invest sas investment in land buildings invest sd2 also contains additional variables for and equipment over 3 waves rate of capacity utilisation purpose investment capital ratios and and ype Oba esencat sources of investment financing finance sas firm s involvement in financial finance sd2 markets borrowing and lending total net indebtedness sources of firm financing worker1 sas series of SAS programs which work1 sd2 see additional explanatory notes on the create key worker level merge12 sd2 worker data in the text workpool sas workmerge sas variables across the 3 waves merge23 sd2 also contains povled ear ings functions i and descriptive statistics total of 1990 including occupational work 123 sd2 observations in the pooled dataset workpool sas categories skilled unskilled ed work123 and earnings Tanzania RPED Dataset User Guide CSAE University of Oxford 3 Raw Data Files The data files are in SPSS portable por format Data analysis has been undertaken using SAS hence it will be necessary to convert the SPSS files into SAS data files sd2
21. ivate foreign owners only PRIVDF Dummy for Private foreign and domestic owners SPRIVDOM Dummy for Private domestic owners amp some state ownership SPRIVFOR Dummy for Private foreign owners amp some state ownership SPRIVDF Dummy for Private foreign and domestic owners amp some state ownership PCFOR Percentage of foreign ownership 13 Tanzania RPED Dataset User Guide CSAE University of Oxford Legal Status of Firm SOLO Dummy for sole trader PARTNER Dummy for partnership LLE Dummy for limited liability enterprise PRIVCORP Dummy for private corporation STATCORP Dummy for state corporation COOP Dummy 1 for cooperative SUBDOM Dummy for subsidiary of Tanzanian firm SUBFOR Dummy for subsidiary of foreign firm Owners Ethnicity AFRICAN Dummy for African owners ASIAN Dummy for Asian Indian owners MIDEAST Dummy for owners of Middle East Arabian origin OTHER Dummy for other ethnicity including European and Chinese Firm Age Dummies created using the continuous variable FIRMAGE which is based on the year in which the firm first commenced operations STYEAR OLD Dummy for firms gt 20 years MATURE Dummy for firms 11 20 years inclusive YOUNG Dummy for firms 6 10 years inclusive NEW Dummy for firms lt 6 years c Worker Level Variables amp Characteristics AGESCH Age left school YRENDSC Year left school EDUC Number of years of education 0 16 years calculated
22. mp hardware 100 116 165 241 272 276 278 269 3812 Metal Furniture 100 116 165 241 272 276 278 269 3813 Metal structures 100 119 154 222 282 292 316 319 3819 Fabricated metal products 100 119 154 222 282 292 316 319 3821 Engines amp Turbines 100 110 132 159 169 160 160 159 3822 Agric Machinery 100 105 120 143 241 235 251 251 3823 Metal amp wood machinery 100 105 120 143 241 235 251 251 3824 Industrial Machinery 100 105 120 143 241 235 251 251 3829 Other machinery 100 105 120 143 241 235 251 251 3831 Electrical machinery 100 110 132 159 169 160 160 159 3833 Electric appliances 100 110 132 159 169 160 160 159 3839 Other Electrical mach 100 110 132 159 169 160 160 159 3843 Motor vehicles 100 115 140 156 151 152 152 162 3844 Bicycles amp motorcycles 100 115 140 156 151 152 152 162 3849 Transport equipment 100 115 140 156 151 152 152 162 Source NBS Producer Price Indices unpublished data 1996 99 Tanzania RPED Dataset User Guide CSAE University of Oxford Table 3 Alternative Capital Stock Deflators ER CPI Cap Defl 1 Cap Defl 2 Cap Defl 3 0 5 ER 0 5 CPI 0 8 ER 0 2 CPI 0 2 ER 0 8 CPI 1988 99 29 56 52 77 91 1 00 90 74 1 00 65 07 1 00 1989 143 38 73 62 108 50 1 39 129 43 1 43 87 57 1 35 1990 195 06 100 00 147 53 1 89 176 05 1 94 119 01 1 83 1991 219 16 128 70 173 93 2 23 201 07 2 22 146 79 2 26 1992 297 71 156 80 227 26 2 92 269 53 2 97 184 98 2 84 1993 405 27 196 40 300 84 3 86 363 50 4 01 238 17 3 66 1994 509 63 261 40
23. nstitute a panel which is intended to be broadly representative of the size distribution of firms across the major sectors of Tanzanian manufacturing industry These sectors include food processing textiles and garments wood products and furniture metal products and machinery It has been possible to obtain data over the three waves for the majority of the original sample of firms However in Wave 2 40 of the firms that dropped out of the survey for a variety of reasons were replaced in the sample by similar firms A fourth wave of the Tanzania Manufacturing Enterprise Survey TMES was conducted in late 1999 involving CSAE and the Economic and Social Research Foundation ESRF in Dar es Salaam This succeeded in revisiting 89 of the remaining RPED firms and interviewing an additional 106 replacement firms Similar data was obtained for 1996 1997 and 1998 thus giving a maximum number of six observation per firm This data is also available on the CSAE website see below A fifth wave of the TMES is currently being conducted by a team from CSAE and ESRF January February 2002 and will seek to revisit all surviving firms from the Wave 4 sample and gather comparable data for 1999 2000 and 2001 thus giving a rich panel dataset covering an entire decade The first four waves of this dataset are being made available to potential users in two ways firstly they will be added to the collection of datasets held by The Data Archive at the Univer
24. r we have used is a weighted average of the national CPI weight 0 5 and the nominal US dollar exchange rate weight 0 5 We have some evidence from the producer price series for domestically produced machinery that capital goods prices have risen in line with changes in the CPI A comparison of alternative deflators is presented in Table 3 Exchange Rates Table 4 shows trends in the nominal exchange rate of the Tanzanian Shilling against the US dollar the benchmark currency for cross country comparative work to date There has been a substantial devaluation since 1992 although the major nominal devaluation took place from 1988 92 with the move from a fixed to floating rate mechanism We also calculate a simple real exchange rate measure which suggests an appreciation over the survey period due to the high levels of domestic price inflation This will have served to make production for the domestic market more attractive in comparison to export markets 16 Table 2 Producer Price Series by 4 Digit ISIC Categories ISIC Activity 1992 1993 1994 1995 1996 1997 1998 1999 3111 Meat products 100 114 152 201 214 230 239 254 3112 Dairy products 100 113 145 174 251 321 331 304 3113 Fruit amp Veg Canning 100 90 119 142 134 155 146 160 3114 Fish amp sea products 100 114 152 201 243 310 337 354 3115 Vegetable oils amp fats 100 119 127 161 195 225 260 286 3116 Grain Mill products 100 110 129 168 175 165 166 178 3117 Bakery
25. s grouping questions related to different aspects of firm level structure and performance and also a section of supplementary labour market information gathered from interviews with a sample of up to 10 workers within each firm Interviewers were instructed to select workers from the range of occupational categories within each firm in order to attempt to capture the range of earnings and human capital characteristics of the workforce These sections are organised as follows in the Wave HI Tanzania questionnaire Questionnaire Sections Wave III Firm Identification amp Contact Details Visit Details Entrepreneurship General Firm Competition Technology Labour Markets Training Financial Markets amp Contractual Relations Infrastructure 9 Regulation 10 Adjustment 11 Investor Confidence 12 Business Support Services Oo TATO ol E a Appendix to Labour Markets Section Survey for a Sub Sample of Workers amp Apprentices Variables within the data files e g output employment exports earnings can be identified and extracted by referring to the unique variable code used in the data entry process See notes on variable coding structure in Section 3 above Please note that information on the actual names of firms amp persons interviewed and their contact details have been excluded from the data files made available here Tanzania RPED Dataset User Guide CSAE University of Oxford 5 Basic Data Anal
26. sity of Essex tel 01206 872001 secondly they can be downloaded from the CSAE s website at the following address www csae ox ac uk data Details of the sampling process for the RPED surveys can be found in Blanc 1997 Report on the Third Wave of the RPED survey in Tanzania Tanzania RPED Dataset User Guide CSAE University of Oxford 2 List of Files amp their Contents The following files and folders are included in this release with a brief description of their contents Please refer to subsequent sections of this user guide for a more detailed description of the files how they link together and how they can be used Raw Data Wave folder tzl por all firm level data from Wave 1 original tz1 and cleaned tzlnew tzlnew por versions of files note some changes were made to these files after Wave 4 of survey hence they date from 03 00 wk1 por all worker level data from Wave 1 original wk1 and cleaned wk1new por wk1new versions of files Wave 2 folder tz2 por all firm level data from Wave 2 as above tz2new por wk2 por all worker level data from Wave 2 as above wk2new por Wave 3 folder tz3 por all firm level data from Wave 3 as above tz3new por wk3 por all worker level data from Wave 3 as above wk3new por tz123 por this combined data file created by Helsinki School of Economics contains some additional variables which are not available in the individu
27. vations lost due to lack of information on worker human capital variables we have undertaken a matching procedure using all available information The steps in this rather convoluted process are contained in the SAS program files worker1 sas workmerge sas workpool sas Also please note that questions on worker allowances were not included in the Wave 1 questionnaire In order to allow comparability of earnings wages plus allowances over waves we have used information on aggregate allowances paid in Section 5 of the firm questionnaire and estimated average monthly allowances per employee This method unfortunately does not allow us to identify differences in allowances between workers in different occupational categories within the firm 11 Tanzania RPED Dataset User Guide CSAE University of Oxford 6 Variable Definitions a Continuous Variables used in Production Functions Output OUTPUT Value of firm s total production during previous year Values are annualised from monthly or weekly figures reported by some firms Note that where there is a missing observation output is set as the value of firm s total sales in the previous year SALES In Wave 1 we have chosen to use sales figures instead of output as there were many missing observations on output and the sales figure is more consistent Raw Material Costs RAWMAT Cost of raw materials used in producing firm output per period Note that this differs from r
28. ve 1 of survey firm genfirm1 sd2 production data revisions made to outputs and inputs capital components of Value Added VAD k l i d and capital stock CAP in light of stock employment etc an sets better information in Wave 4 survey up size and sector dummies genfirm2 sas organises general firm variables w2chars sd2 some firms excluded due to unreliable from Wave 2 of survey size and genfirm2 sd2 production data revisions made to escror dummies components of Value Added VAD and capital stock CAP in light of better information in Wave 4 survey genfirm3 sas organises general firm variables w3chars sd2 some firms excluded due to unreliable from Wave 3 of survey size and genfirm3 sd2 production data revisions made to sector dummies components of Value Added VAD and capital stock CAP in light of better information in Wave 4 survey mergel23 sas pools 3 waves of general firm pooll sd2 contains a total of 519 observations data calculation of constant out of theoretical maximum of 582 ie fi observations given no exclusions over price US and USSppp series 3 waves produces basic descriptive statistics and pooled production functions ownership sas defines dummy variables for owners sd2 firm ownership domestic foreign state private legal status age and unionisation firmchars sas combines firm characteristics for fchars1 sd2 incorporates the variables created in all 3 waves including size ownership sas and makes furt
29. y to modify the library names given at the top of each program which reference the paths of the directory in which the respective raw data files are stored At the moment the reference is to a directory path d tanzrped Each of the program files contains a series of data steps which in most cases organise the various waves of the data pool the three waves and create a whole range of variables of interest to the researcher some of the main ones are listed below At the end of each program a permanent data file is created which retains the main variables generated for further analysis These constructed SAS datasets with file extension sd2 are included in the SAS Data folder A list of some of the main variables in these data files and the methods used in their construction is included in Section 6 of this guide Additional Note on Worker Data Use of the worker data for Waves 1 3 is complicated by a flaw in the questionnaire design In Wave 1 each worker interviewed was assigned a worker ID number WID but their names were not recorded Questions were asked about time invariant gt Where subsequent information indicated clearly that erroneous data had been entered for specific firms changes were made in the original data files new Some additional revisions to components of value added and the capital stock mainly designed to ensure consistency between waves have been undertaken for a number of firms in the SAS programming
30. ysis As already mentioned over the three waves of the survey the structure and contents of the questionnaire have been modified Hence while the data collected over the three years is broadly comparable there are some differences which have subsequently entailed substantial work to improve the data s organisation and derive consistent variables across the three waves It is also important to take into account various differences in reporting conventions between firms in the sample for example firms which report their production and sales data for different time periods i e weeks months years This work of organising and cleaning the raw data in order to ensure that it can be used successfully in variable identification and analysis has been undertaken using the SAS programming language Further revisions to some of the variables for some firms have been undertaken in the light of information obtained in Wave 4 of the survey We also discovered that some of the firms originally interviewed were not manufacturing firms but trading or service firms and hence these have been excluded from subsequent analysis As a result of this work standardised variables across the three waves have been created in several SAS program files with file extension sas and have been broadly separated according to the types of variables that these files generate In order to be able to run these programs on the data it will of course be necessar
Download Pdf Manuals
Related Search