Mini-study project (MBA)--University of Stellenbosch, 2003. / ENGLISH ABSTRACT: This study considers the importance of research and development (R&D) investment for enhanced productivity in the agricultural sector. In this regard, some detail of R&D investments will be discussed. Focus is given to research intensity ratios and government spending intensities in different parts of the world, for both developed and developing countries. Furthermore, the extent of focus on investment activities by the public and private sector, respectively, will be considered. Aspects of agricultural R&D in Africa, such as the crisis of funding, incentives to perform, priorities and procedures will be deliberated. Measures for reform will be addressed. Thereafter, the sustainability of investment in agricultural R&D will be emphasized. Roles and responsibilities of the public and private sectors, funding adequacy, funding trends and mechanisms of funding are discussed. Attention will be given to the impact public and agricultural policies have on both the level and nature of agricultural investment. Furthermore, R&D efforts may impact a wide audience and in this regard welfare implications for different interest groups are evaluated. Previous econometric studies are discussed. The relationship between R&D investment and productivity in the sugar industry in South Africa is studied for the period 1925-2001 within the context of a production function regression. Data were obtained from the South African Sugar Association Experiment Station. The dependent variable was sucrose yield, while the independent variables were rainfall, land used in sugar production, technology and production costs. The techniques of ordinary least squares and Ridge Regression, respectively, were used. Using ordinary least squares, it was determined that the length of the R&D lag was three years. The short lag length may be justified by the fact that research done by the Experiment Station is of an applied nature and research, training and extension are integrated by this commodity organisation. Furthermore, the education level of South African commercial farmers is relatively high. All the variables were economically relevant and statistically significant at the 1% level, except the land variable, which was statistically significant at the 5% level. In addition, it was established that there was a structural change in the model during the period concerned. The impact of technology on sucrose yield was higher during 1959-1979, than either before or after this period. Remedies and violations to the classic normal linear regression model were considered. In particular, there was evidence of multicollinearity, as indicated by the high condition index and variance inflation factors. Multicollinearity was addressed using the technique of Ridge Regression (k=O,15).The R2 determined for this model was 83,43%. All the t-statistics, except for land, were statistically significant at the 1% level. The t-statistics describing the technology and production cost variables were higher in the Ridge Regression model. However, the land variable proved to be unstable using Ridge Regression, as the sign changed. This made it difficult to draw firm conclusions about the contribution of this variable. It was concluded that the Ridge regression model provided the best fit to the data. Finally, the rate of return to R&D investment in the sugar industry in South Africa was calculated. Cost-benefit analysis yielded a benefiUcost value of 1,61, when allowing for direct spillover effects to millers. Furthermore, the export realisation price for sugar was used in the calculation. Costs included those of production, harvesting and land. Using the three-year lag determined in this study, the internal rate of return in the South African sugar industry was estimated as 17{20%. This exceeds the discount rate in South African agriculture, the latter that is equal to 6%. In conclusion, this study makes the argument that even if a commodity displays public good characteristics, there may still be incentives for the private sector to invest, especially if research benefits can be internalised in the industry. A prime example of this situation may be seen in the sugar industry in South Africa, which is privately-funded by industry levies. This concept can be stretched to include other industries. Furthermore, this study established that there was sufficient statistical evidence to conclude, that there is a relationship between investment in R&D and increased productivity in the sugar industry in South Africa, at the 1% level of significance. In addition, it was determined that these investments yielded a high . rate of return. / AFRIKAANSE OPSOMMING: Hierdie studie oorweeg die belangrikheid van investering in navorsing en ontwikkeling vir verbeterde produktiwiteit in die landbousektor. In hierdie verband, sal sekere aspekte van navorsing en ontwikkelingsbeleggings bespreek word. Fokus· sal gegee word aan navorsingsintensiteit verhoudings and regeringsbestedings intensiteite in verskillende dele van die wêreld, vir beide ontwikkelde en ontwikkelende lande. Verder, sal die mate waarin die publieke- en privaatsektore betrokke is by investering in navorsing en ontwikkeling, oorweeg word. Verskillende aspekte van navorsing en ontwikkeling in landbou in Afrika word bespreek. Voorbeelde hiervan is die krisis in befondsing, motivering om te presteer, prioriteite en prosedures. Stappe vir hervorming word oorweeg. Vervolgens, word die volhoubaarheid van investering in navorsing en ontwikkeling beskou. Daar is verskillende rolle en verantwoordelikhede vir beide die publieke- en privaatsektore in hierdie verband. Verder, lê aspekte van befondsing, soos die genoegsaamheid daarvan, verandering in maniere van befondsing en verskillende meganismes van befondsing, onder die riem. Fokus sal gegee word aan die impak van publieke - en landbouregulasies op beide die vlak en tipe landbou investering wat waargeneem word. Navorsing en ontwikkelingsaktiwiteite beïnvloed 'n wye verskeidenheid van belangegroepe. Die mate waarin hierdie groepe geaffekteer word, word oorweeg. Die verwantskap tussen investering in navorsing en ontwikkeling en produktiwiteit in die suikerindustrie in Suid-Afrika word vervolgens beskou. Vorige ekonometriese artikels word bespreek, waarna modelle van meervoudige lineêre regressie bepaal word. Data vir die periode 1925-2001, wat produksie in die suikerindustrie beskryf, is verkry van die Suid-Afrikaanse Suiker Assosiasie Eksperiment Stasie. Die afhanklike veranderlike is opbrengs van sukrose, terwyl die onafhanklike veranderlikes na reënval, land onder suikerproduksie, produksiekostes en tegnologie verwys. Die tegnieke van gewone kleinste kwadrate en rif regressie (Ridge Regression) is gebruik. Deur die tegniek van gewone kleinste kwadrate te gebruik, is dit bepaal dat die sloer voor die voordele van navorsing en ontwikkeling inskop, drie jaar is. Dat die sloer wat bepaal is so kort is, kan gemotiveer word deur die feit dat die navorsing wat gedoen word deur die Eksperiment Stasie, van 'n toegepaste aard is. Verder is navorsing, opleiding en ekstensie is geïntegreer. Die vlak van opvoeding van Suid- Afrikaanse kommersiële boere is ook relatief hoog. AI die veranderlikes was ekonomies relevant en statisties beduidend by die 1% vlak, behalwe die veranderlike wat land onder suikerproduksie beskryf, wat beduidend was by die 5% vlak. Verder is dit bepaal, dat daar 'n strukturele verandering in die model ingetree het. Die impak van tegnologie was groter tydens die 1959-1979 periode, as vóór of ná hierdie tydperk. Daar is verskeie aannames wat gemaak word met die gebruik van 'n lineêre regressie model. Daar was tekens van multikolineariteit, soos duidelik deur die hoë waardes vir die kondisie indeks en variansie inflasie indeks aangedui is. Multikolineariteit is aangespreek deur die tegniek van rif regressie (k=0,15) te gebruik. Die R2 waarde wat vir hierdie model bepaal is, is 83,43%. AI die t-waardes, behalwe dié van land, was statisties beduidend by die 1% vlak. Dié veranderlike se koëffisiëntleken het met rif regressie, verander van negatief na positief. Siende dat die veranderlike geblyk het om onstabiel te wees, was dit moeilik om ferm beslissings oor die bydrae daarvan te lewer. Wanneer die rif regressie model vergelyk word met die model wat strukturele verandering beskryf met gewone kleinste kwadrate, is die t-waardes wat tegnologie en produksiekoste beskryf, heelwat hoër in die eersgenoemde model. Dit was beslis dat die rif regressie model die beste passing vir die data gebied het. Vervolgens is die terugkeerkoers tot investering in navorsing en ontwikkeling vir die suikerindustrie in Suid-Afrika bereken. Die beraamde waarde vir die voordeel/koste van navorsing en ontwikkeling was 1,61. Die voordeel wat deur suikerindustriemeulenaars verkry is, is in berekening gebring. Verder is die uitvoer realisasieprys vir suiker gebruik. Die kostes het dié van produksie, oes en geleentheidskoste van land ingesluit. Die interne terugkeerkoers vir die suikerindustrie in Suid-Afrika is bereken as 17,20%, inaggenome die drie jaar sloer voor die voordele van navorsing en ontwikkeling realiseer. Hierdie waarde is heelwat hoër as die verdiskonteringskoers in die Suid-Afrikaanse landboubedryf, wat beraam word op 6%. Ter samevatting, in hierdie studie word geargumenteer dat al sou 'n kommoditeit eienskappe van openbare goedere toon, daar tog motivering is vir die privaatsektor om te investeer, veral as die voordele van navorsing en ontwikkeling in die industrie self ge-internaliseer word. 'n Welsprekende geval hiervan is die suikerindustrie in Suid-Afrika wat privaat befonds word. Hierdie beginsel kan gestrek word om ander industrieë in te sluit. Verder was daar in hierdie studie oorweldigende statistiese bewyse, by die 1% vlak van beduidendheid, van 'n verwantskap tussen investering in navorsing en ontwikkeling in die suikerindustrie in Suid-Afrika en verhoogde produktiwiteit. Vervolgens was dit bepaal dat hierdie investerings 'n goeie terugkeerkoers gelewer het.