Thesis (MAgricAdmin (Agricultural Economics))--University of Stellenbosch, 2005. / The wine industry is one of the oldest commercial activities in South Africa. The South African wine grape industry annually produces more than a million tonnes of grapes, making the country the ninth largest producer in the world. The total area under wine grape production is divided into eight regions for administrative purposes. These boundaries are a legacy of the era of controlled marketing and there is continued meaningfulness of having various classifications such as ‘wine of origin’ scheme. The purpose of this study is to identify the factors that determine the producers’ choice of wine grape cultivars in the wine regions in South Africa. Time series data for the period 1990-2003 were used to estimate the parameters of linear regression models. Two equations for each wine grape cultivar in each region were postulated and estimated using Ordinary Least Squares as applied with Eviews. Further, a stepwise regression as applied in STATISTICA was used to eliminate the parameters that were not statistically significant at five percent significant level. In identifying the factors that determine the choice of wine grape cultivars in the regions, the results showed that each wine grape cultivar in each region has its own factors influencing the producers’ choice of that specific wine grape cultivar. Same wine grape cultivars in different regions similarly have its own factors determining the producers’ choice. The implication of this is that there are differences in terms of the requirements and types of crops and wine grape cultivars grown in each region. However, the most important result that emerged with regular frequency is that, the factors determining the producers’ choice of a specific wine grape cultivar for each region is price of other wine grape cultivars and competitive products in that wine region. The price of specific wine grape cultivars only had an influence on few wine grape cultivars. The implication is that the producers in South Africa appears to consider the prices of other wine grape cultivars and competitive products before making a choice of whether to plant or uproot a specific wine grape cultivar more than the price of the specific wine grape cultivar. This supports the theory that farm prices play a key role in allocating resources and in rewarding efficient producers.

Thesis (MComm (Statistics and Actuarial Science))--University of Stellenbosch, 2006. / In recent times many theories have surfaced posing challenging threats to the Efficient Market Hypothesis. We are entering an exciting era of financial economics fueled by the urge to have a better understanding of the intricate workings of financial markets. Many studies are emerging that investigate the relationship between stock market predictability and efficiency. This paper studies the existence of calendar-based patterns in equity returns, price momentum and earnings momentum in the South African securities market. These phenomena are commonly referred to in the literature as security market imperfections, financial market puzzles and market anomalies. We provide evidence that suggests that they do exist in the South African context, which is consistent with findings in various international markets. A vast number of papers on the subject exist in the international arena. However, very few empirical studies on the South African market can be found in the public domain. We aim to contribute to the literature by investigating the South African case.

Thesis (MBA)--Stellenbosch University, 2008. / ENGLISH ABSTRACT: Repurchasing of shares by South African companies were legalised on 1 July 1999. This introduced an alternative to dividends for distributing cash to shareholders. Although dividends and share repurchases realise the same value in a perfect efficient market, the inefficiencies of the South African stock market require managers to carefully evaluate factors like taxation and stock price valuation when selecting appropriate distribution methods. This research report aims to update shareholder distribution trends for industrial JSE listed companies over the past 10 years in order to determine the impact of share repurchases on dividend payouts. Furthermore, this research report examines the factors that may have had an impact on shareholder distribution choices in order to provide some guidelines for choosing appropriate distribution methods. An initial analysis of SENS share repurchase announcements revealed that 121 JSE listed companies repurchased about R50 billion worth of shares up to 30 June 2007. The bulk of the shares, 65% by value, were repurchased on the open market, while 35% was repurchased through specific fixed price offers. However, a comparison of accurate share repurchase data obtained from a sample of company annual reports, indicate that repurchase announcements understate actual repurchases by more than 20% on average. Further analysis of distribution trends were therefore based on actual repurchase data published in annual reports rather than SENS announcements. After the legalisation of share repurchases in South Africa, a decline in dividend paying companies was expected similar to that experienced by the United States since the 80's. However, a detailed analysis of 132 industrial listed companies indicated that the proportion of dividend paying companies increased from a level of 50% to almost 75% since the introduction of share repurchases. On the other hand, the proportion of companies repurchasing shares initially rose to over 25%, but then declined to below 20% by 2007. Ordinary dividends are the dominant shareholder distribution choice with 64% of companies opting for this method. Open market share repurchases have been well adopted with 17% of companies using this method, while only 5% and 4% of companies using special dividends or specific repurchases respectively. Dividends paid out of share premium (capital distributions) have also emerged as a favourite over recent years with almost 20% of companies using this shareholder distribution method. Current tax legislation do not provide all the advantages usually enjoyed by share repurchases internationally and have largely prevented dividends from being substituted by share repurchases. The decline in share repurchases up to 2007 also indicates that share repurchases become less effective as share prices increase to overvalued levels. While tax implications and stock price valuation remain the dominant determinants of shareholder distribution choice, this study shows that shareholder diversity, dividend preferences, size of distribution, and BEE requirements also have significant influences on the choice of distribution method in the South African context. / AFRIKAANSE OPSOMMING: Die terugkoop van aandele deur Suid-Afrikaanse maatskappye is wettig sedert 1 Julie 1999. Dit het 'n alternatief tot dividende in werking gestel om kontant aan aandeelhouers uit te keer. Alhoewel dividende en aandele-terugkoop dieselfde waarde in 'n perfekte doeltreffende mark realiseer, vereis die tekortkominge van die Suid-Afrikaanse aandelemark dat bestuurders faktore soos belasting en aandeelpryswaardasie versigtig moet oorweeg tydens die keuse van geskikte uitkeringsmetodes. Die doelwit van hierdie navorsingsverslag is om die tendense van uitkerings aan aandeelhouers te hersien vir industriele JSE-genoteerde maatskappye oor die laaste 10 jaar om sodoende die effek van aandele-terugkope op dividenduitbetalings te bepaal. Verder ondersoek hierdie navorsingsverslag ook die faktore wat moonlik 'n invloed op aandeelhouers-uitkeringskeuses gehad het, om sodoende riglyne vir die keuse van geskikte uitkeringsmetodes saam te stel. 'n Voorlopige analise van SENS-terugkoopaankondigings toon dat 121 JSE-genoteerde maatskappye ongeveer R50 miljard se aandele teruggekoop het tot en met 30 Junie 2007. Die grootste gedeelte van hierdie aandele, 65% se waarde, is op die ope mark teruggekoop terwyl 35% deur spesifieke vasteprys terugkope verkry is. 'n Vergelyking met terugkoopsyfers wat uit 'n steekproef van maatskappyjaarverslae geneem is, dui egter daarop dat aankondigings die ware terugkope met gemiddeld 20% onderskat. Verdere ontleding van aandeelhouers-uitkeringstendense word derhalwe gebaseer op syfers wat in jaarverslae gepubliseer is, eerder as SENS-aankondigings. Na die wettiging van aandele-terugkoop in Suid-Afrika, is verwag dat dividenduitbetalings sou daal soortgelyk aan dit wat in die Verenigde State ondervind is sedert die 80's. Die ondersoek van 132 genoteerde industriele maatskappye toon egter dat die persentasie van maatskappye wat dividende betaal van 50% tot bykans 75% toegeneem het sedert aandele-terugkoop 'n beskikbare opsie is. In teenstelling hiermee, het die persentasie maatskappye wat aandele terugkoop aanvanklik tot 25% gestyg, maar sedertdien afgeneem tot onder 20% teen 2007. Gewone dividende is die gewildste aandeelhouers-uitkeringsmetode met 64% van maatskappye wat van hierdie metode gebruik maak. Aandele-terugkope op die ope mark is goed verteenwoordig met 17% van maatskappye wat van hierdie metode gebruik gemaak het, terwyl slegs 5% en 4% van maatskappye onderskeidelik van spesiale dividende en spesifieke aandele-terugkope gebruik gemaak het. Dividende uit aandelepremie (kapitaaluitkerings) het ook na vore getree as 'n gunsteling keuse in die laaste paar jaar met bykans 20% van maatskappye wat hierdie uitkeringsmetode gebruik het. Huidige belastingswetgewing bied nie al die belastingvoordele aan aandele-terugkope wat normaalweg deur internasionale maatskappye benut word nie en het grotendeels verhoed dat dividende deur aandele-terugkoop vervang is. Die afname in aandeleterugkope tot en met 2007 is ook 'n aanduiding dat dit minder effektief raak soos wat aandeelpryse oor gewaardeerde vlakke styg. Terwyl belasting-oorwegings en aandeelpryswaardasies steeds die dominante drywers van aandeelhouersuitkeringskeuses bly, bevind hierdie studie dat faktore soos aandeelhouers se diversiteit, dividendvoorkeure, grootte van uitkerings, en vereistes van swart ekonomiese bemagtiging ook 'n noemenswaardige invloed op uitkeringskeuses binne die Suid-Afrikaanse konteks het.

Thesis (MBA)--Stellenbosch University, 2011. / This study analyses the effects of generic medicine competition on the market share growth and pricing of originator brand medicine in the South African private pharmaceutical market. The study is based on five years (2005 to 2011) of IMS Health market share data for 39 originator brand drugs that have been exposed to competition from generic substitutes from 2001. The results show that, for all the drug molecules included in the study pooled together, the price of an originator brand medicine relative to the weighted average price of its generics has a significant negative impact on the change of its market share. Results for the molecules pooled according to anatomical classes, as well as each molecule separately, show that in four out of the nine classes represented in the study and nine out of the 39 molecules the relative price of the originator brand medicine had a significant negative impact on its change in market share. The manufacturers and marketers of generic medicines would be well advised to offer their medicines at significantly discounted prices compared to the originator brands, as the results suggest that the market penetration of the generic product may depend heavily on the price the generics are offered at. Investigations into the prices of the originator brands in relation with the number of generic equivalents in the market show that the number of generics available in a specific market has a significant positive impact on the relative price of originators, thereby making originators relatively more expensive compared with their generic competitors, while at the same time the results show that the absolute price of the originator brand medicines declines as the number of generic equivalents in the market increases. This indicates that, from a policy perspective, reducing the barriers to entry for generic medicine once originator patents expire may have a significant role to play in reducing the cost of pharmaceutical drugs in the South African market.

Thesis (MBA)--Stellenbosch University, 2002. / ENGLISH ABSTRACT: A unique data set consisting of economists' expectations on key economic indicators was examined within the context of the controversial normal backwardation theory of Keynes. The economists' expectations were regarded as the expected future spot price and the relationship between them and the corresponding futures contracts was analysed. The respective economic indicators were: i) the yield from aparastatal Bond, ii) the yield from Government Bonds, iii) the rate of the 90 day Banker's Acceptance (BA) Deposit Rate and iv) the Rand/Dollar (R/$) Exchange Rate for the past seven years, i.e. 1995 to 2001. The accuracy of the economists' predictions was tested both on a visual basis and the relationship between the expected values and the futures prices was plotted in a graphical format. A nonparametric statistical procedure was used to determine whether the economists' expectations were of any value. To put it differently, the question being posed is: do these economists, as a group, possess some superior forecasting skills? Two different conclusions were reached from the analysis: First conclusion: by accepting the normal backwardation theory, it implies that the contango theory also holds. Therefore, when analysing the data set visually - depending on which theory it supports - the futures price must trade consistently below or above the expected future spot price. For this particular analysis the yield of the bond, and not its price, was the important factor. In most cases the plotted relationships between the expected values and the futures prices were found to support the contango theory and, to a lesser extent, the normal backwardation theory. Hence, speculators were, in order to make profits, predominately sellers of futures contracts. Second conclusion: the strongest conclusion, however, follows from the statistical tests conducted on the expected values. It was found that economists do possess some superior forecasting skills and if they had used their predictions and had taken the corresponding market positions, they would have been consistent winners in the futures market. Their reward would be mainly for their ability to forecast eventual spot prices and, to a lesser extent, for their risk bearing. It was impossible to link the two conclusions to confirm the normal backwardation theory, for the particular South African data set. The evidence is thus consistent with the hypothesis that the futures price is an unbiased estimate of the expected future spot price. / AFRIKAANSE OPSOMMING: 'n Unieke datastel, bestaande uit ekonome se vooruitsigte van kern ekonomiese aanwysers, is ondersoek binne die konteks van die omstrede normale terugwaardasie-teorie (d.i. "normal backwardation theory") van Keynes. Die ekonome se vooruitsigte is aanvaar as die verwagte toekomstige kontantprys en die verhouding hiertussen en die ooreenstemmende termynpryse is ontleed. Die onderskeie ekonomiese aanwysers was: i) die opbrengs op 'n Semi-Staatseffek, ii) die opbrengs op Staatseffekte, iii) die koers van die negentig-dae-Bankaksepte (BA) Depositokoers en iv) die Rand/Dollar (R/$) Wisselkoers oor die afgelope sewe jaar, d.w.s. 1995 tot 2001. Die akkuraatheid van die ekonome se vooruitskattings is op 'n visuele basis vergelyk, en die verhouding tussen die verwagte prys en die termynpryse is in grafiese formaat gekarteer. 'n Nie-parametriese statistiese prosedure is gebruik om vas te stel of hierdie ekonome se vooruitsigte van enige waarde was. Anders gestel, die vraag is: beskik hierdie ekonome as 'n groep oor sekere superieure vooruitskattingsvaardighede? Die volgende twee afsonderlike gevolgtrekkings is geformuleer: Eerste gevolgtrekking: deur die normale terugwaardasie-teorie te aanvaar, impliseer dit dat die contango-teorie (d.i, "contango theory") ook geldig is. Dus, wanneer die datastel visueel getoets word - afhangende van watter teorie dit ondersteun - moet die termynprys konsekwent bo of onder die verwagte toekomstige kontantprys verhandel. Vir hierdie bepaalde analise was die opbrengs van die staatseffek die belangrike faktor en nié die prys daarvan nie. In die meeste gevalle het die gekarteerde verhouding tussen die verwagte prys en die termynprys getoon dat dit die contango-teorie ondersteun het en, in 'n mindere mate, die normale terugwaardasie-teorie. Derhalwe was spekulante, ten einde wins te maak, oorwegend die verkopers van termynkontrakte. Tweede gevolgtrekking: die belangrikste gevolgtrekking volg egter uit die statistiese toetse wat uitgevoer is op die verwagte pryse. Daar is bevind dat ekonome wel oor superieure vooruitskattingsvaardighede beskik en dat, indien hulle hul vooruitskattings gebruik en die ooreenstemmende markposisies ingeneem het, hulle konsekwent wenners in die termynmark sou gewees het. Hulle vergoedings sou hoofsaaklik gewees het vir hulle vermoë om uiteindelike kontantpryse te voorspel en, in 'n mindere mate, vir hulle risiko-blootstelling. Dit was onmoontlik om hierdie twee vergelykings met mekaar te verbind om sodoende die normale terugwaardasie-teorie te onderskryf vir die betrokke Suid-Afrikaanse datastel. Die bewyslewering is dus konsekwent met die hipotese dat die termynprys 'n onsydige skatting van die verwagte toekomstige kontantprys is.

This thesis is an implicit farmland value study which explores the possible effects of agricultural commodity prices, interest rate and land reform issues on farmland values. The study examines the impacts of these fundamental factors (interest rates and returns to farmland as determined by crop prices) on sugar cane farmland values, maize farmland values, on deciduous fruit (apples and pears) farmland values, and on aggregate South African farmland values. Expectations are that land reform influences the demand for farmland. Since farmland prices are demand driven, changes in the demand for farmland (as influenced by land reform issues) may result in changes in farmland prices. The study thus seeks to empirically examine, to a larger extent, the long-run influence of endogenous factors on farmland prices. Causes of cyclical behaviour in farmland prices are also examined. The study draws on cross-sectional and time series studies of previous research on farmland values. The maximum likelihood Johansen (1991) procedure of cointegration is used to estimate the relationship between fundamental factors and farmland values. The logit model is used to estimate the influence of land reform on the demand for farmland, hence farmland prices. Unit root and the Johansen cointegration test results proved that long-run relationships exist between farmland values and returns to farmland; the use of cointegration methods was thus recommended. Long-run changes in farmland prices are caused by fundamental factors. Short-run variations in farmland prices are caused by exogenous factors that affect net farm income and this lead to boom-bust cycles in farmland values. / Thesis (M.Comm.) - University of KwaZulu-Natal, Pietermaritzburg, 2009.

Thesis (MA (Geography and Environmental Studies))--University of Stellenbosch, 2010. / ENGLISH ABSTRACT: The research problem addressed in this study is how to ascertain the primary determinants of house prices in Hout Bay. This overarching aim encompasses three interwoven aspects. The research attempts first to determine which factors generally affect property prices in Hout Bay; second, to assess the extent to which individual factors affect house prices; and third, to discover the role variables collectively play in determining house prices in Hout Bay. Four objectives emerge from this subdivision of the aim, namely identify potential house priceinfluencing factors in Hout Bay; quantify the selected locational variables; statistically analyse the variables to distinguish the significant and insignificant ones; and use regression analysis to deduce the collective and individual influences of the significant factors on house prices. Structured interviews were conducted with representatives of 12 estate agencies in Hout Bay to uncover factors affecting the local property market. Through insights gleaned from the literature, manipulation of municipal valuation and cadastral data and the structured interviews, 39 structural and site-related variables, 18 distance variables and 11 socioeconomic variables were constructed. Several preliminary and descriptive analyses performed on the variables gave a general impression of the distribution of data and assisted in identifying statistically significant variables for determining house prices. These analyses included measures of central tendency (mean, median and mode); measures of dispersion (minimum and maximum values, range, standard deviation, skewness and kurtosis); the compilation of histograms for each variable; analysis of variance (ANOVA) on nominal data variables; and the creation of 2D scatterplots for ordinal data variables. Spearman rank order correlation was performed on the nominal and ordinal data variables. Statistically weak variables and those exhibiting signs of multicollinearity were eliminated. A best-subsets regression analysis was executed on the remaining variables. The regression model performed adequately, explaining close to 54% of the variation in house prices in Hout Bay. Among the individual factors, the size of the erf was the strongest predictor of the house price dependent variable, house size was the second most important factor, while distance to busy roads and quality of the house shared similar importance. Regression residuals were also mapped to expose spatial patterns. It is recommended that comparable research be conducted on a citywide scale, that variables be quantified differently and that new GIS techniques be incorporated in future studies. / AFRIKAANSE OPSOMMING: Die navorsingsprobleem wat hierdie studie aanspreek, is hoe om vas te stel wat die primêre faktore is wat huispryse in Houtbaai bepaal. Hierdie oorkoepelende doelwit vervat drie onderling verwante aspekte. Eerstens, poog die navorsing om te bepaal watter faktore in die algemeen huispryse in Houtbaai beïnvloed; tweedens, om te assesseer tot watter mate individuele faktore huispryse affekteer; en derdens, om te ontdek watter kollektiewe rol veranderlikes in die bepaling van huispryse in Houtbaai speel. Vanuit hierdie onderverdeling van die navorsingsdoelwit het vier doelstellings ontstaan, naamlik identifiseer die potensiële faktore wat huispryse in Houtbaai beïnvloed; kwantifiseer die geselekteerde liggingsveranderlikes; voer verskeie analises uit op die veranderlikes om die beduidende en onbeduidende veranderlikes te identifiseer; en benut regressie-analise om die kollektiewe en individuele invloed van beduidende faktore op huispryse in die studiegebied vas te stel. Gestruktureerde onderhoude is met verkoopslui van 12 eiendomsagentskappe in Houtbaai gevoer om die faktore te bepaal wat die plaaslike eiendomsmark beïnvloed. Deur middel van insigte verkry uit die akademiese literatuur, manipulasie van munisipale waardasie- en kadastrale data en die gestruktureerde onderhoude is 39 strukturele en liggingsverwante veranderlikes, 18 afstandsveranderlikes en 11 sosio-ekonomiese veranderlikes geskep. Verskeie analises wat op die veranderlikes uitgevoer is, het ‘n algemene indruk van die verspreiding van die data verskaf en het die identifisering van statistiesbeduidende veranderlikes bevorder. Hierdie analises het maatstawwe vir sentrale neiging (rekenkundige gemiddelde, mediaan en modus); maatstawwe vir dispersie (minimum en maksimum, variasiewydte, standaardafwyking, skeefheid en kurtose); die samestelling van histogramme vir elke veranderlike; die analise van variansie (ANOVA) op veranderlikes met nominale data; en die skep van 2D-spreidingstippe vir veranderlikes met ordinale data behels. Spearman se rangorde korrelasie is op beide die nominale en ordinale data uitgevoer. Statistiesonbeduidende veranderlikes, of dié wat tekens van multikollineariteit met ander veranderlikes getoon het, is geëlimineer. ‘n Beste deelversameling regressie-analise is uitgevoer op die oorblywende veranderlikes. Die regressiemodel het gepaste resultate behaal deurdat dit byna 54% van die variasie in Houtbaai se huispryse verklaar het. Van die individuele veranderlikes was die grootte van die erf die sterkste voorspeller van die huisprys afhanklike veranderlike, huisgrootte was die tweede belangrikste faktor, terwyl afstand van besige paaie en die kwaliteit van die huis soortgelyke invloed gedeel het. Die regressiemodel se residu’s is gekarteer om ruimtelike patrone vas te stel. Dit word aanbeveel dat soortgelyke navorsing op ‘n stadswye skaal uitgevoer word, dat die veranderlikes op ander wyses gekwantifiseer word en dat nuwe GIStegnieke in toekomstige studies aangewend word.

The study examined the impact of earnings announcements on the share price of selected mining companies using the most recent data from the Johannesburg Stock Exchange. This study covered a period from 1 January 2011; to 31 December 2015. Using the classical event study methodology, the speed of reaction of the market to annual earnings information releases for a sample of 27 companies listed on the exchange is tested. Over the sample period, the Abnormal Returns (AR), Average Abnormal Returns (AAR) and Cumulative Average Abnormal Returns (CAAR) were calculated. The AR, AAR and CAAR show positive results obtained during the earnings announcement period. The returns yielded from these results are significantly different from zero. / Financial Accounting / M. Phil. (Accounting Sciences)

The current study examines the extent and magnitude by which global and regional shocks are transmitted to the volatility of returns in the stock markets of South Africa, Egypt, Nigeria, Botswana, Mauritius and Egypt. This is done so as to make inferences on the level of the domestic market‟s integration into the regional and world capital markets. By applying multivariate and univariate GARCH models, using weekly data from June 1995 to May 2010, the main empirical findings are threefold. Firstly, the volatility analytical framework finds statistically significant and time-varying volatility spillover effects from the regional and global markets to the South African market. Global shocks are generally stronger and account for up to 23.9 percent of the volatility of South Africa‟s equity market compared to weaker regional factors which account for less than 1 percent of domestic variance. Only in countries with strong bilateral trade and economic links with South Africa, such as Botswana and Namibia, is it found that regional factors are more dominant than global factors for domestic volatility. Compared to the other African markets, the joint influence of foreign shocks on domestic volatility is highest in South Africa and Egypt, two of Africa‟s largest and most developed markets. The results further demonstrate that for all the African markets the explanatory power of both regional and global factors for domestic volatility is not constant over time and tends to increase during turbulent market periods. Secondly, the analysis of the determinants of South frica‟s second moment linkages with the global market suggests that the volatility of the exchange rate plays a cardinal role in influencing the magnitude by which global shocks affect domestic volatility. The increased global integration in the second moments cannot be attributed to either increased trade integration, convergence in inflation rates or to convergence in interest rates between South Africa and the global markets. Lastly, tests were conducted to examine whether there have been contagion effects from the regional and global markets to South Africa from the 1997 Asian crisis and the 2007/8 global financial crisis. The results show no evidence of contagion during either the East Asian currency crisis or the recent global financial crisis to South Africa, while some African markets, such as Egypt, Mauritius and Botswana, exhibit contagion effects from either crisis. Overall, the empirical findings generally support the view that African markets are segmented both at the regional and global levels as domestic volatility is more influenced by local idiosyncratic shocks (the proportion not attributable to either global and regional factors). However, the volatility of South Africa, and to a lesser extent Egypt, remains relatively more open to global influence. This implies that the potential for gains from international portfolio diversification and the scope for success of policies aimed at the stabilisation of equity markets in these markets exist.

This study analyses and investigates the impact of crude oil price vitality on unemployment in South Africa. This is done by firstly surveying theoretical and empirical literature on the crude oil price-unemployment relationship before relating it to South Africa. Secondly, crude oil and unemployment trends with their causes are overviewed. The study employs a Johansen co-integration technique based on VAR to model unemployment against crude oil prices, real effective exchange rate, real interest rates and real gross domestic product. Using quarterly data for the period 1990-2010, econometric results show that crude oil prices are positively related to unemployment in the long run while the opposite is true in the short run. Parameter estimates and variables are statistically significant; hence there are also policy recommendations which are related to both empirical and theoretical literature. Lastly, impulse response functions show that unemployment returns to equilibrium in the long run when crude oil price changes whereas real interest rates followed by crude oil prices explain most of unemployment changes compared to other variables in the long run.