However, drought, soil degradation caused by overgrazing, deforestation, excessive taxation, and insufficient infrastructure continue to plague these industries

1 CONTRIBUTION FORM Student ID Parts contributed Contribution % Nguyen Anh Thi S3915129 Part + 100% Nghiem Xuan Vu S3918159 Part 100% Vu Phuong Linh S3915216 Leader, Part + 100% Hoang Minh Thang S3920340 Support part 1+2+5 100% Tran Ho Viet S3915241 Part 1+2+6 100% Full name Signature Acknowledgement: We would like to thank Miss Trang - our lecturer in the course who deserves a particular thank you from all of us Her invaluable input and guidance helped shape our report and we wouldn't know as much as we now without her This paper would have had a lot more problems if she hadn't been there Sincerely, Team 04 Table of Contents I INTRODUCTION .4 II DESCRIPTIVE ANALYSIS III MULTIPLE REGRESSION IV REGRESSION CONCLUSION 15 V: TIME SERIES AND CONCLUSION 16 VI OVERALL CONCLUSION 21 VII APPENDICES 24 VIII REFERENCES LIST 36 I INTRODUCTION a Overview Agriculture contributes significantly to Ethiopia's economy and GDP, accounting for 83.9% of exports and providing 80% of total employment (worldometers n.d.) Most of the country's output is consumed domestically, and a sizable portion of its commodity exports come from a small agricultural cash-crop sector (J Abbink & Jean-Nicolas Bach 2017) Ethiopia's agriculture industry has grown rapidly over the previous few decades, becoming a major contribution to poverty reduction (Bachewe et.al 2014) However, drought, soil degradation caused by overgrazing, deforestation, excessive taxation, and insufficient infrastructure continue to plague these industries (J Abbink & Jean-Nicolas Bach 2017) Agriculture is a key activity for Nigeria's economy after oil, employing approximately 35% of the population as of 2020 (Inegbedion et al 2019) The sector is divided into four subsectors: crop production, livestock production, forestry, and fishing, however, livestock and crop farming are still considered as the backbone of the industry (Onesmus Semalulu et al 2020) Nigeria’s agriculture is still expected to grow in the future as of the increase in demand from an increasing population (Statista 2021) However, it still must face a number of challenges, including an outdated land tenure system that restricts access to land, a low level of irrigation development, slow adoption of technologies, high input costs, and insufficient storage facilities (Inegbedion et al 2019) b Data collection Nigerian and Ethiopian data used in this report originate entirely from genuine and reputable sources and range the years 1997 to 2018: Variables Sources AAGR% of GDP) (Use agriculture, forestry, and fishing, value added (% of GDP from world bank indicator) GDP per capita growth rate (annual %) Foreign direct investment, net inflow (% of GDP) Employment in employment) Agriculture (% of World Bank (n.d) total Agricultural raw materials exports (% of merchandise exports) Inflation, consumer prices (annual %) Average rainfall (per year) Average temperature (per year) CO2 emission (metric tons per capita) Urban population (% of total population) Table Data and source From 2004 to 2005, the 'Agricultural raw materials exports' dataset was unavailable due to Nigerian government limitations and sanctions (FAO 2013) Thus, to arrive at the final dataset, this report used data replacement, choosing the years 2003 or 2006 The 2003 data is 0.01% while the 2006 data is 0.36%, resulting in a 0.3% change This report will use 2003 data for better consistency and more straightforward for calculations However, the replacement will result in some misleading data II DESCRIPTIVE ANALYSIS a Central tendency Nigeria Comparison Ethiopia Mean 24.778 < 41.350 Median 24.569 < 41.348 Mode #N/A Table 2: The central tendency of two countries Agricultural Annual Growth Rate Minimum Compare value Lower threshold Maximum value Compare Upper threshold Observation result Ethiopia 31.112 > 29.527 54.029 > 53.615 Nigeria 19.990 > 13.229 36.965 > 34.597 Table 3: Outliers identification for the datasets After applying the IQR rule, it is determined that there are two outliers in the data set Therefore, the median will be used to measure the central tendency as it is not sensitive by outliers The median of Nigeria (24.569%) is lower than Ethiopia (41.348%), indicating a lower agricultural growth rate b Measure of variation Agriculture, forestry, and fishing, value added (% of GDP) Nigeria Comparison Ethiopia Range Unit: % of GDP 16.975 < 22.917 Interquartile range (IQR) Unit: % of GDP 5.342 < 6.022 Sample variance (SV) Unit: % of GDP < 18.432 Standard deviation (SD) Unit: % of GDP Coefficient of variation (CV) Unit: % 27.087 4.293 < 17% > 5.205 13% Table Table of comparison with all the measures of two countries of Variation in terms of t% of GDP of Agricultural Annual Growth Rate The Interquartile Range (IQR) – the spread of the 50% central data that excludes outliers – is selected as the most appropriate measure of variation for two countries Nigeria's IQR (5.342) is smaller than Ethiopia's (6.022), indicating that the central proportion of Ethiopia has higher spread c Measure of Shape Nigeria Compare (Unit: % of GDP) Left side Right side Conclusion Skewness Q1-Min vs Max-Q3 1.252 < 10.381 Rightskewed Median-Q1 vs Q3-Median 3.327 > 2.015 Leftskewed Median vs Mean 24.569 < 24.778 Rightskewed Rightskewed Table 5: Table of comparison between Nigeria Box-and-Whisker plot values to determine skewness Ethiopia Compare (Unit: % of GDP) Left side Q1-Min vs Max-Q3 7.448 Median-Q1 vs Q3-Median Median vs Mean Right side Conclusion < 9.447 Rightskewed 2.788 < 3.234 Rightskewed 41.348 < 41.350 Rightskewed Skewness Right-skewed Table 6: Table of comparison between Ethiopia Box-and-Whisker plot values to determine skewness Figure Box-plot of Nigeria and Ethiopia Tables 3,4 show that two countries box-plot are positively-skewed However, figure depicts that Nigeria values are more concentrated in the left tail of the distribution than Ethiopia values In other words, the initial 25% of data in Nigeria is sparsely distributed, whereas the last 25% is very scattered Ethiopia's box plot appears to be less distributed, with approximately equal right and left tails III MULTIPLE REGRESSION Multiple regression – a statistical technique for examining the relationship between a single dependent variable and a collection of independent variables – will be used to assess if the variables mentioned above have an influence on Ethiopia's and Nigeria's agricultural annual growth rates (Moore et al.) Agricultural Annual Growth Rate (AAGR) is taken as dependent variable (Y), and the remaining nine factors are identified as independent variables (X) (Table 5,6) To arrive at the final regression model, this report uses the backward elimination method (Appendix 1) to omit independent variables (X) that are not significant at 0.05 and only include those related to dependent variables (Y) Additionally, since outliers might have an effect on the results of regression analyses, both nations will exclude the years with outliers – 1997 and 2002 (Pratomo & Seohono 2017) Ethiopia Agriculture, forestry, and fishing, value added (% of GDP) / Agricultural Annual Growth Rate Foreign direct investment, net inflows (% of GDP) Employment in agriculture (% of total employment) (modeled ILO estimate) X Urban population (% of total population) X Average of Temperature - (Celsius) X Y Nigeria Agriculture, forestry, and fishing, value added (% of GDP) / Agricultural Annual Growth Rate Y Urban population (% of total population) X X Table Ethiopia independent and dependent variables  Ethiopia a Table Nigeria independent and dependent variables Final regression output Figure 2: Ethiopia's final regression output applying the application of the backward elimination technique Scatter plot and line fit plots Figure Line fit plot between FDI (% of GDP) and AAGR (% of GDP) 10 Figure Line fit plot between Average of Temperature (Celsius) and AAGR (% of GDP) Figure Line fit plot between Urban population (% of total population) and AAGR (% of GDP) 10 24 VII APPENDICES Appendix A: Backward elimination Definition and process The backward elimination technique is used to eliminate features that have no substantial effect on the dependent variable or on the output prediction Beginning with the full model, which includes all independent effects, a search is conducted using t (or p-) values and α to identify any nonsignificant variables If the model contains non-significant independent variables (p-value > α), the predator with the smallest absolute value is deleted from the process and a new model is developed The process of backward elimination will be repeated until all independent variables are significant In scenario in which no insignificant variables are found (p-value < α), the process end (Black & Asafu-Adjaye 2013) Since Nigeria missed 2014 and 2015 data for “Agriculture raw materials exports”, this report will used 2013 data instead (as explained above) Additionally, the multiple regression will omit 1997 and 2002 as they contain outliers Conduction The report will apply the Backward Elimination method for Ethiopia and Nigeria with the variables (X1 to X9) to find the significant independent variables that have a relationship with Agricultural Annual Growth Rate (AAGR) (Y) by comparing the p-value with � = 0.05  Ethiopia: ❖ Step 1: State the Null and Alternative Hypothesis H0: ��= (� � ���������ℎ�� � ��ℎ AAGR) H1: �� ≠ (�� ����� ��� ���������ℎ�� ���� ��� AAGR ��� ��������� � ) - Do not reject Null when p-value larger than or equal to � = 0.05 - Reject Null when p-value smaller than � = 0.05 ❖ Step 2: Run the Regression Model with the full variables 24 25 Figure 15 The Regression Model of Ethiopia with full variables Most of the variables in the figure have p-value larger than 0.05, except urban population and FDI Notably, CO2 emission has the highest p-value (0.883) and lowest absolute t-Stat (0.883) Hence, CO2 emission will be eliminated first ❖ Step 3: Run the Regression Model with k-1 (eight) variables: Figure 16 The Regression Model of Ethiopia with variables There are changed in the p-value after eliminating C02 emission, however, urban population and FDI continues to be the only two variables that have p-value smaller than 0.05 GDP per capita growth has the highest p-value (0.796) and lowest absolute t-Stat (0.265) Hence GDP per capita growth will be eliminated ❖ Step 4: Run the Regression Model with k-2 (seven) variables: 25 26 Figure 17 The Regression Model of Ethiopia with variables After removing two variables, the model shows that only three out of seven variables have p-value greater than 0.05 With the highest p-value (0.302) and lowest absolute t-Stat (1.079), Agricultural raw materials exports will be eliminated next ❖ Step 5: Run the Regression Model with k-3 (six) variables: Figure 18 The Regression Model of Ethiopia with variables At this point, only two variables have p-value greater than 0.05, the backward elimination process continues With the highest p-value (0.218) and lowest absolute t-Stat (1.293), Inflation, consumer price will be eliminated next ❖ Step 6: Run the Regression Model with k-3 (five) variables: Figure 19 The Regression Model of Ethiopia with variables 26 27 Average rainfall is the only variable with p-value greater than 0.05 The report will eliminate it ❖ Step 7: Run the Regression Model with k-3 (four) variables: Figure 20 The Regression Model of Ethiopia with four variables As can be seen, there are four significant variables left, with p-value smaller than 0.05 Therefore, the Backward elimination process ends In conclusion, FDI, agricultural employment, urban population, and temperature average all have a relationship with AAGR In other words, at the 5% level of significance, there are four significant variables that affect the AAGR in Ethiopia: FDI, agricultural employment, urban population, and average temperature  Nigeria: ❖ Step 1: State the Null and Alternative Hypothesis H0: �� = (�� ���������ℎ�� ���ℎ AAGR) H1: �� ≠ (�� ����� ��� ���������ℎ�� ������� AAGR ��� ��������� �) - Do not reject Null when p-value larger than or equal to � = 0.05 - Reject Null when p-value smaller than � = 0.05 ❖ Step 2: Run the Regression Model with the full variables 27 28 Figure 21 The Regression Model of Ethiopia with full variables Most of the variables in the figure have p-value larger than 0.05, except GDP per capita growth, average of rainfall, and average of temperature Notably, CO2 emission has the highest p-value (0.883) and lowest absolute t-Stat (0.900) Hence, CO2 emission will be eliminated first ❖ Step 3: Run the Regression Model with k-1 (eight) variables: Figure 22 The Regression Model of Ethiopia with variables There are changed in the p-value after eliminating C02 emission and GDP per capita growth, Inflation, consumer prices, and Average of rainfall be the variables that have p-value smaller than 0.05 Employment in agriculture has the highest p-value (0.774) and lowest absolute t-Stat (0.294) Hence Employment in agriculture will be eliminated 28 29 ❖ Step 4: Run the Regression Model with k-2 (seven) variables: Figure 23 The Regression Model of Ethiopia with variables After removing two variables, the model shows that only two out of seven variables have p-value greater than 0.05 With the highest p-value (0.842) and lowest absolute t-Stat (1.204), Foreign direct investment, net inflows exports will be eliminated next ❖ Step 5: Run the Regression Model with k-3 (six) variables: Figure 24 The Regression Model of Ethiopia with variables With the highest p-value (0.735) and lowest absolute t-Stat (0.345), agricultural raw materials exports will be eliminated next 29 30 ❖ Step 6: Run the Regression Model with k-3 (five) variables: Figure 25 The Regression Model of Ethiopia with variables With the highest p-value (0.384) and lowest absolute t-Stat (0.899), Foreign direct investment, net inflows exports will be eliminated next ❖ Step 7: Run the Regression Model with k-3 (four) variables: Figure 26 The Regression Model of Ethiopia with four variables With the highest p-value (0.217) and lowest absolute t-Stat (1.288), average of temperature will be eliminated next ❖ Step 8: Run the Regression Model with k-3 (three) variables: 30 31 Figure 27 The Regression Model of Ethiopia with three variables In this step, the variable GDP per capita growth will be eliminated as it has a p-value (0.069) higher than Significance Level ❖ Step 9: Run the Regression Model with k-3 (two) variables: Figure 28 The Regression Model of Ethiopia with two variables In this step, the variable average of rainfall will be eliminated as it has a p-value (0.073) higher than Significance Level ❖ Step 10: Run the Regression Model with k-3 (two) variables: 31 32 Figure 29 The Regression Model of Ethiopia with one variable As can be seen, there is only one significant variable left, with p-value smaller than 0.05 Therefore, the Backward elimination process ends In conclusion, urban population has a relationship with AAGR In other words, at the 5% level of significance, there is only one significant variable that affect the AAGR in Nigeria: urban population Appendix B: Hypothesis testing for significant trend According to Kenton (2021), a regression model for a dataset's values and times in a time series with a significance F value less than 0.05 is considered statistically significant and meaningful This paper will follow the instruction and focus on the criteria of significance F value smaller than 0.05 for the data set of both Nigeria and Ethiopia  NIGERIA A Linear Model (Linear trend exists between AAGR and Time - Year) (Linear trend does not exist between AAGR and Time - Year) As Significance F(0.001) < 0.05, is rejected and we can conclude at 95% certainty that the Linear trend between AAGR and Time - Year is significant 32 33 Figure 30 Regression output for Nigeria’s linear model B Quadratic Model (Exponential trend exists between AAGR and Time - Year) (Exponential trend does not exist between AAGR and Time - Year) As Significance F(0.004) < 0.05, is rejected However, the P-value for the variables Time Period (0.860) and T^2 (0.287) are much larger than 0.05 Therefore, we can conclude with 95% certainty that the variables of the quadratic trend are insignificant, thus causing the quadratic trend to be insignificant in predicting the variance of AAGR Figure 31 Regression output for Nigeria’s quadratic model C Exponential Model (Exponential trend exists between AAGR and Time - Year) (Exponential trend does not exist between AAGR and Time - Year) 33 34 As Significance F(0.001) < 0.05, is rejected and we can conclude at 95% certainty that the Exponential trend between AAGR and Time - Year is significant Figure 32 Regression output for Nigeria’s exponential model ETHIOPIA A Linear Model (Linear trend exists between AAGR and Time - Year) (Linear trend does not exist between AAGR rate and Time - Year) As Significance F(0.001) < 0.05, is rejected and we can conclude at 95% certainty that the Linear trend between AAGR and Time - Year is significant Figure 33 Regression output for Ethiopia’s linear model B Quadratic Model (Exponential trend exists between AAGR and Time - Year) (Exponential trend does not exist between AAGR and Time - Year) 34 35 As Significance F(0.004) < 0.05, is rejected However, the P-value for the variables Time Period (0.472) and T^2 (0.742) are much larger than 0.05 Therefore, we can conclude with 95% certainty that the variables of the quadratic trend are insignificant, thus causing the quadratic trend to be insignificant in predicting the variance of AAGR Figure 34 Regression output for Ethiopia’s quadratic model C Exponential Model (Exponential trend exists between AAGR and Time - Year) (Exponential trend does not exist between AAGR and Time - Year) As Significance F(0.001) < 0.05, is rejected and we can conclude at 95% certainty that the Exponential trend between AAGR and Time - Year is significant Figure 35 Regression output for Ethiopia’s exponential model 35 36 VIII REFERENCES LIST ‘Acknowledgement to Reviewers of Agriculture in 2018’ 2019, Agriculture, vol 9, no 1, p 19 Agwu, EA & Anugwa, QI 2021, 'Stemming rural-urban migration through agricultural development: Can Nigeria apply the lessons from the COVID-19 pandemic?', Agro Science, October, vol 20, no 4, pp 36-45, viewed January 2022, researchgate database And, A 1997, Nigeria: agriculture for national economic growth and development., Federal Ministry Of Agriculture And Natural Resources, Garki, Abuja Asadullah, MN & Rahman, S 2009, ‘Farm productivity and efficiency in rural Bangladesh: the role of education revisited’, Applied Economics, vol 41, no 1, pp 17–33 Atkeyelsh G M Persson 2021, FOREIGN DIRECT INVESTMENT IN LARGE-SCALE AGRICULTURE IN AFRICA: economic, social and environmental sustainability in ethiopia., Routledge, S.L Bachewe, NF, Berhane, G, Minten, B & Taffesse, AS 2015, 'Agricultural Growth in Ethiopia (2004-2014): Evidence and Drivers', INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE, October Banayo, NPM, Rahon, RE, Sta Cruz, P & Kato, Y 2020, ‘Fertilizer responsiveness of highyielding drought-tolerant rice in rainfed lowlands’, Plant Production Science, vol 24, no 3, pp 279–286 Bartlett, JW & Frost, C 2008, ‘Reliability, repeatability and reproducibility: analysis of measurement errors in continuous variables’, Ultrasound in Obstetrics and Gynecology, vol 31, no 4, pp 466–475 Black, Asafu-Adjaye, J 2013, Australian Business Statistics, 2nd edition, John Wiley & Son Australia, Ltd corporatefinanceinstitute n.d, What is the Coefficient of Determination?, viewed January 2022, Eshetu, F & Mehare, A 2020, 'Determinants of Ethiopian Agricultural Exports: A Dynamic Panel Data Analysis', Review of Market Integration, vol 12, no 1-2, pp 70-94 FAO 2013, ANALYSIS OF INCENTIVES AND DISINCENTIVES FOR MAIZE IN NIGERIA Fredrik SYderbaum, Taylor, I & Nordiska Afrikainstitutet 2008, Afro-regions: the dynamics of cross-border micro-regionalism in Africa, Nordiska Afrikainstitutet, Uppsala Frost, J 2015, How To Interpret R-squared in Regression Analysis, statisticsbyjim, viewed January 2022, georank n.d, Ethiopia vs Nigeria: Economic Indicators Comparison, georank, viewed January 2022, < https://georank.org/economy/ethiopia/nigeria> 36 37 Hoskins, MD 2009, ‘THE GREEN REVOLUTION AND CROPPING INTENSITY’, Institute of Development Studies Bulletin, vol 5, no 4, pp 43–50 Inegbedion, H, Obadiaru, E, Obasaju, B, Asaleye, A & Lawal, A 2019, ‘Financing Agriculture in Nigeria through Agricultural Extension Services of Agricultural Development Programmes (ADPs)’, F1000Research, vol 7, p 1833 ITA 2021, Nigeria - Agriculture Sector, trade.gov, viewed January 2022, J Abbink & Jean-Nicolas Bach 2017, A Decade of Ethiopia: Politics, Economy and Society, 2004-2016, Brill, Leiden; Boston Kenton, W 2021, Why Statistical Significance Matters, Investopedia, viewed February 2022, Ketema, MA & Negeso, KD 2020, 'Effect of climate change on agricultural output in Ethiopia', Jurnal Perspektif Pembiayaan dan Pembangunan Daerah, August, vol 8, no 3, pp 2355-8520 Lenne, JM 2007, ‘Enhancing crop-livestock systems for agricultural productivity, food security and reduced poverty in developing countries’, Proceedings of the British Society of Animal Science, vol 2007, pp 258–258 Mathews, S 2018, Interpreting Regression Output - Without all the Statistics Theory Nori, M & Farinella, D 2020, Migration, agriculture and rural development: IMISCOE Short Reader, Springer Open, Cham, Switzerland Oloukoi, J, Oyinloye, RO & Yadjemi, H 2014, ‘Geospatial analysis of urban sprawl in Ile-Ife city, Nigeria’, South African Journal of Geomatics, vol 3, no 2, p 128 Onesmus Semalulu, Simon Okello, Robinah Gafabusa Naggayi, Phyllis Kasambula & Hongsik Shim 2020, ‘Profitability of Improved versus Traditional Cultivation Practices for African Indigenous Vegetables’, Journal of Agricultural Science and Technology B, vol 10, no Osaki, M, Morikawa, K, Shinano, T, Urayama, M & Tadano, T 1991, ‘Productivity of highyielding crops’, Soil Science and Plant Nutrition, vol 37, no 3, pp 445–454 Pickett, J & Organisation For Economic Co-Operation And Development Development Centre 1991, Economic development in Ethiopia: agriculture, the market, and the state, Development Centre Of The Organisation For Economic Co-Operation And Development; Washington, D.C, Paris Pratomo, DS & Seohono, L 2017, ‘Does Outlier need to be removed from Regression Analysis? Case Study in Economics Research’, Journal of Applied Economic Sciences, vol 12, no 4, pp 1141–1147 Reimers, M & Klasen, S 2013, ‘Revisiting the Role of Education for Agricultural Productivity’, American Journal of Agricultural Economics, vol 95, no 1, pp 131–152 statista n.d, Agriculture in Nigeria and Ethiopia - statistics and facts, viewed January 2022, 37 38 Thirtle, C, Lin, L & Piesse, J 2003, ‘The Impact of Research-Led Agricultural Productivity Growth on Poverty Reduction in Africa, Asia and Latin America’, World Development, vol 31, no 12, pp 1959–1975 Welteji, D 2018, 'A critical review of rural development policy of Ethiopia: access, utilization and coverage', Agriculture & Food Security, August, vol 7, no 55, viewed January 2022, Agriculture & Food Security database worldometers n.d., Ethiopia GDP - Worldometer, www.worldometers.info WTO 2005, WTO | Managing the Challenges of WTO Participation: Case Study, www.wto.org, viewed January 2022, Yalew, WA 2020, 'Urban Agriculture in Ethiopia: An Overview', Universal Wiser, vol 1, no 2, pp 1-8 38 ... (Bachewe et.al 2014) However, drought, soil degradation caused by overgrazing, deforestation, excessive taxation, and insufficient infrastructure continue to plague these industries (J Abbink... 2020 (Inegbedion et al 2019) The sector is divided into four subsectors: crop production, livestock production, forestry, and fishing, however, livestock and crop farming are still considered... GDP and labor, whilst Nigeria continues to favor heavy industries above agriculture 2) Agriculture is crucial to poverty reduction and increased food security in most developing nations However,