38 Research on World Agricultural Economy | Volume 03 | Issue 03 | September 2022 Research on World Agricultural Economy https://ojs.nassg.org/index.php/rwae Copyright © 2022 by the author(s). Published by NanYang Academy of Sciences Pte. Ltd. This is an open access article under the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) License. (https://creativecommons.org/licenses/by-nc/4.0/). *Corresponding Author: Abera Gayesa Tirfi, Department of Agriculture and Animal Health, University of South Africa, Ethiopia Regional Learning Centre, Ethiopia; Email: abera.gayesa@gmail.com DOI: http://dx.doi.org/10.36956/rwae.v3i3.580 Received: 1 July 2022; Received in revised form: 12 August 2022; Accepted: 18 August 2022; Published: 5 September 2022 Citation: Tirfi, A.G., 2022. Determinants of Barley Output Supply Response in Ethiopia: Application of Ardl Bound Cointegration Approach. Research on World Agricultural Economy. 3(3), 580. http://dx.doi.org/10.36956/rwae.v3i3.580 RESEARCH ARTICLE Determinants of Barley Output Supply Response in Ethiopia: Application of Ardl Bound Cointegration Approach Abera Gayesa Tirfi* Department of Agriculture and Animal Health, University of South Africa, Ethiopia Regional Learning Centre, Ethiopia Abstract: This study investigated barley output supply response determinant factors in Ethiopia. An ARDL bound test approach was employed as method using secondary data from 1981-2020. The study demonstrated that barley output supply was affected positively and significantly by zero-order lagged seasonal rainfall and crop growing period temperature. The study supports the findings of researchers who reported that warming temperature followed by an increase in the amount of rainfall had a positive impact on barley output supply. The positive impact of temperature was induced because of a rise in the ocean and earth’s surface average temperature, causing more evaporation that increases overall rainfall while reaching over the highland areas. Studies confirm that ENSO and moist winds coming from the Atlantic and Indian Oceans influence the occurrence of rainfall in the western, southeastern, central, and northern highlands of Ethiopia. The study further exhibited that CSMRR and CGPMT had a positive effect on barley output both in the long-run and short-run, implying that climate parameters have minimal effect on barley production. Non- climatic variables demonstrated that both lagged and current year’s producer prices had a positively significant effect on barley output supply in both the long-run and short-run, implying that barley output supply is highly responsive to any price incentive strategies announced before re-allocation of the area towards barley cultivation. Conversely, the study explored that use of fertilizer in first-order lag had negatively significant impact on barley output supply in both seasons; implying that increased use of fertilizer in lagged period may reduce barley output as a result of inappropriate fertilizers application by farmers. The results generated by this study are useful addendum to the repository of knowledge on elasticity of crop supply at an aggregate level, which can be used in designing strategies and measures for mitigation and adaptation of climate change. Keywords: Changing climate; Supply response; Barley output; ARDL model 1. Introduction Changes in climate factors in terms of warmer tempera- ture and variability in seasonal rainfall patterns have been reported as the main factors reducing agricultural produc- tion [1]. Numerous studies concluded that climate change has posed a strong effect on agricultural output in most of sub-Saharan Africa, including Ethiopia [2]. Researchers reported that the negative effects exerted due to climate change are anticipated to be more severe in developing mailto:abera.gayesa@gmail.com http://dx.doi.org/10.36956/rwae.v3i3.580 https://orcid.org/0000-0002-6267-1366 39 Research on World Agricultural Economy | Volume 03 | Issue 03 | September 2022 countries where food insecurity is a major problem since rainfall is the only source of moisture for soil to meet the water requirement of crops in agricultural production practices [3]. Agricultural production and its performance in Ethiopia also depend on the pattern of seasonal climate parameters out of which seasonal rainfall such as short/ belg-season and long/main-season rainfalls are key fac- tors in local food production systems, including barley crop [4]. Barley (Hordeum vulgare L.) has been reported as the fourth most important cereal crop in the world in terms of production [5]. In terms of volume of production, it ranks fourth in the world and fifth in Ethiopia [6]. It was used by ancient civilizations as food for humans and animals, as well as to make alcoholic beverages. According to CSA [7], the barley crop is considered as a major cereal crop in Ethiopia accounting for 9% in terms of both the area cul- tivated under cereal crops (0.95 million hectares) and the volume of total annual cereal production (2.378 million tons). Shreds of evidence show that Ethiopia is considered a center of barley diversity [8] having diverse landraces and local varieties cultivated under a wide spectrum of land races as a result of its adaptation capability to diverse and harsh climatic conditions and soil types. Such a wide diversity is assumed to be contributed a result of long- term geographic isolation since barley is considered a founder of Old World Agriculture and have been assumed to be cultivated in Ethiopia for the last 5,000 years [9]. In Ethiopia, barley is currently cultivated at altitudes ranging from 1,400 meters to 4,000 meters above sea level (m.a.s.l) which have an extremely variable climatic and edaphic environment [10]. Barley is cultivated in all regions of Ethiopia. The most important barley-producing regions are Shewa, Arsi, Bale, Gojam, Gonder, Welo, and Tigray. “Belg” season barley is also produced in Wollo, Shewa, and Bale. The estimated production of barley between 1981 and 2020 was 1.08 and 2.38 million tons respectively, which showed an increase of about 220% over the years. However, barley production in Ethiopia is constrained by several problems such as climate change (high inter- annual rainfall variability and increasing temperature), unpredictable drought stress, poor soil fertility, water logging moisture stress, low yield potential of currently grown cultivars, and infestation of diseases, insect pests and weeds [11]. Among these factors, climate change significantly affects the production of the barley crops. Nowadays, the incidence of climate change of the world is widely agreed upon among the scientific community. The Intergovernmental Panel on Climate Change (IPCC) assessment has confirmed that anthropogenic activities are changing the climate system of the world and regions which may remain to do so [2]. In the last century, the im- pacts associated with surface temperatures on the physical and biological systems were increasingly being observed. The findings inform that climate changes may lead to environmental changes, such as a rise in sea level, and alterations of climatic zones due to warmer temperatures and variation in rainfall patterns. Some African countries, including Ethiopia, are vulner- able to the severe impacts of changes in climate derived as a result of limitations in capacity and access to miti- gation and adaptive resources. Most of these countries are considered the ones most susceptible and vulnerable to climatic changes in the world [12,13]. In Ethiopia, bar- ley production is highly dependent on rainfall, since the contribution of irrigation is estimated to be less than 1% of the country’s total cultivated land area under barley. Hence, the impact of these climate changes on the produc- tion and supply of barley output should be studied to pro- vide detailed information to researchers and policy plan- ners. There is a scarcity of such empirical studies having a national scope on the impact of climate change on barley production in Ethiopia. The study aimed to investigate the determinant factors influencing barley output supply response in Ethiopia. The results of the study could be used for future planning of the mitigation and adaptation responses to be taken. 2. Materials and Methods 2.1 Description of Study Area Located in the Horn of Africa, Ethiopia’s latitudinal and longitudinal locations are between 3o to 15o N and 33o to 48oE, respectively. According to the World Bank [14], the country is bordered with Sudan, Eritrea, Djibouti, Soma- lia, Kenya, and South Sudan. Ethiopia is administratively divided into four levels: regions/city administrations, zones, woredas, and kebeles; kebele being the last grass- roots administrative unit. According to the population projection of the United Nations Population Funds [15], the Ethiopian population has reached 117.90 million with an annual growth rate of 2.6 percent. Barley, the theme of this study, is among the most important food crops grown in the country. Shreds of evidence show that the major barley growing belts in the country include: Oromia, Amhara, Tigray, and Southern Nation Nationality and Regional State, which supply about 99.9% of the total national barley production [16]. Zone-wise, it is grown mainly in the zones of Arssi, Bale, Shewa, Wollo, Gojam, and Gonder [17] (see Figure 1 for the map). 40 Research on World Agricultural Economy | Volume 03 | Issue 03 | September 2022 According to Muluken and Jemal [18], the wheat crop optimally grows the best at higher locations ranging from 2000 meters to 3500 meters above sea levels. Barley crop is mostly grown during two consecutive seasons: the short/belg-season and main/ meher-season at the higher elevations of Dega Agroecologies. In Ethiopia, the crop is substantially grown and supplied by smallholder subsist- ence farmers, who mostly grow local seed varieties with either little or no application of modern inputs like ferti- lizers, pesticides, and herbicides. In Ethiopia, barley is mainly grown during long-rainy/ meher season from June – September when the amount of crop growing period rainfall ranges between 180 mm to 400 mm depending on the altitudinal and geographic location [19]. Although the crop mainly grows in the highlands, it can also be grown in a subtropical climates characterized by hot, humid summers and cool to mild winters. Barley best suits a temperature of 12 °C ~ 15 °C during the crop growing period and about 30 °C at maturity time. The crop cannot tolerate frost at all stages of growth, particularly at the flowering stage. The incidence of frost at the crop flowering stage highly affects the yield of the barley crops. 2.2 Data Type and Sources The data selected for this study included: barley output, the area allocated under barley cultivation, quantity of chemical fertilizers and improved barley seeds consumed, and producer price of the barley crop. These nationally aggregated time series secondary data were obtained and compiled from Agricultural Sample Survey Reports of the Ethiopian Central Statistical Agency (CSA) for the period 1981 to 2020. Furthermore, secondary data on climatic parameters (seasonal temperature and rainfall) for the ob- servation period have been obtained from the Ethiopian National Meteorological Agency (NMA). Weather sta- tions considered representative from barley crop growing belts were selected (12 stations) and crop growing period rainfall and temperature data have been taken as recorded in the NMA database. Specifically, average monthly data covering the crop growing period (F-S) were taken. His- torical producer prices of barley crops for the observation period have also been compiled from the FAOSTAT data- base and CSA. 2.3 Empirical Model Selection and Specification 2.3.1 Variables Considered for Investigation In this study, the variables considered for investiga- tion included climatic and non-climatic variables that affect output supply response of barley crop. From the climatic variables, temperature and rainfall were included in the study since it was assumed that these variables ex- ert substantial impact on barley output supply response. Furthermore, producer price, fertilizer, and land area were selected from the non-climatic variables. Labor and farm machinery are variable inputs that must have been includ- ed in the study but excluded since there is no time series data that matches with the other variable inputs. The concep- tual structure of this investigation is depicted in Figure 2. Figure 1. Major Barley Growing Belts of Ethiopia 41 Research on World Agricultural Economy | Volume 03 | Issue 03 | September 2022 2.3.2 Model Review and Selection To measure the impacts exerted by climatic and non- climatic factors on crop output supply response, research- ers have employed different analytical models. Among the models, the general circulation models (GCMs), the Cobweb, Ricardian, and ARDL models are models widely used in empirical studies of supply responses. The GCMs are the most complex climate prediction models devel- oped to predict what would happen to climate around the world in response to a wide variety of changes in the concentrations of greenhouse gases in the atmosphere [20]. However, the GCMs have limitations such as poor knowl- edge on the ocean circulation processes, lack of know- how on cloud formation and feedbacks, crude spatial reso- lution, and inability to simulate current regional climate factors accurately [21]. The Cobweb Model is the lag between production deci- sions and the realization of demand and market prices [22]. According to Serena Brianzoni (2018), the cobweb model is a dynamical system that describes price fluctuations as a result of the interaction between demand function, depending on current price, and supply function, depend- ing on expected price [23]. The theory focuses on analysis of price fluctuations in market (demand side) which may lead to reduction of food grains production (supply side). However, the Cobweb model has weaknesses; viz. price divergence being unrealistic and not empirically seen. As this study focuses on supply side factors, particularly tem- perature and rainfall as well as physical inputs like ferti- lizer and area under barley production, the Cobweb model has not been considered for the study. The Ricardian model analyzes a cross section of farms under different climatic conditions and examines the re- lationship between the value of land or net revenue and agro-climatic factors [21]. The model was applied by re- searchers in the valuation of the contributions made by the environmental factors to farm income by regressing land values on a set of environmental inputs where net rev- enue or price of land represent farm income. Though the Ricardin model measures climatic factors against value of land, it possesses weaknesses such as non inclusion of non-climatic factors and impossibility of getting perfect measures for such variables [24], non inclusion of price ef- fects, and does not account the fertilization effect of CO2 concentrations. The ARDL Model is an ordinary least square (OLS) based model which is applicable for both non-stationary time series as well as for times series with mixed order of inte- gration. The ARDL approach developed by Pesaran, et al. [25] as modified from the previous traditional cointegration technique was documented by Johansen and Juseline [26]. The model is considered as the best econometric method compared to others to estimate short-run and long-run impact of explanatory variables on output supply response of crops [27,28]. The ARDL approach enjoys several advan- tages over the others such as its appropriateness for gener- ating short-run and long-run elasticities for a small sample sizes, affords flexibility about the order of integration of the variables, and suitable for the independent variable in the model which is I(0), I(1), or mutually cointegrated [29]. In view of these, the ARDL Model was selected for the current study. 2.3.3 Model Specification This study applied an ARDL bound cointegration ap- proach proposed by Pesaran, et al. [25] to examine the im- pact of climatic and non-climate input variables on barley output supply responses. To find the relationship between dependent and in- dependent variables, the following general form of the ARDL model was constructed: 7 BaProt = ɑ0 + i=1 p β IBaProt-i + i=0 q β iXt-i + Ut (1) where BaProt represents barley production, BaProt-i represents barley output supplied in year t-i, Xt-i represents explanatory variables in year t-i, t represents the time from 1981 to 2020, and β0, βi, … are coefficients of variables included in the model, and Ut is disturbance term. In this study, it was considered that the relationship between the independent and explanatory variables is expected to take the following functional form: BaProt = f (BaProt-1, BaPrit, BaArt, FertQt, CSMRFt, CGPMTt) (2) where BaProt is barley output measured in million tons; BaProt-1 is barley output in first- lag order, BaPrit is producer price of barley output in ETB, BaArt is land area allocated under barley cultivation, FertQt is fertilizer quantity used in barley production, CSMRFt is crop season mean rainfall in millimeters, and CGPMTt is crop growing period mean temperatures in degrees Celsius. By converting all the variables in Equation (2) into natural log form, the model is expressed as below: lnBaProt= β0 + β1lnBaProt-1 + β2lnBaPrit + β3lnBaArt + β4lnFertQt + β5lnCSMRFt + β6lnCGPMTt + εt (3) where lnCSMRFt is the log of crop season mean rainfall in mm, and lnCGPMTt is the log of crop growing period mean temperature in °C. In addition, εt represents the disturbance term. To generate some long-run relationships, Equation (3) is hereby modified as: lnBaProt = ɑ0 + α 1lnBaProt-i + α 2lnBaPrit-i + α 3lnBaArt-i + α 5lnFertt-i + α 6lnCSMRFt- i + α 8lnCGPMTt-i + εt-I (4) In case the variables are found cointegrated, the model exemplifies the existence of an error correction representation. After establishing the above long-run relationship between variables, the Error Correction Model (ECM) can be derived from the ARDL model (Equation (4)) through simple linear transformation to find the short-run elasticity coefficients, which integrates short-run adjustments with long-run equilibrium. The short- run elasticity coefficients can be estimated using the following Dynamic ARDL Error Correction Model (ECM): ∆lnBaProt = β0 + β 1∆lnBaProt-i + β 2∆lnBaPrit-i + β 3∆lnBArt-i+ β 5∆lnFertt-i + β 6∆lnCSMRFt-i + β 8∆lnCGPMTt-i + ψiECT1-i+ ui (5) where ∆ represents the first difference while ψi is the coefficient of ECM for short-run dynamics. ECM shows the speed of adjustment in long-run equilibrium after a shock in the short-run. In this study, the investigator used the general to a specific approach to select an optimal lag length for the ARDL model. Before estimating the ARDL bound test using the models established above, the data series on the selected variables should be tested to detect the presence of unit root and long-run cointegration. To this end, an Augmented Dickey-Fuller (ADF) and Philips- (1) where BaProt represents barley production, BaProt-i repre- sents barley output supplied in year t-i, Xt-i represents ex- planatory variables in year t-i, t represents the time from 1981 to 2020, and β0, βi, … are coefficients of variables included in the model, and Ut is disturbance term. In this study, it was considered that the relationship between the independent and explanatory variables is expected to take 5 humid summers and cool to mild winters. Barley best suits a temperature of 12°C ~ 15 °C during the crop growing period and about 30 °C at maturity time. The crop cannot tolerate frost at all stages of growth, particularly at the flowering stage. The incidence of frost at the crop flowering stage highly affects the yield of the barley crops. 2.2 Data Type and Sources The data selected for this study included: barley output, the area allocated under barley cultivation, quantity of chemical fertilizers and improved barley seeds consumed, and producer price of the barley crop. These nationally aggregated time series secondary data were obtained and compiled from Agricultural Sample Survey Reports of the Ethiopian Central Statistical Agency (CSA) for the period 1981 to 2020. Furthermore, secondary data on climatic parameters (seasonal temperature and rainfall) for the observation period have been obtained from the Ethiopian National Meteorological Agency (NMA). Weather stations considered representative from barley crop growing belts were selected (12 stations) and crop growing period rainfall and temperature data have been taken as recorded in the NMA database. Specifically, average monthly data covering the crop growing period (F-S) were taken. Historical producer prices of barley crops for the observation period have also been compiled from the FAOSTAT database and CSA. 2.3 Empirical Model Selection and Specification 2.3.1 Variables Considered for Investigation In this study, the variables considered for investigation included climatic and non-climatic variables that affect output supply response of barley crop. From the climatic variables, temperature and rainfall were included in the study since it was assumed that these variables exert substantial impact on barley output supply response. Furthermore, producer price, fertilizer, and land area were selected from the non-climatic variables. Labor and farm machinery are variable inputs that must have been included in the study but excluded since there is no time series data that matches with the other variable inputs. The conceptual structure of this investigation is depicted in Figure 2. Figure 2. Conceptual Structure of the Investigation 2.3.2 Model Review and Selection To measure the impacts exerted by climatic and non-climatic factors on crop output supply response, researchers have employed different analytical models. Among the models, the general circulation models (GCMs), the Cobweb, Ricardian, and ARDL models are models widely used in empirical studies of supply responses. The GCMs are the most complex climate prediction models developed to predict what would happen to climate around the world in response to a wide variety of changes in the concentrations of C lim at ic F ac to rs Barley Production N on -C lim at ic F ac to rsCrop Season Rainfall CGP Temperature Fertilizer Use Producer Price Area under barley production Labor & Machinery Figure 2. Conceptual Structure of the Investigation 42 Research on World Agricultural Economy | Volume 03 | Issue 03 | September 2022 the following functional form: 7 BaProt = ɑ0 + i=1 p β IBaProt-i + i=0 q β iXt-i + Ut (1) where BaProt represents barley production, BaProt-i represents barley output supplied in year t-i, Xt-i represents explanatory variables in year t-i, t represents the time from 1981 to 2020, and β0, βi, … are coefficients of variables included in the model, and Ut is disturbance term. In this study, it was considered that the relationship between the independent and explanatory variables is expected to take the following functional form: BaProt = f (BaProt-1, BaPrit, BaArt, FertQt, CSMRFt, CGPMTt) (2) where BaProt is barley output measured in million tons; BaProt-1 is barley output in first- lag order, BaPrit is producer price of barley output in ETB, BaArt is land area allocated under barley cultivation, FertQt is fertilizer quantity used in barley production, CSMRFt is crop season mean rainfall in millimeters, and CGPMTt is crop growing period mean temperatures in degrees Celsius. By converting all the variables in Equation (2) into natural log form, the model is expressed as below: lnBaProt= β0 + β1lnBaProt-1 + β2lnBaPrit + β3lnBaArt + β4lnFertQt + β5lnCSMRFt + β6lnCGPMTt + εt (3) where lnCSMRFt is the log of crop season mean rainfall in mm, and lnCGPMTt is the log of crop growing period mean temperature in °C. In addition, εt represents the disturbance term. To generate some long-run relationships, Equation (3) is hereby modified as: lnBaProt = ɑ0 + α 1lnBaProt-i + α 2lnBaPrit-i + α 3lnBaArt-i + α 5lnFertt-i + α 6lnCSMRFt- i + α 8lnCGPMTt-i + εt-I (4) In case the variables are found cointegrated, the model exemplifies the existence of an error correction representation. After establishing the above long-run relationship between variables, the Error Correction Model (ECM) can be derived from the ARDL model (Equation (4)) through simple linear transformation to find the short-run elasticity coefficients, which integrates short-run adjustments with long-run equilibrium. The short- run elasticity coefficients can be estimated using the following Dynamic ARDL Error Correction Model (ECM): ∆lnBaProt = β0 + β 1∆lnBaProt-i + β 2∆lnBaPrit-i + β 3∆lnBArt-i+ β 5∆lnFertt-i + β 6∆lnCSMRFt-i + β 8∆lnCGPMTt-i + ψiECT1-i+ ui (5) where ∆ represents the first difference while ψi is the coefficient of ECM for short-run dynamics. ECM shows the speed of adjustment in long-run equilibrium after a shock in the short-run. In this study, the investigator used the general to a specific approach to select an optimal lag length for the ARDL model. Before estimating the ARDL bound test using the models established above, the data series on the selected variables should be tested to detect the presence of unit root and long-run cointegration. To this end, an Augmented Dickey-Fuller (ADF) and Philips- (2) where BaProt is barley output measured in million tons; BaProt-1 is barley output in first-lag order, BaPrit is pro- ducer price of barley output in ETB, BaArt is land area al- located under barley cultivation, FertQt is fertilizer quan- tity used in barley production, CSMRFt is crop season mean rainfall in millimeters, and CGPMTt is crop growing period mean temperatures in degrees Celsius. By converting all the variables in Equation (2) into natural log form, the model is expressed as below: lnBaProt= β0 + β1lnBaProt-1 + β2lnBaPrit + β3lnBaArt + β4lnFertQt + β5lnCSMRFt + β6lnCGPMTt + εt (3) where lnCSMRFt is the log of crop season mean rainfall in mm, and lnCGPMTt is the log of crop growing period mean temperature in °C. In addition, εt represents the dis- turbance term. To generate some long-run relationships, Equation (3) is hereby modified as: 7 BaProt = ɑ0 + i=1 p β IBaProt-i + i=0 q β iXt-i + Ut (1) where BaProt represents barley production, BaProt-i represents barley output supplied in year t-i, Xt-i represents explanatory variables in year t-i, t represents the time from 1981 to 2020, and β0, βi, … are coefficients of variables included in the model, and Ut is disturbance term. In this study, it was considered that the relationship between the independent and explanatory variables is expected to take the following functional form: BaProt = f (BaProt-1, BaPrit, BaArt, FertQt, CSMRFt, CGPMTt) (2) where BaProt is barley output measured in million tons; BaProt-1 is barley output in first- lag order, BaPrit is producer price of barley output in ETB, BaArt is land area allocated under barley cultivation, FertQt is fertilizer quantity used in barley production, CSMRFt is crop season mean rainfall in millimeters, and CGPMTt is crop growing period mean temperatures in degrees Celsius. By converting all the variables in Equation (2) into natural log form, the model is expressed as below: lnBaProt= β0 + β1lnBaProt-1 + β2lnBaPrit + β3lnBaArt + β4lnFertQt + β5lnCSMRFt + β6lnCGPMTt + εt (3) where lnCSMRFt is the log of crop season mean rainfall in mm, and lnCGPMTt is the log of crop growing period mean temperature in °C. In addition, εt represents the disturbance term. To generate some long-run relationships, Equation (3) is hereby modified as: lnBaProt = ɑ0 + α 1lnBaProt-i + α 2lnBaPrit-i + α 3lnBaArt-i + α 5lnFertt-i + α 6lnCSMRFt-i + α 8lnCGPMTt-i + εt-I In case the variables are found cointegrated, the model exemplifies the existence of an error correction representation. After establishing the above long-run relationship between variables, the Error Correction Model (ECM) can be derived from the ARDL model (Equation (4)) through simple linear transformation to find the short-run elasticity coefficients, which integrates short-run adjustments with long-run equilibrium. The short- run elasticity coefficients can be estimated using the following Dynamic ARDL Error Correction Model (ECM): ∆lnBaProt = β0 + β 1∆lnBaProt-i + β 2∆lnBaPrit-i + β 3∆lnBArt-i+ β 5∆lnFertt-i + β 6∆lnCSMRFt-i + β 8∆lnCGPMTt-i + ψiECT1-i+ ui (5) where ∆ represents the first difference while ψi is the coefficient of ECM for short-run dynamics. ECM shows the speed of adjustment in long-run equilibrium after a shock in the short-run. In this study, the investigator used the general to a specific approach to select an optimal lag length for the ARDL model. Before estimating the ARDL bound test using the models established above, the data series on the selected variables should be tested to detect the presence of unit root and 7 BaProt = ɑ0 + i=1 p β IBaProt-i + i=0 q β iXt-i + Ut (1) where BaProt represents barley production, BaProt-i represents barley output supplied in year t-i, Xt-i represents explanatory variables in year t-i, t represents the time from 1981 to 2020, and β0, βi, … are coefficients of variables included in the model, and Ut is disturbance term. In this study, it was considered that the relationship between the independent and explanatory variables is expected to take the following functional form: BaProt = f (BaProt-1, BaPrit, BaArt, FertQt, CSMRFt, CGPMTt) (2) where BaProt is barley output measured in million tons; BaProt-1 is barley output in first- lag order, BaPrit is producer price of barley output in ETB, BaArt is land area allocated under barley cultivation, FertQt is fertilizer quantity used in barley production, CSMRFt is crop season mean rainfall in millimeters, and CGPMTt is crop growing period mean temperatures in degrees Celsius. By converting all the variables in Equation (2) into natural log form, the model is expressed as below: lnBaProt= β0 + β1lnBaProt-1 + β2lnBaPrit + β3lnBaArt + β4lnFertQt + β5lnCSMRFt + β6lnCGPMTt + εt (3) where lnCSMRFt is the log of crop season mean rainfall in mm, and lnCGPMTt is the log of crop growing period mean temperature in °C. In addition, εt represents the disturbance term. To generate some long-run relationships, Equation (3) is hereby modified as: lnBaProt = ɑ0 + α 1lnBaProt-i + α 2lnBaPrit-i + α 3lnBaArt-i + α 5lnFertt-i + α 6lnCSMRFt-i + α 8lnCGPMTt-i + εt-I In case the variables are found cointegrated, the model exemplifies the existence of an error correction representation. After establishing the above long-run relationship between variables, the Error Correction Model (ECM) can be derived from the ARDL model (Equation (4)) through simple linear transformation to find the short-run elasticity coefficients, which integrates short-run adjustments with long-run equilibrium. The short- run elasticity coefficients can be estimated using the following Dynamic ARDL Error Correction Model (ECM): ∆lnBaProt = β0 + β 1∆lnBaProt-i + β 2∆lnBaPrit-i + β 3∆lnBArt-i+ β 5∆lnFertt-i + β 6∆lnCSMRFt-i + β 8∆lnCGPMTt-i + ψiECT1-i+ ui (5) where ∆ represents the first difference while ψi is the coefficient of ECM for short-run dynamics. ECM shows the speed of adjustment in long-run equilibrium after a shock in the short-run. In this study, the investigator used the general to a specific approach to select an optimal lag length for the ARDL model. Before estimating the ARDL bound test using the models established above, the data series on the selected variables should be tested to detect the presence of unit root and 7 BaProt = ɑ0 + i=1 p β IBaProt-i + i=0 q β iXt-i + Ut (1) where BaProt represents barley production, BaProt-i represents barley output supplied in year t-i, Xt-i represents explanatory variables in year t-i, t represents the time from 1981 to 2020, and β0, βi, … are coefficients of variables included in the model, and Ut is disturbance term. In this study, it was considered that the relationship between the independent and explanatory variables is expected to take the following functional form: BaProt = f (BaProt-1, BaPrit, BaArt, FertQt, CSMRFt, CGPMTt) (2) where BaProt is barley output measured in million tons; BaProt-1 is barley output in first- lag order, BaPrit is producer price of barley output in ETB, BaArt is land area allocated under barley cultivation, FertQt is fertilizer quantity used in barley production, CSMRFt is crop season mean rainfall in millimeters, and CGPMTt is crop growing period mean temperatures in degrees Celsius. By converting all the variables in Equation (2) into natural log form, the model is expressed as below: lnBaProt= β0 + β1lnBaProt-1 + β2lnBaPrit + β3lnBaArt + β4lnFertQt + β5lnCSMRFt + β6lnCGPMTt + εt (3) where lnCSMRFt is the log of crop season mean rainfall in mm, and lnCGPMTt is the log of crop growing period mean temperature in °C. In addition, εt represents the disturbance term. To generate some long-run relationships, Equation (3) is hereby modified as: lnBaProt = ɑ0 + α 1lnBaProt-i + α 2lnBaPrit-i + α 3lnBaArt-i + α 5lnFertt-i + α 6lnCSMRFt-i + α 8lnCGPMTt-i + εt-I In case the variables are found cointegrated, the model exemplifies the existence of an error correction representation. After establishing the above long-run relationship between variables, the Error Correction Model (ECM) can be derived from the ARDL model (Equation (4)) through simple linear transformation to find the short-run elasticity coefficients, which integrates short-run adjustments with long-run equilibrium. The short- run elasticity coefficients can be estimated using the following Dynamic ARDL Error Correction Model (ECM): ∆lnBaProt = β0 + β 1∆lnBaProt-i + β 2∆lnBaPrit-i + β 3∆lnBArt-i+ β 5∆lnFertt-i + β 6∆lnCSMRFt-i + β 8∆lnCGPMTt-i + ψiECT1-i+ ui (5) where ∆ represents the first difference while ψi is the coefficient of ECM for short-run dynamics. ECM shows the speed of adjustment in long-run equilibrium after a shock in the short-run. In this study, the investigator used the general to a specific approach to select an optimal lag length for the ARDL model. Before estimating the ARDL bound test using the models established above, the data series on the selected variables should be tested to detect the presence of unit root and 7 BaProt = ɑ0 + i=1 p β IBaProt-i + i=0 q β iXt-i + Ut (1) where BaProt represents barley production, BaProt-i represents barley output supplied in year t-i, Xt-i represents explanatory variables in year t-i, t represents the time from 1981 to 2020, and β0, βi, … are coefficients of variables included in the model, and Ut is disturbance term. In this study, it was considered that the relationship between the independent and explanatory variables is expected to take the following functional form: BaProt = f (BaProt-1, BaPrit, BaArt, FertQt, CSMRFt, CGPMTt) (2) where BaProt is barley output measured in million tons; BaProt-1 is barley output in first- lag order, BaPrit is producer price of barley output in ETB, BaArt is land area allocated under barley cultivation, FertQt is fertilizer quantity used in barley production, CSMRFt is crop season mean rainfall in millimeters, and CGPMTt is crop growing period mean temperatures in degrees Celsius. By converting all the variables in Equation (2) into natural log form, the model is expressed as below: lnBaProt= β0 + β1lnBaProt-1 + β2lnBaPrit + β3lnBaArt + β4lnFertQt + β5lnCSMRFt + β6lnCGPMTt + εt (3) where lnCSMRFt is the log of crop season mean rainfall in mm, and lnCGPMTt is the log of crop growing period mean temperature in °C. In addition, εt represents the disturbance term. To generate some long-run relationships, Equation (3) is hereby modified as: lnBaProt = ɑ0 + α 1lnBaProt-i + α 2lnBaPrit-i + α 3lnBaArt-i + α 5lnFertt-i + α 6lnCSMRFt-i + α 8lnCGPMTt-i + εt-I In case the variables are found cointegrated, the model exemplifies the existence of an error correction representation. After establishing the above long-run relationship between variables, the Error Correction Model (ECM) can be derived from the ARDL model (Equation (4)) through simple linear transformation to find the short-run elasticity coefficients, which integrates short-run adjustments with long-run equilibrium. The short- run elasticity coefficients can be estimated using the following Dynamic ARDL Error Correction Model (ECM): ∆lnBaProt = β0 + β 1∆lnBaProt-i + β 2∆lnBaPrit-i + β 3∆lnBArt-i+ β 5∆lnFertt-i + β 6∆lnCSMRFt-i + β 8∆lnCGPMTt-i + ψiECT1-i+ ui (5) where ∆ represents the first difference while ψi is the coefficient of ECM for short-run dynamics. ECM shows the speed of adjustment in long-run equilibrium after a shock in the short-run. In this study, the investigator used the general to a specific approach to select an optimal lag length for the ARDL model. Before estimating the ARDL bound test using the models established above, the data series on the selected variables should be tested to detect the presence of unit root and (4) In case the variables are found cointegrated, the model exemplifies the existence of an error correction represen- tation. After establishing the above long-run relationship between variables, the Error Correction Model (ECM) can be derived from the ARDL model (Equation (4)) through simple linear transformation to find the short-run elasticity coefficients, which integrates short-run adjustments with long-run equilibrium. The short-run elasticity coefficients can be estimated using the following Dynamic ARDL Er- ror Correction Model (ECM): 7 BaProt = ɑ0 + i=1 p β IBaProt-i + i=0 q β iXt-i + Ut (1) where BaProt represents barley production, BaProt-i represents barley output supplied in year t-i, Xt-i represents explanatory variables in year t-i, t represents the time from 1981 to 2020, and β0, βi, … are coefficients of variables included in the model, and Ut is disturbance term. In this study, it was considered that the relationship between the independent and explanatory variables is expected to take the following functional form: BaProt = f (BaProt-1, BaPrit, BaArt, FertQt, CSMRFt, CGPMTt) (2) where BaProt is barley output measured in million tons; BaProt-1 is barley output in first- lag order, BaPrit is producer price of barley output in ETB, BaArt is land area allocated under barley cultivation, FertQt is fertilizer quantity used in barley production, CSMRFt is crop season mean rainfall in millimeters, and CGPMTt is crop growing period mean temperatures in degrees Celsius. By converting all the variables in Equation (2) into natural log form, the model is expressed as below: lnBaProt= β0 + β1lnBaProt-1 + β2lnBaPrit + β3lnBaArt + β4lnFertQt + β5lnCSMRFt + β6lnCGPMTt + εt (3) where lnCSMRFt is the log of crop season mean rainfall in mm, and lnCGPMTt is the log of crop growing period mean temperature in °C. In addition, εt represents the disturbance term. To generate some long-run relationships, Equation (3) is hereby modified as: lnBaProt = ɑ0 + α 1lnBaProt-i + α 2lnBaPrit-i + α 3lnBaArt-i + α 5lnFertt-i + α 6lnCSMRFt-i + α 8lnCGPMTt-i + εt-I In case the variables are found cointegrated, the model exemplifies the existence of an error correction representation. After establishing the above long-run relationship between variables, the Error Correction Model (ECM) can be derived from the ARDL model (Equation (4)) through simple linear transformation to find the short-run elasticity coefficients, which integrates short-run adjustments with long-run equilibrium. The short- run elasticity coefficients can be estimated using the following Dynamic ARDL Error Correction Model (ECM): ∆lnBaProt = β0 + β 1∆lnBaProt-i + β 2∆lnBaPrit-i + β 3∆lnBArt-i+ β 5∆lnFertt-i + β 6∆lnCSMRFt-i + β 8∆lnCGPMTt-i + ψiECT1-i+ ui where ∆ represents the first difference while ψi is the coefficient of ECM for short-run dynamics. ECM shows the speed of adjustment in long-run equilibrium after a shock in the short-run. In this study, the investigator used the general to a specific approach to select an optimal lag length for the ARDL model. 7 BaProt = ɑ0 + i=1 p β IBaProt-i + i=0 q β iXt-i + Ut (1) where BaProt represents barley production, BaProt-i represents barley output supplied in year t-i, Xt-i represents explanatory variables in year t-i, t represents the time from 1981 to 2020, and β0, βi, … are coefficients of variables included in the model, and Ut is disturbance term. In this study, it was considered that the relationship between the independent and explanatory variables is expected to take the following functional form: BaProt = f (BaProt-1, BaPrit, BaArt, FertQt, CSMRFt, CGPMTt) (2) where BaProt is barley output measured in million tons; BaProt-1 is barley output in first- lag order, BaPrit is producer price of barley output in ETB, BaArt is land area allocated under barley cultivation, FertQt is fertilizer quantity used in barley production, CSMRFt is crop season mean rainfall in millimeters, and CGPMTt is crop growing period mean temperatures in degrees Celsius. By converting all the variables in Equation (2) into natural log form, the model is expressed as below: lnBaProt= β0 + β1lnBaProt-1 + β2lnBaPrit + β3lnBaArt + β4lnFertQt + β5lnCSMRFt + β6lnCGPMTt + εt (3) where lnCSMRFt is the log of crop season mean rainfall in mm, and lnCGPMTt is the log of crop growing period mean temperature in °C. In addition, εt represents the disturbance term. To generate some long-run relationships, Equation (3) is hereby modified as: lnBaProt = ɑ0 + α 1lnBaProt-i + α 2lnBaPrit-i + α 3lnBaArt-i + α 5lnFertt-i + α 6lnCSMRFt-i + α 8lnCGPMTt-i + εt-I In case the variables are found cointegrated, the model exemplifies the existence of an error correction representation. After establishing the above long-run relationship between variables, the Error Correction Model (ECM) can be derived from the ARDL model (Equation (4)) through simple linear transformation to find the short-run elasticity coefficients, which integrates short-run adjustments with long-run equilibrium. The short- run elasticity coefficients can be estimated using the following Dynamic ARDL Error Correction Model (ECM): ∆lnBaProt = β0 + β 1∆lnBaProt-i + β 2∆lnBaPrit-i + β 3∆lnBArt-i+ β 5∆lnFertt-i + β 6∆lnCSMRFt-i + β 8∆lnCGPMTt-i + ψiECT1-i+ ui where ∆ represents the first difference while ψi is the coefficient of ECM for short-run dynamics. ECM shows the speed of adjustment in long-run equilibrium after a shock in the short-run. In this study, the investigator used the general to a specific approach to select an optimal lag length for the ARDL model. 7 BaProt = ɑ0 + i=1 p β IBaProt-i + i=0 q β iXt-i + Ut (1) where BaProt represents barley production, BaProt-i represents barley output supplied in year t-i, Xt-i represents explanatory variables in year t-i, t represents the time from 1981 to 2020, and β0, βi, … are coefficients of variables included in the model, and Ut is disturbance term. In this study, it was considered that the relationship between the independent and explanatory variables is expected to take the following functional form: BaProt = f (BaProt-1, BaPrit, BaArt, FertQt, CSMRFt, CGPMTt) (2) where BaProt is barley output measured in million tons; BaProt-1 is barley output in first- lag order, BaPrit is producer price of barley output in ETB, BaArt is land area allocated under barley cultivation, FertQt is fertilizer quantity used in barley production, CSMRFt is crop season mean rainfall in millimeters, and CGPMTt is crop growing period mean temperatures in degrees Celsius. By converting all the variables in Equation (2) into natural log form, the model is expressed as below: lnBaProt= β0 + β1lnBaProt-1 + β2lnBaPrit + β3lnBaArt + β4lnFertQt + β5lnCSMRFt + β6lnCGPMTt + εt (3) where lnCSMRFt is the log of crop season mean rainfall in mm, and lnCGPMTt is the log of crop growing period mean temperature in °C. In addition, εt represents the disturbance term. To generate some long-run relationships, Equation (3) is hereby modified as: lnBaProt = ɑ0 + α 1lnBaProt-i + α 2lnBaPrit-i + α 3lnBaArt-i + α 5lnFertt-i + α 6lnCSMRFt-i + α 8lnCGPMTt-i + εt-I In case the variables are found cointegrated, the model exemplifies the existence of an error correction representation. After establishing the above long-run relationship between variables, the Error Correction Model (ECM) can be derived from the ARDL model (Equation (4)) through simple linear transformation to find the short-run elasticity coefficients, which integrates short-run adjustments with long-run equilibrium. The short- run elasticity coefficients can be estimated using the following Dynamic ARDL Error Correction Model (ECM): ∆lnBaProt = β0 + β 1∆lnBaProt-i + β 2∆lnBaPrit-i + β 3∆lnBArt-i+ β 5∆lnFertt-i + β 6∆lnCSMRFt-i + β 8∆lnCGPMTt-i + ψiECT1-i+ ui (5) where ∆ represents the first difference while ψi is the coefficient of ECM for short-run dynamics. ECM shows the speed of adjustment in long-run equilibrium after a shock in the short-run. In this study, the investigator used the general to a specific approach to select an optimal lag length for the ARDL model. Before estimating the ARDL bound test using the models established above, the data series on the selected variables should be tested to detect the presence of unit root and 7 BaProt = ɑ0 + i=1 p β IBaProt-i + i=0 q β iXt-i + Ut (1) where BaProt represents barley production, BaProt-i represents barley output supplied in year t-i, Xt-i represents explanatory variables in year t-i, t represents the time from 1981 to 2020, and β0, βi, … are coefficients of variables included in the model, and Ut is disturbance term. In this study, it was considered that the relationship between the independent and explanatory variables is expected to take the following functional form: BaProt = f (BaProt-1, BaPrit, BaArt, FertQt, CSMRFt, CGPMTt) (2) where BaProt is barley output measured in million tons; BaProt-1 is barley output in first- lag order, BaPrit is producer price of barley output in ETB, BaArt is land area allocated under barley cultivation, FertQt is fertilizer quantity used in barley production, CSMRFt is crop season mean rainfall in millimeters, and CGPMTt is crop growing period mean temperatures in degrees Celsius. By converting all the variables in Equation (2) into natural log form, the model is expressed as below: lnBaProt= β0 + β1lnBaProt-1 + β2lnBaPrit + β3lnBaArt + β4lnFertQt + β5lnCSMRFt + β6lnCGPMTt + εt (3) where lnCSMRFt is the log of crop season mean rainfall in mm, and lnCGPMTt is the log of crop growing period mean temperature in °C. In addition, εt represents the disturbance term. To generate some long-run relationships, Equation (3) is hereby modified as: lnBaProt = ɑ0 + α 1lnBaProt-i + α 2lnBaPrit-i + α 3lnBaArt-i + α 5lnFertt-i + α 6lnCSMRFt-i + α 8lnCGPMTt-i + εt-I In case the variables are found cointegrated, the model exemplifies the existence of an error correction representation. After establishing the above long-run relationship between variables, the Error Correction Model (ECM) can be derived from the ARDL model (Equation (4)) through simple linear transformation to find the short-run elasticity coefficients, which integrates short-run adjustments with long-run equilibrium. The short- run elasticity coefficients can be estimated using the following Dynamic ARDL Error Correction Model (ECM): ∆lnBaProt = β0 + β 1∆lnBaProt-i + β 2∆lnBaPrit-i + β 3∆lnBArt-i+ β 5∆lnFertt-i + β 6∆lnCSMRFt-i + β 8∆lnCGPMTt-i + ψiECT1-i+ ui (5) where ∆ represents the first difference while ψi is the coefficient of ECM for short-run dynamics. ECM shows the speed of adjustment in long-run equilibrium after a shock in the short-run. In this study, the investigator used the general to a specific approach to select an optimal lag length for the ARDL model. Before estimating the ARDL bound test using the models established above, the data series on the selected variables should be tested to detect the presence of unit root and (5) where ∆ represents the first difference while ψi is the co- efficient of ECM for short-run dynamics. ECM shows the speed of adjustment in long-run equilibrium after a shock in the short-run. In this study, the investigator used the general to a specific approach to select an optimal lag length for the ARDL model. Before estimating the ARDL bound test using the mod- els established above, the data series on the selected vari- ables should be tested to detect the presence of unit root and long-run cointegration. To this end, an Augmented Dickey-Fuller (ADF) and Philips-Perron (PP) tests have been considered the best approach [30,31] and used the mod- els to test the presence of unit root in the data series. To estimate the bound tests, all the variables included in the model must be stationary at I(0), I(1), or both. Neverthe- less, researchers noted that the presence of unit root in data series implies that the analyst may obtain spurious results from analyzing them at their original level [32,33]. Next to the stationarity test, a cointegration test has been conducted to detect the presence of a stable equi- librium relationship between the variables included in the model as proposed in Enders [34]. If the presence of cointegration is confirmed with the model for at least two I(1) series and some I(0), the variables can be added to the ARDL model for the estimation which may not alter the I(0) characteristics of the error term. In this study, cointegration analysis was carried out using the Johansen procedure as recommended by Akter and Hong [35], which first defines an unrestricted vector autoregression (VAR). All of the analyses have been conducted using Eviews 9 Econometric Software. 3. Results and Discussion 3.1 Results of Preliminary Time Series, Specifica- tion, and Robustness Tests Before the estimation of the ARDL model, appropri- ate tests have been carried out to detect the existence of unit root and long-run co-integration in the data series. Table 1 presents the results of unit root tests conducted on the time series data using ADF and PP approaches. The results imply that log barley output and log fertilizer quantity used in barley production exhibited stationarity at the first-order difference (I(1)). Conversely, log producer barley price, log area under barley crop, log CSMRF, and log CGPMT were stationary at level (I(0)). The result, therefore, demonstrated a mixture of level-order (I(0)) and first-order (I(1)) integration of variables [36]. Whenever the time series data exhibit a mixture of I(0) and I(1), most investigators propose to apply ARDL modeling as the best approach to estimate the coefficients of the parameter included in the models [37]. To apply the ARDL approach, cointegration bounds test, model sta- bility test, and variance error correction model (VECM) has to be conducted to test the presence of long-term co- integration, models’ goodness of fit, presence of serial correlation, and model misspecification [37]. Table 2 presents the outcomes of the cointegration bound test. It can be seen from the table that a linear com- bination of the variables in the regression model was sta- tionary since the F-statistics exceeds the upper bound at the 5% critical value. This implies that barley output and its determinants are cointegrated, exemplifying the exist- ence of a long-run relationship among the variables in the model. 43 Research on World Agricultural Economy | Volume 03 | Issue 03 | September 2022 To test the robustness of the ARDL model, diagnostic tests such as non-normality, serial correlation, and hetero- scedasticity were conducted. The results of the diagnostic tests for the barley output response equation are presented in Table 3. It can be seen from the table that the p-values for normality (Jarque-Bera), serial correlation (Breush– Godfrey Lagrange Multiplier (LM)), and heteroscedastic- ity are greater than a 5% level of significance. The results imply that the residuals are normally distributed; there is no evidence of serial correlation; no autoregressive condi- tional heteroscedasticity (ARCH). Table 3. Residual properties of barley output response equation Type of test Test statistic Test statistic value Probability Normality test - Histogram Jarque- Bera 3.43526 0.17794 Serial Correlation (LM ) Obs*R- squared 13.4639 0.05720 Heteroscedasticity (ARCH ) Obs*R- squared 11.3876 0.5784 In addition to the above diagnostic tests, the stability of long-run estimates has been tested using the cumula- tive sum of recursive residuals (CUSUM) and cumulative squares of recursive residuals (CUSUMQS) test. Table 4 shows CUSUM stability test results. As can be seen from the table, the model does not suffer from any form of mis- specification. Equally, the plot of CUSUM test shown in Figure 2 reveals that the estimated parameters are stable over the observation period at a 5% level of significance. Table 4. CUSUM stability test results Dependent Variable F – statistic Probability Conclusion Log barley output 0.46382 0.6349 No indication of misspecification -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 98 00 02 04 06 08 10 12 14 16 18 20 CUSUM of Squares 5% Significance Figure 3. Plot of cumulative sum of squares of re- cursive residuals. 3.2 Impact of Climate and Non-Climate Param- eters on Barley Output Supply Response This study was organized to examine the long-run and short-run impacts that climatic and non-climatic factors exert on barley output supply response. An ARDL model selected for the study has been estimated for both climatic (CSMRF and CGPMT) and non-climatic parameters (lagged barley output, barley producer price, land area allocated for barley cultivation, and quantity of fertilizer Table 1. Time Series Unit Root Test Results for Barley Output and Related Independent Variables Variable ADF PP ResultLevel First Difference Level First Difference t-Stat P-value C. Value t-Stat P-value C. Value t-Stat P-value C. Value t-Stat P-value C. Value LnBaPro –2.2735 0.4378 –4.2187 –8.6119*** 0.0001 –4.2119 –2.0962 0.5317 –4.2119 –9.3237*** 0.0000 –4.2191 I(1) LnBaPri –2.2660*** 0.4416 –4.2119 –5.7388*** 0.0002 –4.2191 –2.2649*** 0.4422 –4.2119 –6.2363*** 0.0000 –4.2119 I(0) LnBaAr –3.6455** 0.0387 –3.5297 –8.5007 0.0000 –4.2191 –3.6395** 0.0392 –3.5298 –19.265*** 0.0000 –4.2191 I(0) LnFert –2.9416 0.1613 –3.1964 7.2393*** 0.0000 –4.2191 –2.9228* 0.1668 –3.1964 –13.1899*** 0.0000 –4.2191 I(1) LnCSMRF –4.9210** 0.0015 –4.2119 –6.5636*** 0.0000 –4.2436 –4.8858*** 0.0017 –4.2119 –20.979*** 0.0000 –4.2191 I(0) LnCGPMT –18.986*** 0.0000 –4.2119 –20.0798 0.0000 –4.2191 –16.239 0.0000 –4.2119 –54.222*** 0.0000 –4.2191 I(0) Note: ***, ** and * indicates 1%, 5% and 10% level of significance. Table 2. Estimated cointegration equations for barley output response Dependent variable Type of test Test statistics Critical values Conclusion Barley output response Wald test 4.2844** 4.130 Long-run cointegration exists Note: ** Statistically Significant at 5% level 44 Research on World Agricultural Economy | Volume 03 | Issue 03 | September 2022 consumed on barley production). Main-season/meher rainfall and improved barley seed have been initially in- corporated into the model but dropped since both have high serial autocorrelation and multicollinearity with other variables. Equally, the irrigated area under barley cultiva- tion was dropped because the total land area allocated under barley cultivation encompassed an irrigated area as well. The effect of the irrigated area under barley cultiva- tion should be treated separately from the total land area allocated under barley crop to avoid double counting and to know their impact contribution individually. The ARDL model with lag length (1, 0, 0, 1, 0, 0) was selected as optimum to estimate the regression coefficients for the variables included in the model. It was found that the ARDL regression model demonstrated good fitness to the barley output supply time series data, with high values of adjusted R2 (0.779). Based on the value of adjusted R2, the explanatory variables explained 77.9% of the variation in barley output supply. Furthermore, the Durban-Watson showed no evidence of serial autocorrelation. The F-test does not show the presence of any heteroscedasticity of the residual. The tests, therefore, exemplify that the model becomes viable and fits at lag length 1 and first-order dif- ferences. Table 5 presents the estimated regression coefficients of the ARDL model for barley output supply against the de- terminant variables. The result shows that climatic factors had a positive impact on the current barley output supply both during previous and current years. In this respect, the current year’s barley output supply was affected positively and significantly by the amount of zero-order difference crop season mean rainfall (CSMRF) and crop growing pe- riod mean temperature (CGPMT). The result implies that a 1% increase in CSMRF and CGPMT individually boosts barley output supply by 0.47% and 2.27% respectively. This shows that crop season rainfall is among the main determinants of barley output supply in Ethiopia. This result is inversely related to research results re- corded globally by other investigators, the majority of which demonstrated increasing temperature associated with decreasing rainfall [38]. The current study finding supports the findings of researchers who reported that a warming temperature followed by an increase in the amount of rainfall had a positive impact on barley output supply. As can be seen from Figure 3, crop growing period mean temperature in this study exhibited a significant (at 1% level) rising trend in barley growing areas followed by increasing crop season mean rainfall in the same areas. The positive impact of temperature can be explained that as the average surface temperature rise, more evaporation arises, which increases the overall rainfall while reaching the highland and mid-highland areas of Ethiopia. Some study reports confirm that the so-called “El Nino-Southern Oscillation (ENSO)” and the moist wind coming from the Atlantic and Indian Oceans influence the western, south- eastern, central, and northern highlands of Ethiopia; which bring moisture from the oceans [39]. This finding is also in conformity with Fischer and Velthuizen [40] who in their examination of the impact of climate change on Kenya re- ported that higher temperatures exert a positive impact in the highland areas. Among the non-climatic inputs, regression coefficients were estimated for lagged barley output (previous year’s output), current year’s barley price, land area under bar- ley cultivation, and fertilizer used in barley production. The estimates demonstrated that the current year’s barley price, land area under barley cultivation, and fertilizer used had a positive impact on the current year’s barley crop output supply. However, only the producer price of barley had a significant impact on the current year’s bar- ley production. The result implies that a 1% increase in the current year’s producer price of barley will increase barley output supply by 0.7%. Equally, the previous year’s (zero-order difference) producer price demonstrated a highly significant positive impact on the current year’s barley output supply in which a 1% increase in producer price of barley last year will increase the current year’s barley output by 0.82%. Conversely, the use of fertilizer in its first-order lag (previous year) had demonstrated a negative and highly significant (at 5% level) impact on barley output supply, in which a 1% change (increase or decrease) in fertilizer quantity used leads to a decrease of barley output by 0.0.17%. Furthermore, lagged barley output (first-order lag) exerted a positive and significant (at 1% level) impact on the current year’s barley output sup- ply. The result shows that a 1% change in the quantity of previous years’ barley output would decrease the volume of the current year’s barley output by 0.38%. From the results of non-climatic factors, it can be concluded that the current barley output supply is posi- tively and significantly responsive to both the current and previous year’s producer prices. Barley producer price change or price incentives announced before land area allocation to specific crops had a significant and positive contributions in boosting the current year’s barley output supply. 45 Research on World Agricultural Economy | Volume 03 | Issue 03 | September 2022 Figure 4. Trend of CGP Mean Temperature and CSM Rainfall in Barley Growing Areas Table 5. Estimates of Regression Coefficients for ARDL Model of Barley Output Supply Response Variable Coefficient Std. Error t-Statistic Prob. Cons –3.24905 3.56716 –0.91083 0.3714 LNBAPRO(-1) –0.2928*** 0.09994 –2.93004 0.0073 LNBAPRI(-1) 0.6955*** 0.16548 4.20280 0.0003 LNBAAR(-1) 0.0252 0.03266 0.77163 0.4479 LNFERT(-1) 0.1054 0.07360 1.43225 0.1650 LNCSMRF(-1) 0.2319 0.21385 1.08438 0.2890 LNCGPMT(-1) 0.4676 1.05533 0.44306 0.6617 D(LNBAPRO(-1)) –0.3787*** 0.11631 –3.25630 0.0034 D(LNBAPRI) 0.8220*** 0.10730 7.66077 0.0000 D(LNBAAR) -0.0278 0.06776 –0.40977 0.6856 D(LNFERT) 0.0193 0.06897 0.27907 0.7826 D(LNFERT(-1)) –0.174** 0.07787 –2.23457 0.0350 D(LNCSMRF) 0.4715*** 0.14154 3.33140 0.0028 D(LNCGPMT) 2.26702** 0.85321 2.65706 0.0138 R-squared 0.8565 Mean dependent var 0.02455 Adjusted R-squared 0.7787 S.D. dependent var 0.13759 S.E. of regression 0.0647 Akaike info criterion –2.36011 Sum squared resid 0.1006 Schwarz criterion –1.75679 F-statistic 11.0159 Hannan-Quinn criter. –2.14545 Prob(F-statistic) 0.0000 Durbin-Watson stat 2.28889 Note: *, ** & *** indicate significance at 10%, 5% and 1% level. 46 Research on World Agricultural Economy | Volume 03 | Issue 03 | September 2022 Since the cointegration test confirmed the presence of long-run cointegration among the variables included in the model, long-run elasticity coefficients have been esti- mated for the barley output determinant variables. Table 6 presents estimated long-run elasticity coefficients of the climate and non-climate variables included in the barley output ARDL model with a lag length of (1, 0, 0, 1, 0, 0). The estimated elasticity coefficients of all climatic and non-climatic variables, except the log of lagged barley output, demonstrated a positive relationship with the de- pendent variable in the long-run. Nevertheless, only bar- ley producer price had demonstrated a positively signifi- cant impact on the supply of barley output in the long-run. This implies that a 1% rise in producer price of barley out- put will boost barley output supply by 0.54%. Conversely, log barley output included as lagged (first-order lag) ex- planatory variable exerted a negative and significant (1% level) impact on the current year’s barley output supply in the long-run. This implied that a 1% change (decrease or increase) in first-order lagged output will decrease barley output by 0.23%. The study result exhibited that average crop season rainfall and crop growing period mean temperature exhib- ited a positive influence on the supply of barley output, al- though it is statistically non-significant. This implies that climate parameters have minimal impact on the supply of barley output in the long-run. The reason for this result is that barley is grown during the main rainy season when rainfall is relatively plentiful and the temperature is rela- tively cool. Furthermore, barley is grown in the highlands and mid-highlands where the temperature is naturally cool and has a moderate to a high amount of rainfall. The Error Correction Model which engages short-term fluctuations in the long-run has been estimated employ- ing the ARDL bounds test approach. The outcomes of the elasticity coefficients for the variables with lag length (1, 0, 0, 1, 0, 0) model are presented in Table 7. The results show that CSMRF and CGPMT had a positive and signifi- cant influence on the current barley output supply in the short-run. This indicates that a 1% increase in CSMRF and CGPMT would boost barley output supply by 0.34% and 0.32% respectively. Furthermore, non-climatic factors included in the mod- el showed mixed results in the short-run. Accordingly, the log producer price of barley (at zero-order difference) showed a positively significant effect on barley output supply in the short-run. This indicates that a 1% increase in log producer price at zero-order difference would lead to an increase in barley output by 0.66%. This specifically implies that barley output supply is highly responsive to any strategy of price incentive announced before re- allocation of the land area towards barley cultivation in the short-run. Conversely, the elasticity coefficients of log area and log fertilizer used at zero-order differences demonstrated a negative effect on barley output supply in the short-run. Nevertheless, only fertilizer quantity used in barley production had a significant effect on barley output supply. The result indicates that a 1% increase or decrease in fertilizer quantity used would decrease the quantity of barley output supply by 0.13% in the short-run. The result implies that increased use of fertilizer in lagged period (last year) will reduce the current year’s supply of barley out- put. This was achieved since farmers in Ethiopia do not use fertilizer as per recommendations of the extension ser- vice. Equally, any incentivized barley price in zero-order difference will affect barley output supply positively, i.e. price incentive announced during the previous year will encourage producers to allocate more land and boost bar- ley output supply. On the other hand, the lagged error correction term which captures the speed of adjustment towards long-run equilibrium exemplified the correct sign and magnitude. The speed of adjustment was found to be -0.948 which is highly significant (1% level) and indicates the speed of adjustment to be back to the long-run equilibrium after a short-run shock on barley crop output and climate vari- ables. It is crucial to note that the coefficient of –0.948 precisely means that it takes 1.05 years (1/0.948) for the barley crop output to return to its equilibrium position following a shock. The estimated coefficient (ECTt-1) also portrays that 94.8% of the disequilibrium created will be corrected within 1 year. 3.3 Comparison of the Study with Others Studies The results of this study are analogous to the study re- sults of various researchers based in the country as well as in other countries. Among these researchers, Dumrul and Kilicarslan [41] in their study on the economic impacts of changes in climate on agriculture in Turkey reported that log average temperature had a positive and significant im- pact on agricultural GDP. In contrast to theory, positive ef- fects of warmer temperatures on selected crops have also been demonstrated by Lobell, et al. [42] and Schlenker and Roberts [43], although only below the threshold of tempera- ture. Equally, Chandio, et al. [44] in their study on the rela- tionship between climatic and wheat production in Turkey reported that rainfall has exerted a positive influence on wheat production in the long-run, although insignificant. The result implies that a 1% increase in precipitation level would lead to an increase in wheat production by 0.06% in the long-run. However, their findings are contrary to 47 Research on World Agricultural Economy | Volume 03 | Issue 03 | September 2022 the result of this study since average temperature had negatively and significantly impacted wheat production in Turkey. This implies that a 1% rise in the average level of temperature will lead to a decline in wheat production by 0.29% in the long-run. This implies that a rising level of mean temperature in Turkey had adversely affected the production of wheat. It is evident that a decrease in wheat production leads to reduced growth in the agricultural sec- tor and creates a challenge to food security in the country. On the other hand, Ketema [45] in his study on determinants of agricultural output in Ethiopia reported that rainfall had a positive and significant impact on agricultural output, which is similar to the current study finding. The result implied that a 1% increase in the amount of rainfall boosts agricultural output by 0.56%. The study findings of Taye, et al. [46] are also congruent with the result of the current study; studying the impacts that a changing climate and fertilizer exert on barley production in Ethiopia reported that rainfall had a positive and significant impact on bar- ley production at zero-order difference both in the short- and long-run. These indicate that a 1% increase in the amount of rainfall during the short- and long-run boosts barley production by 0.03% and 0.41% respectively. The current study result is further consistent with that Table 6. Long-run estimated elasticities of parameters in barley output supply model Variable Coefficient Std. Error t-Statistic P-value Constant –2.51316 2.73879 –0.91762 0.3679 lnBaPro(-1) –0.22649*** 0.059793 –3.788003 0.0009 lnBaPri 0.53797*** 0.102486 5.249228 0.0000 lnBaAr 0.01949 0.024653 0.790720 0.4369 lnFert 0.08154 0.05632 1.44784 0.1606 LnCSMRF 0.17938 0.168013 1.067625 0.2963 LnCGPMT 0.36167 0.810883 0.446021 0.6596 R-squared 0.85646 Mean dependent var 0.02455 Adjusted R-squared 0.77872 S.D. dependent var 0.13759 S.E. of regression 0.06472 Akaike info criterion –2.36011 Sum squared resid 0.10054 Schwarz criterion –1.75679 F-statistic 11.01587 Hannan-Quinn criter. –2.14545 Prob(F-statistic) 0.000000 Durbin-Watson stat 2.28889 Note: ** indicates significance at 5% level Table 7. Short-run elasticities of variables in barley output dynamic ECM model Variable Coefficient Std. Error t-Statistic Prob. Cons 0.16093 0.15513 1.03739 0.3073 ECT(-1) –0.948*** 0.06303 –15.0414 0.0000 D(LNBAPRI) 0.6646*** 0.10845 6.12856 0.0000 D(LNBAAR) –0.12303 0.07689 –1.59999 0.1194 D(LNFERT) –0.13170* 0.07390 –1.78212 0.0842 D(LNCSMRF) 0.33554** 0.12665 2.64937 0.0124 D(LNCGPMT) 0.32123* 0.16913 1.89927 0.0666 R-squared 0.65255 Mean dependent var 0.02035 Adjusted R-squared 0.58741 S.D. dependent var 0.13827 R-squared 0.65255 Mean dependent var 0.02035 Adjusted R-squared 0.58741 S.D. dependent var 0.13827 S.E. of regression 0.08882 Akaike info criterion –1.84334 Sum squared resid 0.25243 Schwarz criterion –1.54475 F-statistic 10.0168 Hannan-Quinn criter. –1.73621 Prob(F-statistic) 0.000003 Durbin-Watson stat 2.27756 Note: *, ** and *** indicates significance at 10%, 5% and 1% level, respectively 48 Research on World Agricultural Economy | Volume 03 | Issue 03 | September 2022 of Chandio, et al. [47], who, in their study on short- and long-run impacts exerted by changing climate on the pro- duction of agricultural outputs in China have reported that log temperature and log rainfall had a positive influence on agricultural production in the short-run, although sta- tistically insignificant. Equally, Taye, et al. [46] who studied the impact of change in climate and fertilizer use on the production of barley output in Ethiopia reported that pre- cipitation and rainfall have a positive and significant influ- ence on the supply of barley output. The results indicated that a 1% rise in current precipitation increases barley output supply by 2.8% in the long-run. Furthermore, the finding related to the producer price of barley is similar to that of Elbeydi, et al. [48], who in their study on the re- sponse of barley in Libya, reported that the coefficient of producer price of barley is positive (0.543) and significant in the long-run. Conversely, Taye, et al. [46] reported that fertilizers (DAP and UREA) demonstrated a negatively significant influence on the production of barley output in the long run. This indicates that a 1% increase in the use of DAP and UREA fertilizers decreases the supply of barley output by 28.8% and 3.4% respectively in the long- run. Similarly, Taye, et al. [46] in their study on the impacts exerted by changing climate and fertilizer use on the production of barley output in Ethiopia exemplified that current year barley output is negatively and significantly affected by the use of current year DAP fertilizer in the short-run. This implies that a 1% increase in the use of current DAP fertilizer would decrease barley output by 4.44% in the short-run. Conversely, they reported that barley production is affected positively by the current and previous year’s (first-order lag) quantity of UREA ferti- lizers consumed in the short-run. In this respect, every 1% rise in the use of current and previous year’s UREA fertilizer boosts barley output supply by 6.87% and 6.57% respectively. 3.4 Implication and Explanation of Findings The study results demonstrated that climatic factors had a positive impact on the current barley output supply both during previous and current years. In this context, the current year’s barley output supply was affected posi- tively and significantly by the amount of zero-order dif- ference average crop season rainfall (CSMRF) and crop growing period mean temperature (CGPMT). This result is inversely related to the research results recorded glob- ally by other investigators, the majority of which reported that increasing temperature is associated with decreasing rainfall [38]. The current study result supports the findings of researchers who reported that a warming temperature followed by an increase in the amount of rainfall had a positive impact on barley output supply. The positive impact of temperature on barley output supply can be explained by the fact that as the average temperature on the earth’s surface rise, more evaporation occurs, which in turn, increases overall rainfall mostly while reaching the highland and mid-highland areas of Ethiopia. Study reports confirm that El Nino-Southern Oscillation (ENSO) and the Atlantic and Indian Oceans influence the occurrence of rainfall in the western, south- eastern, central, and northern highlands of Ethiopia; which bring moisture from the oceans [39]. However, the study result exhibited that CSMRF and CGPMT had a positive impact on the supply of barley output, which implies that climate parameters have mini- mal impact on the supply of barley output in the long-run. This was realized since barley is mainly grown during the main rainy season when rainfall is relatively plentiful and the temperature is cool. Furthermore, barley is grown in the highlands and mid-highlands where the temperature is naturally cool and has a moderate to a high amount of rainfall. In the short-run, CSMRF and CGPMT revealed a positively significant influence on the current barley out- put supply indicating that a 1% increase in CSMRF and CGPMT will boost barley output supply by 0.34% and 0.32% respectively. Similar investigations demonstrated that previous and current year’s barley producer prices had a positively sig- nificant influence on the current year’s barley production. From this result, it can be concluded that barley output supply is positively and significantly responsive to both current and previous year’s (first-order lag) own producer prices. Change (increase) in barley producer price or price incentives announced before land area allocation to spe- cific crops had a significant and positive contribution to boost the current year’s barley output supply. Conversely, the study result explored that use of fertilizer in its first-order lag (previous year) had exerted a negatively significant (at a 5% level) influence on barley output supply. On the other hand, the producer price of barley dem- onstrated a positively significant influence on the supply of barley output in both the long- and short-term. This specifically exemplifies that barley output supply is highly responsive to any strategy of price incentive announced before re-allocation of the land area towards barley cul- tivation. Conversely, the study result exemplified that fertilizer used at zero-order difference had a negatively significant influence on barley output supply in the short- run, implying that increased use of fertilizer in lagged period (last year) will reduce the current year’s barley output supply. This may be due to the inappropriate use of 49 Research on World Agricultural Economy | Volume 03 | Issue 03 | September 2022 fertilizers among farmers in Ethiopia. 4. Conclusions This study aimed to investigate the determinant factors influencing barley output supply response in Ethiopia. The study applied an ARDL model proposed by Pesaran, et al. [25] to examine the impact of climatic and non-climate input variables on barley output supply responses. The study used secondary time series data covering the period from 1981-2020. The study results demonstrated that climatic factors had a positive impact on the current barley output supply both during zero-order lag (previous) and current years. In this context, the current year’s barley output supply was affected positively and significantly by the amount of zero-order lag CSMRF and CGPMT. The result is inversely related to research results recorded globally by other investigators, the majority of which reported that rising temperature is associated with decreasing rainfall [38]. The current study result supports the findings of research- ers who reported that a warming temperature followed by an increase in the amount of rainfall had a positive impact on barley output supply. The positive impact of tempera- ture on barley output supply can be explained by the fact that as average surface temperature rise, more evaporation would be created which increases overall rainfall while reaching the highland and mid-highland areas of Ethiopia. Studies by Conway [39] confirm that El Nino-Southern Os- cillation (ENSO) and moist winds from the Atlantic and Indian Oceans influences the occurrence of rainfall in the western, southeastern, central, and northern highlands of Ethiopia; which bring moisture from the oceans. Furthermore, the study result exhibited that CSMRF and CGPMT had a positive impact on the supply of barley output in the long-run, although non-significant, which implies that climate parameters have got a minimal im- pact on the supply of barley output. This result has been realized because barley crop is mainly grown during the main rainy season when rainfall is relatively plentiful and the temperature is cool. Furthermore, barley is grown in the highlands and mid-highlands where the temperature is naturally cool and has a moderate to a high amount of rainfall. Conversely, CSMRF and CGPMT revealed a pos- itively significant influence on current barley production in the short-run, indicating that a 1% increase in CSMRF and CGPMT will boost barley output supply by 0.34% and 0.32% respectively. Similar investigations on non-climatic variables dem- onstrated that the previous year (first-order lag) and the current year’s barley producer price have had a positively significant influence on the current year’s barley produc- tion. From this result, it can be concluded that barley output supply is positively and significantly responsive to both current and previous year’s (first-order lag) producer prices. An increase in barley producer price or price in- centives announced before land area allocation to specific crops had a significant and positive contribution to boost the current year’s barley output supply. Conversely, the study result explored that use of fertilizers in its first-order lag (previous year) induced a negatively significant (at 5% level) influence on barley production. On the other hand, the producer price of barley dem- onstrated a positively significant effect on the supply of barley output both in the long- and short-run, specifically indicating that barley output supply is greatly responsive to any strategy of price incentive announced ahead of re- allocation of the land area towards barley cultivation. Conversely, the study result exemplified that fertilizer used at zero-order difference had a negatively significant effect on barley output supply in the short-run, which im- plies that increased use of fertilizer in lagged period (last year) will reduce the current year’s barley output supply. This may be due to the inappropriate use of fertilizers among farmers in Ethiopia. In conclusion, the elasticity estimates of climatic and non-climatic variables presented in this study can be a useful addition to the repository of knowledge on the sup- ply elasticity of agricultural commodities in the country at an aggregate level. The results can also be used to design appropriate mitigation and adaptation strategies and meas- ures in the future. Acknowledgments The author is grateful to the different Institutions that made available the dataset used in this study. Funding The article did not have a donation or other funding sources. The main author has covered all the expenses in- volved, including the purchase of meteorology data from NMA. Conflict of Interest The author declares that there are no competing inter- ests. Declaration on Non-Submission or Published with Other Journals The author declares that this submission has not been previously published, nor is it before another journal for consideration. 50 Research on World Agricultural Economy | Volume 03 | Issue 03 | September 2022 Data Availability The data used for this study can be made available upon request provided there is going to be compliance with the owners’ policy concerning sharing. References [1] Gardi, M.W., Memic, E., Zewudu, E., et al., 2022. Simulating the effect of climate change on barley yield in Ethiopia with the DSSAT-CERES-Barley model. Agronomy Journal. 114, 1128-1145. [2] IPCC (Intergovernmental Panel on Climate Change), 2014. Climate change report adaption, impact and vulnerability regional aspects synthesis report. Pachauri R.K, & L.A Meyer (eds), Contribution of working groups I, II and III to ‘Fifth Assessment re- port of intergovernmental panel on climate change’, Switzerland: IPCC Geneva. [3] Martinez, R., Hemming, D., Malone, L., et al., 2012; Improving climate risk management at local level: Techniques, case studies, good practices and guidelines for World Meteorological Organization members. N. Banaitiene (Ed.), Risk management – Current issues and challenges, September 2012. Inte- chOpen. DOI: https://doi.org/10.5772/51554# [4] Winthrop, M., Kajumba, T., McIvor, S., 2018. Ethio- pia country climate risk assessment report, Irish Aid, available at: https://www.climatelearningplatform. org/sites/default/files/resources/ethiopia_country_ climate_risk_assessment_report. [5] Yawso, D.O., Adu, M.O., Armah, F.A., 2020. Impacts of climate change and mitigation policies on malt barley supplies and associated virtual water flows in the UK. Scientific Reports. www.nature.com. [6] CSA (Central Statistical Agency), 2012. Annual Sta- tistics Bulletin. Addis Ababa, Ethiopia. [7] CSA (Central Statistical Agency), 2020. Report on area and production of major crops in Ethiopia. Agricultural Sample Survey 2019/2020, Addis Ababa, Ethiopia. [8] Lakew, B., Semeane, Y., Alemayehu, F., et al., 1997. Exploiting the diversity of barley landraces in Ethi- opia. Ethiopian Institute of Agricultural Research, Holettta Research Centre, Addis Ababa, Ethiopia. [9] Bekele, B., Alemayehu, F., Lakew, B., et al., 2005. Food Barley in Ethiopia. Grando, S. and H. G. Macpher- son (Eds.), Food Barley: Importance, Uses and Local Knowledge. ICARDA, Aleppo, Syria. pp. 53-82. [10] Asfaw, Z., 2000. The barleys of Ethiopia. S. B. Brush (edit), 2000; GENES in the FIELD: On-Farm Con- servation of Crop Diversity, International Develop- ment Research Centre and International Plant Genet- ic Resources Institute, Lewis Publishers, New York. [11] Wosene, G.A., Berhane, L., Bettina, I.G.H., et al., 2015. Ethiopian barley landraces show higher yield stability and comparable yield to improved varieties in multi-environment field trials. Academic Journals. 7(8), 275-291. [12] Samson, J., Berteaux, D., McGill, B.J., et al., 2011. Geographic disparities and moral hazards in the pre- dicted impacts of climate change on human popula- tions. Globa Ecology and Biogeography. DOI: https://doi.org/10.1111/j.1466-8238.2010.00632.x [13] Bekele, B., Wu, W., Yirsaw, E., et al., 2019. Climate change and its effect on land use change in the Cen- tral Rift Valley of Ethiopia. Applied Ecology and Environmental Research. 17(4), 7693-7713. [14] World Bank, 2021. Overview: Retrieved from: https://www.worldbank.org/en/country/ethiopia/ overview. [15] United Nations Population Funds, 2021. World pop- ulation dashboard Ethiopia. Retrieved from: https:// www.unfpa.org/data/world-population/ET. [16] CSA, 2018. Report on area and production of ma- jor crops in Ethiopia. Agricultural sample survey 2017/18, Addis Ababa. [17] Gashaw, G., Tura, K., 2015. Review of Barley Value Chain Management in Ethiopia. Journal of Biology, Agriculture and Healthcare. 5(10), 84-98. [18] Muluken, B., Jemal, E., 2011. Achievements in food barley breeding research in the early production systems of northwest, Ethiopia. Mulatu, B. and Grando, S. (eds), 2011; Review of Barley Research and Development in Ethiopia, Proceedings of the 2nd National Barley Research and Development Review Workshop, 28-30 November 2006, HARC, Holetta, Ethiopia. [19] Cammarano, D., Hawes, C., Squire, G., et al., 2019. Rainfall and temperature impacts on barley (Hordeum vulgare L.) yield and malting quality in Scotland. Field Crop Research. 241. DOI: https://doi.org/10.1016/j.fcr.2019.107559 [20] IPCC (Intergovernmental Panel on Climate Change), 2001. Climate change: The scientific basis, http:// www.ipcc.ch/. [21] Deressa, T.T., 2007. Measuring The Economic Im- pact of Climate Change on Ethiopian Agriculture: Ricardian Approach; Policy Research Working Paper 4342, EDRI, http://econ.worldbank.org. [22] Zaremba, Ł., 2018. Cobweb theorem in relation to the fruit market. AgEcon Search. 20(3), 190-195. http://ageconsearch.umn.edu. DOI: https://doi.org/10.5604/01.3001.0012.1516 https://doi.org/10.5772/51554 https://www.climatelearningplatform.org/sites/default/files/resources/ethiopia_country_climate_risk_assessment_report https://www.climatelearningplatform.org/sites/default/files/resources/ethiopia_country_climate_risk_assessment_report https://www.climatelearningplatform.org/sites/default/files/resources/ethiopia_country_climate_risk_assessment_report http://www.nature.com https://www.worldbank.org/en/country/ethiopia/overview https://www.worldbank.org/en/country/ethiopia/overview https://www.unfpa.org/data/world-population/ET https://www.unfpa.org/data/world-population/ET https://doi.org/10.1016/j.fcr.2019.107559 http://www.ipcc.ch/ http://www.ipcc.ch/ http://econ.worldbank.org http://ageconsearch.umn.edu 51 Research on World Agricultural Economy | Volume 03 | Issue 03 | September 2022 [23] Brianzoni, S., Mammana, C., Michetti, E., et al., 2008. A Stochastic Cobweb Dynamical Model. Dis- crete Dynamics in Nature and Society. Article ID 219653, pp 1-19. DOI: https://doi.org/10.1155/2008/219653 [24] Mendelsohn, R., Tiwari, D., 2000. Two essays on cli- mate change and agriculture: A developing country perspective, FAO Economic and Social Development Paper 145. Rome, Italy. [25] Pesaran, M.H., Shin, Y., Smith, R.J., 2001. Bounds testing approaches to the analysis of level relation- ships. Journal of Applied Economics. 16(3), 289-326. [26] Johansen, S., Juselius, K., 1990. Maximum likeli- hood estimation and inference on cointegration – with application to the demand for money. Oxford Bulletin of Economics. 52, 169-210. [27] Hassler, U., Wolters, J., 2006. Autoregressive distrib- uted lag models and cointegration; Working Paper, No. 2005/22, Free University of Berlin, School of Business and Economics, Berlin. [28] Duasa, J., 2010. “Real Exchange Rate and Trade Variables”: Asymmetric Cointegration Relationship. International Journal of Applied Business and Eco- nomic Research (IJABER). 8(1), 1-14. [29] Frimpong, M.J., Oteng, E.F., 2006. Bound Testing Approach: An Examination of Foreign Direct Invest- ment, Trade and Growth Relationships, MPRA Paper No. 352, pp. 1–19. [30] Dickey, D.A., Fuller, W.A., 1979. Distribution of the estimators for autoregressive time series with a unit root. Journal of the American Statistical Association. 74(366a), 427-431. [31] Gujarati, D., 2004. Basic econometrics. New York: McGraw Hill. [32] Heij, C., de Boer, P., Franses, P.H., et al., 2004. Econometric methods with applications in business and economics. New York: Oxford University Press. [33] Wooldridge, J.M., 2013. Introductory econometrics. A modern approach (pp. 868): South Western Cen- gage Learning. [34] Enders, W., 2010. Applied econometric time series (3rd ed.). Hoboken, New Jersey: John Wiley & Sons, Inc. Publishing. [35] Akter, N., Hong, S.J., 2011. Education and economic growth in Haryana (India): Application of cointe- gration analysis and vector error correction model on supply response of pulses in Bangladesh. South Asian Studies. 17(1), 277-295. [36] Ssekuma, R., 2011. A study of cointegration models with applications; Unpublished Master of Commerce Dissertation, University of South Africa. [37] Sharma, S., Singh, S., 2019. The Validity of Wag- ner’s Law in India: A Postliberalisation Analysis, M.J.P. Rohilkhand University, Uttar Pradesh, India. [38] IPCC, 2013. Summary for Policy Makers; Climate Change 2013: The Physical Science Basis, Contribution of the Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)], Cambridge University Press, Cam- bridge, UK and New York, NY, USA. [39] Conway, D., 2000. Some aspects of climate variabil- ity in the North East Ethiopian Highlands, Wolli and Tigray. Ethiopian Journal of Science. 23(2), 139-161. [40] Fischer, G., Velthuizen, H.T.V., 1996. Climate Change and Global Agricultural Potential Project: A case study of Kenya, Laxenburg: International Insti- tute for Applied Systems Analysis. [41] Dumrul, Y., Kilicarslan, Z., 2017. Economic impacts of climate change on agriculture: empirical evidence from ARDL approach for Turkey. Journal of Busi- ness, Economics and Finance (JBEF). 6(4), 336-347. [42] Lobell, D.B., Schlenker, W., Costa-Roberts, J., 2011. Climate trends and global crop production since 1980. Science. 333, 616-620. [43] Schlenker, W., Roberts, M., 2009. Nonlinear tem- perature effects indicate severe damages to U.S. crops yields under climate change. Proceeding of the National Academy of Science. 106, 15594-15598. [44] Chandio, A.A., Gokmenoglu, K.K., Ahmad, F., 2021. Addressing the long- and short-run effects of climate change on major food crops production in Turkey. Environmental Science and Pollution Research, Springer. pp. 1-17. [45] Ketema, A.M., 2020. Determinants of agricultural output in Ethiopia: ARDL approach to co-integration. International Journal of Business and Social Re- search. 10(3), 01-10. [46] Taye, B.A., Asfaw, F.F., Yirsaw, B.G., et al., 2021. Modeling the Impact of Climate and Fertilizer on Barley Production. American Journal of Biological and Environmental Statistics. 7(2), 44-51. [47] Chandio, A.A., Rehman, A., Jiang, Y., et al., 2020. Short and long-run impacts of climate change on agriculture: an empirical evidence from China. In- ternational Journal of Climate Change Strategies and Management. 12(2), 201-221. [48] Elbeydi, K., Eljadi, A.A., 2007. Measuring the Sup- ply Response Function of Barley in Libya. African Crop Science Conference Proceedings. 8, 1277-1280.