INTRODUCTION
The agriculture sector is one of the major industry of the country’s economy and the government of Myanmar puts a high priority on it. The potential of agricultural development offers the most significant opportunity for inclusive economic growth in Myanmar. The vision of the agricultural policy in Myanmar is achieving sustainable development, international standard, food security, and climate smart agricultural system ensuring wellbeing of rural livelihood and the socioeconomy development of smallscale farmers, and the country’s economic growth by 2030 (MoALI, 2018).
The economy of Myanmar is primarily based on three sectors that include agriculture, industry, and services. The agriculture sector includes crops, livestock, and fishery subsectors which contribute 21.9%, 26.6% to total export earnings in 2019 and 2020, respectively, and the sector employs 67.3% of the labor force (MoALI, 2020). The agriculture sector plays a critical role in allinclusive development, rural poverty eradication, and food security enhancement for the country. It is also important to promote the employment and the livelihood of the rural population, especially for some 36% of those living below the poverty line (ADB, 2013).
Myanmar's government is dedicated to achieving the Sustainable Development Goals (SDGs), with the objective of delivering a balance of development in the economic, social, and environmental spheres. For these reasons, sustainability in every area should be considered such as crosscutting, and integrated into all aspects of Myanmar’s Sustainable Development Plan (MSDP) implementation. Myanmar's fulfillment of the SDGs will be guided by the 2030 Sustainable Development Agenda throughout the same time frame (MoPF, 2018).
Myanmar receives international development assistance with largest amount nowadays (Carr, 2018). Many development agencies give a priority, especially, to social infrastructure and services. In recent years, Myanmar becomes the 3^{rd} largest recipient country of Official Development Assistance (ODA) per capita followed by Cambodia and Lao PDR, although it was the 7^{th} largest recipient country back in 2015. Therefore, it is necessary to analyze the effectiveness of the ODA for Myanmar and to find out better ways to enhance sustainability for the future cooperation.
There have been many studies to consider the effectiveness of ODA on agricultural development. Most of the studies analyzed the contribution and the impact of agricultural ODA for Myanmar and other countries utilizing various methodologies such as Autoregressive Distributed Lag (ARDL) method, Vector Error Correction Model (VECM), and Generalized Method of Moments (GMM). These studies found out that the agricultural ODA had a positive and significant contribution, in particular, to value chain and technical transformation. Moreover, Korea’s ODA to agricultural production and marketing areas did not show significant effect but had a positive relationship with the regional economy (Htay and Lee, 2021). Nonetheless, Alabi (2014) and Ssozi et al. (2019) argued that bilateral aid in agriculture was more effective in productivity than multilateral aid. Zewide (2017) investigated on agricultural growth for Ethiopia using VECM and revealed that ODA had a negative and significant effect on agricultural growth in the short run, but it had a significant and positive influence in the longrun. In particular, Ighodaro & Chii analyzed the effectiveness of ODA on the agricultural sector growth in Nigeria applying ARDL and VECM and pointed out that ODA had an insignificant and negative relationship on the agricultural output both in the longrun and shortrun. Oppositely, both technological trend and savings had a significant and positive relationship with agricultural output in terms of the shortrun and the longrun (Ighodaro & Chii, 2015). The results of the previous researches showed there has been a positive relationship between agricultural growth and ODA. However, few studies investigated the effectiveness of ODA in the agricultural sector.
Korea’s Bilateral ODA Allocation
The Korean government has tried to establish its ODA implementation system more effective, transparent, and sustainable. Its implementing principle of ODA has ensured that ODA should contribute to the global community’s efforts to achieve SDGs. The significance of supporting ODA is aligned with the government’s international policies, such as diplomatic diversification and strategic economic cooperation, and a greater emphasis is being placed on the need for ODA to contribute to national interests of the partner countries (MoFA, 2021).
In 2010, the Committee on International Development Cooperation (CIDC) prioritized 26 partner countries for a targeted and concentrated approach to improve development effectiveness. Korea's bilateral aid to 24 priority partner countries continued to be selected in 2019, with a total amount of USD 1,060.37 million for a fiveyear period from 2015 to 2019 (55.7 percent). Bangladesh and Lao PDR, two of the top ten recipients of ODA, received a total of USD 78.81 million. Bangladesh was the largest recipient country in 2019, with the amount of USD 98.38 million, followed by Vietnam ($93.21 million), Myanmar ($83.39 million), and Ethiopia ($82.15 million), as shown in Table 1 (OPC, 2020).
Korea’s bilateral ODA was primarily focused on Asian countries in 2019, with USD 1.0 billion. Africa was received USD 516.3 million was contributed to African countries, accounting for 25.2% of the bilateral ODA. Moreover, Asia and Africa were also the major regional recipients of Korea’s multilateral aid, based on its policy priorities of overall strategy. However, gross bilateral ODA was not broken down by region in 2019. Furthermore, Korea also committed USD 1.4 billion (40.4% for bilateral aid) to promote assistance for commerce and trade performance of developing countries and combine into the world economy (OECD, 2021).
International Development Cooperation of Myanmar
Distribution and allocation policies of resources has been considered as a key component for poverty reduction and equitable economic growth, and Myanmar also has a similar policy that emphasizes on the role of greater access to public goods and services to reduce poverty. Development partners should also ensure that their technical advice is explicitly provided to assist the government policy that efficiently utilizes resource allocation for the smallscale agriculture.
In 2016, the total amount of ODA from international organizations to Myanmar was approximately $1.5 billion, of which ODA for the agriculture and forestry sector was only 5.9%. The amount of ODA was $235 million from Japan, $183 million from U.S., $124 million from Korea, $72 million from Germany, and $48 million from France. The international organizations were the World Bank with $300 million and Asian Development Bank with $263 million for various development programs and projects.
In general, Myanmar receives bilateral and multilateral development ODAs. These assistances focus on various areas. For instance, Japan provides technical cooperation with private partners, the U.S. provides ODA programs for agricultural production of smallholders to enhance productivity and marketing capability in response to climate change. Furthermore, ACIAR assists for agricultural research, ADB focuses on promoting agricultural value chains and technical assistance for rural development, the World Bank Group supports financial access for rural development, and FAO focuses on promoting agricultural value chain development through mitigating natural disasters coping with climate change (Moore, 2016).
In the meantime, Korea designated Myanmar as one of the priority partner countries since 2010. Korea is also one of the major donors to Myanmar with various ODA projects. In addition, there are a number of assistant organizations of Korea in Myanmar, such as KOICA (Korea International Cooperation Agency) and KOPIA (Korea Program on International Agriculture). KOICA focuses on new village development programs, postharvest research and agricultural marketing. KOPIA provides technical assistance for productivity enhancement in the agricultural sector. Regarding rural development of Myanmar, the government has benchmarked Korea’s experience of ‘Saemaul Undong, SMU’, which the Korean government launched a comprehensive rural development policy in the early 1970s, with a focus on rural livelihood and agricultural infrastructure such as village development, irrigation, farmland expansion, and arable land development. SMU entails the experience and lessons learned from enhancing agricultural productivity, building infrastructure, establishing systems of processing and distribution of agricultural products, and developing advanced agriculture technologies. Moreover, the Korean government would continue to share its development experience and knowledge with Myanmar to support the socioeconomic development strategies of the nation.
Regarding Korea’s Country Partnership Strategy with Myanmar, Korean government is willing to promote sustainable and inclusive rural development, agriculture infrastructure, and value chain development by providing necessary ODA programs. The strategy also highlights enhancing agricultural productivity and promoting exports of Myanmar’s agricultural products. The details of the strategy focus on the followings: 1) promotion of efficient irrigation management for predictable and systematic agricultural production, flood prevention, and water management; 2) policy consultation and technical assistance for agricultural product distribution and exports; 3) formation of independent farmers’ cooperatives to improve their bargaining power in the markets; 4) rural community development while reflecting the lessons learned from SMU; and 5) establishment of accessibility to agricultural financial services (CPS, 2021).
Therefore, it is necessary to examine the effectiveness of ODA, in this case, between Myanmar and Korea to make sure that the future development cooperation take a further step toward the right direction for both countries and possibly for others driven by implications of the study. This study aims to build a winwin situation for both countries and to promote an effective way for development cooperation to contribute to SDGs. The main objective of the study is to analyze the effectiveness of bilateral development cooperation in the agricultural sector between Myanmar and Korea.
Analytical Framework and Model Specification
To achieve the main objective of the study, this study utilizes VECM to examine the existence of cointegration among the relevant factors investigating the effectiveness of bilateral cooperation in the agriculture sector. The methodological approach of the study is depicted in the Figure 1. The reason for using the model is that it is relevant due to its closeness to reality and its ability to produce better shortterm forecasts and, more importantly, longterm effect can be aggregated in economically meaningful respects for macroeconomic perspectives.
Data Source
Regarding the general objective, this study utilizes the secondary data. The data used is annual time series from 1991 to 2018 (28 years) from different reliable sources such as Korea International Cooperation Agency (KOICA) and Central Statistical Organization (CSO) of Myanmar for different variables for ODA and the agricultural sector in Myanmar.
Time series analysis is explored in this study. To analyze the effectiveness of bilateral development cooperation of Korea on the agricultural development of Myanmar, the interest variables used in this study are various. For example, the influential factors on agricultural development, agriculture GDP(AGDP) can be explained by agricultural official development assistance (AODA), government expenditure on agriculture sector (GEA) and foreign direct investment (FDI).
Stationary Test
Doing empirical analysis with time series data, it is essential to determine whether data is stationary or not. The data with unit roots cause problems of statistical inference and unpredictable results might occur in time series. There are several ways to test a unit root. Among these, this study uses ADF test. It is more powerful than the DickeyFuller (DF) test and can handle more complex models. The ADF test determines that y_{t} is I (1) when compared to the alternative hypothesis of I (0), it can be assumed y_{t} is an Autoregressive Integrated Moving Average (ARMA) process and it has not only autoregressive but also moving average terms (Mccarthy & Permanente, 2015).
Using the ADF test for testing the existence of a unit root, the estimation of regression is required as follows:
Where

y_{t} = series that should be tested

β_{0} = a constant term

β_{1} = test coefficient

t = trend or time

n = lag length chosen for ADFtest

u_{t} = white noise error term

Ho: β_{1} = 0 and ${\sum}_{i=1}^{n}{\alpha}_{i}\Delta {y}_{ti}=1$, nonstationary, the time series with unit root

H1: β_{1} < 0, stationary, the time series with no unit root
If the absolute value of computed t statistics is larger than the critical values of ADF, we reject the null hypothesis, β_{1} = 0, in which there is no unit root and it is stationary and vice versa.
CoIntegration Test
As the second stage in the VECM model, determining the existence of cointegration among the time series variables is critical. For assessing longrun relationship in the series, the Johansen’s (1988) initial multivariate procedure is considered for the study. As a part of the procedure, a maximum likelihood procedure is used to determine the existence of cointegrating vectors in nonstationary time series dataset. We can accept the null hypothesis if there is no cointegration between the variables. Otherwise, if the series have at least one cointegration, we can accept the alternative.
If the interest variables are cointegrated, this suggests the presence of a longterm or equilibrium relationship. There is a longrun relationship in the system if there is cointegration among the variables in nonstationary variables. Since differencing the variables to become stationary in the regression does not reflect longrun behavior of the variables, checking for the presence of cointegration is crucial. To assess whether or not nonstationary variables in a particular model have a longrun equilibrium relationship, empirical analysis for evaluating cointegration is required. In order to conduct cointegration test, Johanson maximum likelihood estimation procedure is applied.
In the VECM modeling, the first step in Johansen cointegration analysis defines an unrestricted vector autoregression (VAR). It is critical to choose the correct VAR order for a cointegration test. The following form of VAR is considered:
where Y_{t} is vector of I(1) variables, A_{t} is matrix of parameters, X_{t} is vector of stationary I(0) exogenous variables, A_{X} is matrix of parameters and U_{t} is white noise errors.
Vector Error Correction Model (VECM)
If the variables are cointegrated or have a long run relationship, an error correction mechanism can be proceeded. The reason for using VECM is that it is relevant due to its closeness to reality and its ability to produce better shortterm projection and, more importantly, longterm forecasts aggregated in economically meaningful respects for macroeconomic policy analysis. In addition, the error correction mechanism (ECM) is used to accurate the variables of any deviation in shortrun from the condition of longrun equilibrium. If there is the existence of longterm or equilibrium relationship between the two variables of Y and X, ECM can be used. The ECM can be expressed as follows:
Where Δ indicates the first difference operator, εt is a random error term, and U_{t−1} = Y_{t}_{− 1} − β_{1} − β_{2}X_{t}_{−1}, the oneperiod lagged value of the error term.
The above equation indicates that ΔY_{t} depends on ΔX_{t} and the equilibrium error term. The model is out of equilibrium condition when the error term is nonzero. Therefore, the study assume that ΔX_{t} is zero and U_{t}_{−1} is positive. Since b_{2} is expected to be negative, the term b_{2}U_{t}_{− 1} is negative and ΔY_{t} would be negative to restore the equilibrium.
It is necessary to check unbiased estimation and robustness of the model by conducting diagnostics tests. The robustness of estimated coefficients can be determined by conducting Diagnostic tests. Lagrange multiplier (LM) test, residual diagnostics is utilized in this study for testing residual. Then the diagnostics test for stability also applied to examine whether the parameters are stable or not in the estimated model across various data used.
Empirical Results
Testing the presence of unit root in time series estimation is necessary before running the regression. For testing stationarity, ADF unit root test is applied. The test finds that there are no variables that are stationary. Therefore, first difference of the variables is used to examine the presence of stationarity. Based on the ADF test statistics at first difference, the absolute values of all variables are statistically significant at the 5% level. As a result, the variables are stationary at first difference and we can accept alternative hypothesis. It means that the data and variables can be used for the analysis. The empirical test for ADF does not include the constant term for the simplicity due to the equivalent results.
In time series analysis, the error correction model is crucial due to better understand longrun dynamics. Since VECM is the multivariate extension of the ECM, the existence of cointegration is essential for the longrun relationships in time series data. Therefore, we should choose the appropriate lag length to consistently test for cointegration. The three major tests that can be applied to choose appropriate lag length are Akaike Information Criteria (AIC), the HannanQuinn Information Criteria (HQIC) and the Schawarz Information Criteria (SBIC). For determining the optimal lag length for the study, AIC is used in this study and the result indicates that the optimal lag length is lag order 2 as shown in the Table 3.
Cointegration tests identify situations in which two or more nonstationary time series variables are integrated so as not to deviate from the longrun equilibrium. To identify the longterm relationships between sets of variables, this study used Johansen test, it allows for more than one cointegrating relationship. Based on the results of the Table 4, the values of trace statistic are larger than 5% critical value at 0 cointegration equation, we can accept the alternative hypothesis. In this model, there is one cointegrating equation and this indicates that there is cointegration among the variables and they have long run relationship, so we can apply VECM for the further analysis based on the Johansen’s test result of maximum rank of 1 with the eigenvalue of 0.69.
The model equation is as follows:
When ECT_{t}_{− 1} = 0, the normalized co integration equation can be written as:
In Table 5, the results indicate that the coefficient of Korea’s ODA for the agriculture sector (lnAODA) is negative (1.775215) and government expenditure on the agriculture sector (lnGEA) is also negative (1.186621) and statistically significant at 1% level, respectively. However, the coefficient of foreign direct investment (lnFDI) is positive (1.965333) and it is also statistically significant at 1% level.
In an empirical sense, there is elasticity among variables; a percentage change in AODA and GEA results in a 1.78% and 1.19% increase in AGDP, respectively. But a 1% change in FDI causes a 1.97% decrease in AGDP. The output of the longrun relationship of the vector errorcorrection model states that independent variables AODA and GEA have a positive impact and are statistically significant on AGDP. On the other hand, FDI is statistically significant but has a negative impact on AGDP.
As it is expected, the outcomes confirm that the effect of Korea’s ODA for agricultural growth and government expenditure on the agriculture sector have a positive influence on the agriculture sector. Nevertheless, FDI has a negative relationship with agricultural sector growth. The coefficients of all variables are statistically significant at the 1% level.
The results in Table 6 indicate that there is a longrun causal effect in AGDP at a statistically significant level of 5%. The coefficient is .186 and the pvalue is 0.025. That shows the presence of a longrun causality effect in the AGDP equation. In the short run, the coefficient of FDI is positive (.311) and it has a causality effect on AGDP at a 5% level of significance. But the coefficient of AODA is negative (.245) and GEA is positive (.003) and they do not have a causal effect on AGDP in the short run as they are not statistically significant. It refers that Korea’s ODA for the agriculture sector has a negative effect and is not statistically significant on the agricultural GDP, and government expenditure on the agriculture sector has a positive impact but is not significant in the short run. Nevertheless, foreign direct investment has a positive impact and is statistically significant on the agricultural GDP in the short run.
Autocorrelation problem can be occurred arises in a regression model when error terms correlate over time. Lagrange multiplier (LM) test is used to determine the existence of autocorrelation problem in Table 7 and the outcome indicate that there is no serial correlation in both lag order one and two as pvalue is larger than 10% critical value and we can accept the nullhypothesis.
The JarqueBera test is usually use for testing normality. It is always positive and if not close to zero, it shows that the data used do not have a normal distribution. As shown in Table 8, the results show that in D_lnAGDP and D_ln GEA equations, the residuals are less than 5% critical level and not normally distributed. Only D_lnAODA and D_lnFDI, the residuals are greater than 5% critical level, which means that the residual are normally distributed. Therefore, we have to accept the alternative hypothesis and the result shows that the residuals are not normally distributed in this scenario.
The VECM specification imposes 3unit moduli. The model’s stability demonstrates the estimated model's validity; thus, it should be tested before proceeding. The longterm stability of the parameters was also checked by plotting recursively limited to 95% of the critical values. As seen in the Table 9, the model satisfies with the stability condition.
CONCLUDING REMARKS
The agriculture sector is critical in boosting rural people's social welfare and overall economic development. It is also important in poverty reduction and ensuring inclusive development of a nation’s economy in the long term. Myanmar is an agriculturebased country and more than half of rural people rely on agriculture for their livelihoods and employment. In that sense, international development cooperation is essential in promoting economic transformation and in accelerating progress towards the SDGs.
This study explores the effectiveness of Korea’s bilateral development cooperation on the agriculture sector of Myanmar using a cointegrated VECM to show both the longrun and shortrun relations. The significant coefficients show that both the explanatory variables of Korea’s ODA and Myanmar’s expenditure on the agriculture sector have a positive effect on the sector's growth in the longrun. In addition, foreign direct investment has a significant impact, but its longterm impact on agriculture is negative. In the meantime, only foreign direct investment has a significant and positive impact in the short run while Korea’s ODA and government expenditure on the sector have a negative effect and are not significant on the agricultural GDP in the short run.
Although Korea's development cooperation to the Myanmar's agricultural sector has been effective, it still needs to be improved efficiently. On the other hand, the government of Myanmar needs to be more transparent, timely, and manageable in order to implement a more efficient and sustainable way. Even though the Myanmar's agricultural sector has great potential to be developed, there are still weaknesses on policies and a lack of access to foreign aid compared to other ASEAN countries due to a lack of stability in investment legislation. If they formulate better agricultural policies, laws, and legislation, there would be more support from the international community including Korea.
적 요

본 연구는 한국이 미얀마 농업 분야에 지원한 ODA 효과성을 1991년부터 2018년까지의 시계열 자료를 활 용하여 분석함.

벡터오류수정모델(VECM) 분석 결과, 미얀마 농업 분야 에 대한 한국의 양자 원조(ODA)와 농업 분야에 대한 미 얀마 정부지출이 농업GDP에 긍정적인 영향이 있는 것으 로 분석되었으며, FDI는 단기적으로 긍정적 영향을 미치 는 것으로 나타났지만 장기적 관점의 분석 결과는 부정적 인 영향을 미치는 것으로 분석되어 ODA활용을 통한 중 장기적 농업개발 정책의 필요성이 대두됨.

본 연구는 미얀마 정부가 보다 효율적이고 지속가능한 농 업개발을 위해 투명하고 효율적인 정책 수립의 필요성을 시사하고 있으며, 이는 각 부문에서 ODA효과성을 강화할 필요성이 강조된 것임.

미얀마의 농업 분야는 발전 가능성이 크지만 제도 등 관 련 정책의 시행이 불안정하기 때문에 개발협력에 대한 접 근성이 열악하여 향후 미얀마 정부는 효과적인 농업 정책 과 규정 수립 등을 통해 국제사회와의 협력을 확대할 필 요가 있음.