Types and Sources of Data

19 export volumes going to the United States, Japan, China, Singapore, the Republic of Korea, Germany, Canada, Brazil, India, and Belgium. The data used were obtained from Statistics Indonesia BPS, the International Rubber Study Group IRSG, the United States Department of Agriculture USDA, the Centre dEtudes et dInformations Internationales prospectives CEPII, UN Comtrade, the World Bank, and FAOSTAT. Types and sources of data for the study are summarized in Table 2. Table 2 Type and Source of Data No Type of Data Unit Source of Data 1 Natural Rubber Export Volume KG UN COMTRADE 2 Natural Rubber Production KG FAOSTAT 3 Real GDP US WDI 4 Distance Km CEPII timeanddate.com 5 Exchange Rate WDI 4.2 Analysis Method 4.2.1 Descriptive Analysis The descriptive analysis is an analytical method that is used to describe the general condition of a phenomenon. The purpose of descriptive analysis is to facilitate the interpretation or explanation in the estimation result of the data. In this study, a descriptive analysis is used to describe a general overview of the international trade of natural rubber specifically the development of the sector from upstream to downstream in the world trade system. Additionally, descriptive analysis is used to analyze the dynamics of international trade flows, as well as the competitiveness of the commodity as a means of support for the result of the quantitative analysis. Some of the techniques are used to compile the data in the form of tables or graphs, along with the interpretation of the data and presented arguments. Descriptive analysis and graphical tabulation are the simplest methods, yet still have a strong capability to qualitatively explain the relationship between the observed variables.

4.2.2 Revealed Comparative Advantage RCA

RCA is a method of analysis for determining the competitiveness of a player, while taking into account the comparative advantage of an individual, organization or country. The performance of natural rubber exports from exporting countries to importing countries is a variable that is measured by calculating the share of natural rubber exports to total exports of all commodities to destination countries, compared with the share of world exports to the destination countries. So that it can be quantitatively the ability or inability of three main exporting countries compete in international trade. The calculation of the RCA method is as follows: 20 ��� = � � � � where X ij is the export value of natural rubber commodity from country j, X it equals the value of total exports for country j, W j represents the world export value of natural rubber commodity and W t is the total value of world exports. The competitiveness index value of a commodity in the RCA has two possibilities, namely: 1. If RCA 1, it indicates that the share of natural rubber commodity in the total exports of country j is greater than the average share of natural rubber export in all countries the world. This means that country j has a comparative advantage, so it is relatively more specialized in natural rubber commodity. 2. If RCA 1, it shows that the share of natural rubber commodity in the total exports of country j is less than the average share of export natural rubber commodity in all countries the world. This means that country j does not have a comparative advantage, so the country will not specialize in natural rubber commodity.

4.2.3 The Least Square Dummy Variable LSDV

The Least Squares Dummy Variable LSDV method aims to represent differences in the intercept, i.e., through the use of dummy variables Firdaus, 2011. To illustrate this approach, for example in the initial equation as in the pooled least square PLS equation and the group of dummy variables d git = 1 g=i. � = � + � � + � By entering a number of d git = 1 g=i, the initial equation becomes: � = � � + � � + ⋯ + � � + � � + � This equation can be estimated with the OLS approach in order to obtain the parameter β LSDV. The advantage of the LSDV approach is that it can generate the estimated parameter, β, which is unbiased and efficient. The weakness of this approach, however, is that if the research contains a large number of observation units, it looks unmanageable. To test whether the intercept is significant or not, the F-test can be used with the following hypotheses: � = � = � = � = ⋯ = � H 1 = one value of α is not the same. The hypothesis can be directly used to test whether it is better to use the PLS or the LSDV approach. The basic rejection of H is to use the F-statistic that is: � = � − � 1 − � . �� − � − � � − 1