Addressing Structural Transformation: Exploring the potential exports using product space

Abstraksi

The Indonesian structural transformation has been recorded with the stagnancy in decades, where a higher growth on service sectors at the expense of lower industry share occurs, signaling what economists coined as premature de-industrialisation (Rodrik, 2015). However, historically, to be an advanced and upper-middle income country, manufacturing sectors should keep its pace of growth, even though service (that is driven by specialization of works nowadays) can be growing fast at the same time. In the case of China, Malaysia, and Korea, the countries once hit the 30% employment of manufacturing sectors, before naturally its decreasing because of further advancement of high-productivity service sectors, while Indonesia’s rate is staggering in 18-22% level in last decade (WDI, 2019). It is notable that two Dutch Diseases happened: in late 1970s (oil boom due to Gulf War) and 2004-2011 (commodities boom fueled by Chinese demand) as well as Asian Financial Crisis in 1997 as remarkable events that slow the structural transformation in Indonesia (Pangestu et al., 2015). The problem of structural transformation is closely related with the facilitation of transition to manufacturing sectors. McMillan and Rodrik (2011) emphasis on two components on the productivity ‘within’, i.e. the improvement of productivity within the sector; and ‘between’, i.e. movement of labour to the sectors with higher productivity. Anandhika and Laksono (2014) exercised this calculation for Indonesia in the year 2001-2011 and found that manufacturing sectors even have a negative ‘productivity between’ growth for -0.58% per annum, indicating subtle ‘deindustrialisation’ happened. Some related factors are identified for this sluggishness and low-productivity structural transformation. First, factor related with the ‘within’ productivity come from the lacking of labor productivity improvement. Both TFP and labor productivity slowly increased but relatively slower than other countries in region (Indonesia’s manufacturing productivity growth is amongst the lowest in ASEAN in 2005-2016, according APO Database in 2019), resulting the low competitiveness of Indonesian labour. This human capital issues could be related to education, health (including child nutrition), and currently is being address by the government whose the results can be harvested in decades ahead. The second factor is from the facilitation of shifting of labour and industry from lower to higher productivity sectors. In labour context, this is related to flexibility of labour movement. Currently, Indonesia ranks 82 in Labour Market Pillar in 2018 Global Competitiveness Index, while our regional competitors such as Philippines, Malaysia, and Thailand ranked in 36, 20 and 44, respectively. However, it is a relatively untouchable topic about the industrial shifting from low value-added industry into higher value-added, probably because of lacking of science-based methodology to estimate (the approach was usually based on downstreaming arguments or industrial clusters, which could not capture the similarity of skills). This research is expected to help government (or planning institute such as Bappenas) to at least understand the existing condition of the competitiveness of sectoral industry and the potential sectors to consider the relevant policies. Research Questions 1. What sectors are having a competitive skill set in Indonesia and what sectors can be extended from current competitive sectors? 2. What competitive sectors are having similar skills with the existing less competitive ones so the industry can diversify to the higher value-added products similar with their current skills? 3. What are the relevant policy recommended for the results of product space mapping? Methodology To map the proximity of skills in Indonesia, we use the ‘Product Space’ indicator developed by Hidalgo et al (2007). The logic is derived from the amount of probability if one product is exported given the other product’s exported. This approach assumes the higher probability of two goods being exported together means the more similar skillset needed to produce the goods. The probability is captured by utilizing Revealed Comparative Advantage (RCA) indicator (Balassa, 1965). In mathematical equation: φ_(i,j,t)=min⁡{P(x_(i,t)│x_(j,t) ),P(x_(j,t)│x_(i,t) )} x_(i,c,t)={■( 1 if 〖 RCA〗_(i,c,t)>1@0 if otherwise)┤ █( @RCA)_(i,c,t)=(〖xval〗_(c,i,t)⁄(∑_i▒〖xval〗_(c,i,t) ))/((∑_c▒〖xval〗_(c,i,t) )⁄(∑_i▒∑_c▒〖xval〗_(c,i,t) )) Where φ_(i,j,t) means the minimum conditional probability between exporting goods i given the exporting of goods j in year t, and vice versa. x_(i,c,t) is 1 if the RCA of product i from country c in year t is higher than 1, while 0 for otherwise. We continue the calculation using the density indicator, i.e. pairwise proximity measures for each element of the country’s entire export basket as follow: 〖density〗_(i,c,t)=((∑_k▒〖φ_(i,k,t) x_(c,k,t) 〗)/(∑_k▒φ_(i,k,t) )) Density has been statistically proven in determining if the skills of producing i can be adopted to the new product (Hausman & Klinger, 2006). The research employs the international trade data from UNCOMTRADE in building RCA database, in 6-digit HS level or 5-digit SITC level. The research will take the latest year of international trade data and 5 years before to see the evolution of RCA and product space. Some comparison study are planned to be exercised with the countries in the region to see the difference of transformation patterns: China, Malaysia, and Vietnam. Policy Implication Broadly speaking, the industrial policymaking is usually driven by downstreaming obsession—which losing its relevance now when trade openness give more access of importing cheaper consumer goods, while the advantage of cheap raw materials often ignoring other significant proportion for production costs such as energy and intermediate input; or industrial cluster—which still have limitation in explaining sectoral development beyond the output product. Mapping the proximity from the product space allows the policymaker to use skills/knowhow-centred perspective in understanding the current condition of the competitiveness and skills similarity between industries, as similarly exercised by Poncet and Waldemar (2013) for China, Hamwey et al (2013) for Brazil and Lo et al (2015) for Turkey. The planner body such as Bappenas can utilize the analysis to direct the sectoral vision on the National Development Plan (e.g. RPJMN) with skills orientation, instead of downstreaming of local commodities. Combining the analysis with the firms database, government will have a more targeted programme to develop the enterprises for product diversification and inter-sectoral transition of the labour force.