Income Inequality in Malaysia: Examining the Labour Income Links.

• Author: Wai, Christopher Choong Weng
• Position: RESEARCH NOTES

1. Introduction

   Tun Dr Mahathir Mohamad, in his second term as Prime Minister of Malaysia, launched the Shared Prosperity Vision 2030 (SPV) on 5 October 2019. The SPV is the new long-term plan to guide the development of Malaysia, with the expectation that the thirty-year Vision 2020, conceived in 1991 under his first term as Prime Minister, would be coming to an end. Vision 2020, introduced when the country was experiencing roaring economic growth, is largely perceived as falling short of its established ideals due to the economic mismanagement of the previous administration. The SPV, announced against a backdrop of moderating economic growth, was meant to craft a new narrative for Malaysia under the first-term Pakatan Harapan (PH) government. Despite the unexpectedly short-lived and truncated tenure of the PH government, controversially replaced by the Perikatan Nasional (PN) government in the first quarter of 2020, the SPV survived the demise of its original architect. The PN government has maintained a stance of policy continuity, embracing the SPV as a long-term strategic document alongside more immediate COVID-19 initiatives and ongoing development of the Twelfth Malaysia Plan (12MP).

   Wealth and income inequalities are prominently featured in the SPV, more so than previous long-term and medium-term plans, forming one of the three key pillars of the strategic plan. In formulating the issue of income inequality, which forms the basis for target setting and action planning, one of the key departures of the SPV from previous development plans--at least up until the Eleventh Malaysia Plan (11 MP)--is its shift towards factor income, beyond the conventional household income, to conceptualize income inequality and assess disparities.

   The SPV makes the case, justifiably, that the share of compensation of employees (CE) to GDP at 35.7 per cent in 2018 is low compared to developed countries. It then goes on to set a target of increasing the CE share to 48 per cent by 2030. While the 1 IMP has a similar target of increasing the CE share to 40 per cent by 2020 (revised downwards to 38 per cent during the 1 IMP Mid-Term Review (MTR)), it was introduced as a broader effort to improve labour productivity and wages through the shift to high-skilled jobs. However, the shift in focus to inequality between labour and capital owners, from inequality between households, articulated as a crude expression of the CE share without attendant considerations of its underlying dynamics, reflects the new emphasis of the PH government and continued by the PN administration. The shift is also alluded to in the 1 IMP MTR, but now framed more noticeably as an inequality issue in the SPV.

   At the same time, the SPV maintains a target of reducing the Gini coefficient to 0.340 by 2030 (from 0.399 in 2016). It is a practice, started in the Tenth Malaysia Plan (10MP), of setting Gini targets in reducing household income inequality. The twin goals of increasing the labour income share (LIS), using the CE share to GDP as a measure, and reducing household income inequality, using the Gini coefficient, are formulated as concrete outcomes of addressing inequalities, implicitly assumed to be complementary to each other and to the broader goal of moving the economy up the value chain. These targets extend pre-existing trends, but their relationships when subsumed under a broader inequality agenda must be given further scrutiny, as their underlying dynamics may contain trade-offs and other non-negligible intermediating factors.

   Against this backdrop, this study undertakes a critical examination of the relationship between income inequality and labour income in Malaysia. In the next section, drawing on relevant literature on Malaysia, we discuss the specific aspect of the LIS and its relationship with income inequality as well as consider the trade-offs. In the third section, we usher in another aspect of labour income into the discourse, i.e., the labour income dispersion (LID) and argue that it is potentially the more important variable in the link to income inequality. This is then analysed with a regression model in the fourth section. The concluding section discusses the results and policy implications.

2. Income Inequality and Labour Income Share

   Interest in the links between income inequality and labour income stems from a wider global resurgence in the interest to research the links between personal income distribution and functional income distribution (Francese and Mulas-Granados 2015). Personal income distribution is usually measured with data on income obtained from household surveys, most commonly with the Gini coefficient, while functional income distribution draws from factor income obtained from national accounts such as labour income and capital income. This study uses income inequality to refer to personal income distribution and labour income to functional income distribution.

   The growing interest to understand these two sets of distributional outcomes is motivated by a desire to find macroeconomic explanations for personal income distribution (Atkinson 2009), and can be traced back to the works of Kuznets in the 1940s and 1950s in which he analysed the distributional aspects of national accounts of some industrialized countries (Piketty, Saez, and Zucman 2017). In more recent literature, Karabarbounis and Neiman (2014), building on the growing literature that study the global decline of the LIS (Ellis and Smith 2007; Azmat, Manning, and Van Reenen 2012; Karabarbounis and Neiman 2013; Piketty and Zucman 2014), raised the question of whether gross or net labour share was more informative in studying welfare-based measure of inequality, and concluded that neither was adequate outside of steady state. A number of approaches have been introduced and used to study the links, which include undertaking standard decomposition and regression analysis as well as constructing and analysing social accounting matrix (SAM) and distributional national accounts (DINA) (Civardi and Lenti 2018).

   In Malaysia, the literature linking income inequality and labour income is generally scarce. There is more focus on the trends and drivers of the LIS. Nagaraj and Goh (2005), for example, estimated the LIS for Malaysia from 1960 to 2004 and showed a decline in the LIS when self-employment was included, but it remained relatively unchanged when self-employment was taken out. A more recent IMF working paper analysed the LIS of fifty economies around the world and showed that Malaysia bucked the global trend of decreasing LIS with the highest increase in its LIS from 1991 to 2014 (Dao et al. 2017). The results were robust to self-employment and capital depreciation adjustments. Compared to Nagaraj and Goh's (2005) estimates, the IMF paper presented a trend reversal in the LIS for Malaysia from the earlier phase of the country's economic development. However, both papers did not establish any link with income inequality in Malaysia. (1)

   Ng, Tan, and Tan (2019) is the only study that has made this connection between income inequality and the LIS in Malaysia, highlighting the negative co-movement of the Gini coefficient and the LIS (Figure 1). The paper aimed to explain the factors behind the increase in the LIS, attributing it to the country's inability to move away from more labour-intensive modes of production, particularly in traditional service subsectors, signalling a persistent trend of deindustrialization. It poses the question on whether the increase in the LIS, although well-disposed towards reducing income inequality, is facilitative of Malaysia's ambitions to move up the technological ladder.

   A box article in Bank Negara Malaysia Annual Report 2018 with the title 'Are Malaysian Workers Paid Fairly? An Assessment of Productivity and Equity" also cautioned that an increase in the LIS does not necessarily imply higher wages for workers (Murugasu, Hakim, and Yau 2019). The authors pointed out that the target to increase the LIS in the 11MP, although needed because the LIS is lower compared to countries with similar productivity levels, should be driven by higher-waged jobs. However, its assessment seems to differ from that of Ng, Tan, and Tan (2019) in that it explained low wages were caused by unfair compensation and not structural features of the economy. Nonetheless, both assessments underscore the multiple pathways that can lead to an increase in the LIS, giving more weight to structural possibilities and policy alternatives that can coalesce the rising LIS with reindustrialization.

   Utilizing the SAM approach to lend more evidence to the relationship between income inequality and the LIS is limited, as the first publicly available SAM was only released in 2017 (for the year 2014). This was followed by the release of the SAM for 2015 at the end of 2019. While there are studies that have analysed the SAM using matrices constructed by the researchers themselves (Mukaramah, Zakariah, and Azali 2012; Masoud et al. 2017; Zakariah 2007), they have mostly focused on the methodology and framework of the SAM. For instance, the link between production and households has been examined using multiplier analysis to assess the impact of an exogenous injection on group-based disparities, most notably ethnic and regional disparities.

   The DINA approach is even newer in Malaysia. Muhammed and Li (2019) produced the first and only set of DINA for Malaysia for the 2002-14 period using national accounts, household surveys, fiscal data and demographic statistics. The study paid more attention to the decline in the capital income share and explained it in terms of the ethnic dynamics in corporate ownership changes.

   Both the SAM and DINA approaches hold a lot of potential to examine the connections between income inequality and labour income, but they are too nascent at this point to further our understanding of the observed co-movement of the rising LIS and...