![Sample our Social Sciences journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/110822/TOC-Subject-Sample-2021-Social-Sciences.jpg"})
ABSTRACT
It has been long believed that prolonged political instability harms economic growth and development. This paper contributes to this growing empirical literature by studying the case in Fiji, which has faced a long period of political instability caused by a series of coups, military administrations and frequent changes in government since 1987. The impact of political instability on growth is hard to identify empirically because the counterfactual is unobserved and it is difficult to find valid comparisons. To solve this problem, we use the recently developed Synthetic Control Method to construct a counterfactual (or synthetic Fiji) that predicts the growth of a politically stable Fiji. The difference in per capita growth trajectories of the synthetic and the actual Fiji can thus be attributed as the impacts of political instability. Our findings show that the political instability caused by a series of coups since 1987 has indeed led Fiji onto a lower growth path, and that the accumulated effect is getting larger.
Notes on contributors
Xiaodong Gong is an Associate Professor, the Institute of Governance and Policy Analysis (IGPA), University of Canberra. Maheshwar Rao is an independent scholar. Both authors have equally contributed to the paper.
Notes
1 Per capita GDP is not given in the Penn Tables. We divide real GDP in PPP 2005 US dollars by population (both available in Penn Tables) to calculate the per capita GDP in PPP 2005 US dollars.
2 Available at www.barrolee.com. We used education attainment (defined as secondary and tertiary school completions) for the percentage of population aged 25 years and over. See Barro and Lee (Citation1993, Citation2013) for the estimation and description of the dataset.
3 Available at www.databank.worldbank.org.
4 The demographic structure of Fiji has changed dramatically since the coup of 1987. In 2007 census, the share of the Indo-Fijian population comprised around 37% (both due to falling birth rates and an increase rates migration) and the Fijian share had risen to 58%.
5 The Synthetic Control Method has found wide applicability in estimating the effects of events and interventions across many disciplines and areas of interest. Examples include: Liou and Musgrave (Citation2013), the effects of resource income on democratisation; Mideksa (Citation2013), the economic impact of natural resources; Fowler (Citation2013) and Carey and Horiuchi (Citation201 Citation4), the effect of compulsory voting on voter turnout; Montalvo (Citation2011), the effect of timing of terrorist attacks on democratic election outcomes; Nannicini and Ricciuti (Citation2010), the effect of autocratic transitions on growth; Billmeier and Nannicini (Citation2013), the effect of trade liberalisation; Cavallo et al. (Citation2013) and Coffman and Noy (Citation201 Citation2), the effect of natural disasters; Singhal and Nilakantan (Citation2012), security response on terrorism; and Abadie, Diamond, and Hainmueller (Citation2015), the impact of the 1990 German reunification.
6 A large literature exists using the differences-in-difference methods. See, for example, Card (Citation1990), Card and Kruger (Citation1994), and a recent literature review by Imbens and Wooldridge (Citation2009).
7 Per capita GDP is not given in the Penn Tables. We divide real GDP in PPP 2005 US dollars by population (both available in Penn Tables) to calculate the per capita GDP in PPP 2005 US dollars.
8 Available at www.barrolee.com.
9 Available at www.databank.worldbank.org.
10 For some countries, a few indicators are missing for a few years in the early period. In such cases, the shortened series are used. For details of the data availability, see in the Appendix.
11 First, we take the average value of the pre-coup characteristics for each country in the donor pool (similar to the calculation of the average values for actual Fiji). The mean is simply the average of these values of the countries in the donor pool. The standard deviations (in brackets in ) are the variations among the countries in the donor pool from their pre-coup average values.
12 From 1987 to 2011, Fiji’s real per capita GDP declined from $4300.4 to $3395. The average growth rate was about −0.9% annually. The real per capita GDP for the Synthetic Fiji was $7961.0 in 2011, growing from $5126 in 1987. The average growth rate is about 2.3% annually. Thus the difference is about 3.2%.
13 We also conducted the placebo test using 1980 as an alternative hypothetical coup year. Results are similar.
14 For in-time placebo test, the data for the predictors are required for the (pre-intervention) 1970–1977 period. Data for a few countries are not available for the whole 1970–1977 period. In such cases, the data available for the closest year are used as the average of the 1970–1977 period. For instance, for Belize, the data for agriculture share of GDP are available from 1979. In this case, the 1979 value for the agriculture share of GDP is used as the average of the 1970–1977 period for the in-time placebo test.