The impact of institutions on poverty

Размещено на http://allbest.ru

¹Kyrgyz Russian Slavic University

²American University of Central Asia

The impact of institutions on poverty

¹Osmonova Ainur Anvarovna

Doctor of Science, Economics, professor

²Osmonova Tinatin Anvarovna Master of Science, Economics,

²Emirova Altynai Emirovna Master of Business Administration

Summary

This paper analyses the impact of six different measurement of institutional quality on FGT class of poverty indices. By employing two estimation techniques as fixed effects with instrumental variable and Arellano-Bond Generalized Method of Moments the study tests how “Control of Corruption”, “Government Effectiveness”, “Political Stability and Absence of Violence”, “Regulatory Quality”, “Rule of Law”, “Voice of Accountability” affect poverty headcount, poverty gap and severity of poverty. The panel data models are constructed for 15 lower middle-income countries, 22 upper middle-income countries and 27 high-income countries and the work covers the period between 2006 and 2016. The data on countries with different income levels estimated separately. The main hypothesis of the work is that institutional quality is a key to prosperity and economic growth, whereas low quality of institutions leads to stagnation and poverty. The findings demonstrate that effective governance, regulatory quality and rule of law might reduce poverty in lower middle-income and upper middle-income countries. In case of high-income countries, all six institutional quality measures show negative and significant impact on poverty.

Key words: economics, institutions, institutional quality, economic growth, poverty, development, IV approach

Introduction and literature review

Economic point of view considers poverty as “lack of money or other material possession” (Webster, 1993) to meet basic needs like food, clothing and housing. The effects of poverty are very dramatic. Thus, the World Bank defines poverty as hunger, sickness and not having access to education. However, poverty in each country might have different causes and effects.

One countries experience absolute poverty, while in other countries poverty is relative term. Thus, there are no standard poverty reduction strategies. Providing basic school education, food and medical treatment might boost the economy of extreme poor African countries (UNDP).

Nevertheless, the same way of poverty reduction might not have any effects to the economy of the lower middle-income country as Kyrgyzstan, where almost 3.5 million out of 6.14 million people are literate (NSC, 2017). Despite these impressive statistics in educational sphere, poverty in the country still exists. More than 1.5 million poor people live in Kyrgyzstan (NSC, 2016). Hence, the causes of poverty in the Kyrgyz Republic are different from the ones determined in the other countries.

Such kind of distinctions in poverty definition, provoked scientist to analyze the problem deeply. The recent growth theories revealed an institutional approach as one of the key reasons for economic development and poverty reduction (North, 1991; Acemoglu et al. 2002; Rodrik, 2008; Tebaldi and Mohan, 2010).

The approach explains the principle of effective regulations and policies that are necessary for prosperity.

The concept states that factors as democracy, regulatory quality, effectiveness of government, control of corruption, strong property rights or rule of law are important in creating welfare and material benefits (North, 1991; Sen, 1999; Acemoglu et al., 2003; Rodrik, 2000). Institutional theory specifies good and poor institutions. According to North (1991), good institutions lead to prosperity and growth, whereas low quality of institutions lead to stagnation and poverty. Rodrik (2008) points out that structured governance and good institutions are required, since society or markets are not self-stabilized, self-legitimized and self-regulated.

In general, the idea states that the quality of institutions have significant effect to growth and poverty reduction. Since the institutional theory is still new and many aspects related to the approach might be discovered, the current work, based on the existing literature aims to check the impact of institutional quality on poverty of 64 both rich and poorer countries for the period between 2006 and 2016.

The literature observed in the paper helped define proper measurement for poverty and institutional quality.

According to Easterly and Levine (2003), Fabro and Aixala, (2009), Tebaldi and Mohan (2010), Perera and Lee (2013), “Control of Corruption”, “Government Effectiveness”, “Political Stability and Absence of Violence”, “Regulatory Quality”, “Rule of Law”, “Voice of Accountability” are used as measurements of institutional quality. Based on Chong and Calderon (2000), Tebaldi and Mohan (2010), Perera et al. (2013), poverty headcount ratio, poverty gap and severity of poverty are chosen as main indicators of poverty. All above mentioned authors take poverty line at 1 USD or 1.9 USD per day for set of developing and developed countries.

However, the World Bank's data for the last 10 years show that there are no many poor people living below 1.9 USD per day in high income counties. Hereby, by taking into account all possible shortcomings of previous studies, the current research estimates the data at different poverty lines for countries with different level of income.

Specifically, it takes into account the World Bank's segregation and examines the data at 1.9 USD, 3.2 USD and 5.5 USD poverty lines for lower-middle income countries, for upper-middle income counties and for high income ones respectively. The other measurements of poverty are not examined in this work due to complexity of data collection for large panel of countries.

The current study contributes to the literature on poverty and institutions, since it tests six measurements of institutional quality on three FGT poverty indices. All possible shortcomings of other empirical works are overviewed, thus new estimation methods are applied to run regressions with panel data.

Moreover, the work classifies countries into three main groups by income level. Precisely, the current paper checks how institutional quality affects poverty in lower middle-income countries, in upper middle ones, and in high-income countries for the period between 2006 and 2016.

Based on the World Bank's classification by income (2017), the data on 15 lower middle-income countries, 22 upper middle-income and 27 high-income countries is empirically tested. This way of doing analysis helps define real influence of proper regulations, control of corruption, and political conditions to the level of poverty in both rich and poor countries.

To calculate the data, an econometric model is built. The model includes poverty measure as a dependent variable and measurement of institutional quality as one of the independent variables.

To avoid multicollinearity problem, all six measures of institutional quality are tested separately, not simultaneously. Following existing theory and available literature, the current work employs fixed effects with instrumental variable and Arellano-Bond Generalized Method of Moments (GMM) estimation techniques. Since week instruments might bring incorrect confidence intervals and wrong tests of significance (Chao and Swanson, 2005), it is necessary to diagnose whether an instrument is week or not. Following Tebadi and Mohan (2010), the Anderson-Rubin test is employed. Perera et al. (2013) suggest performing unit root test and applying one of the GMM techniques according to the test results. The current work follows their recommendations. Two estimation methodologies are chosen for the robustness check and both estimation techniques are suitable to control possible endogeneity issues.

The main research question and hypothesis is that strong institutions are key factors to reduce poverty, increase prosperity and economic growth of a country. Good institutions lead to better performance in any economic spheres, whereas low quality of institutions causes stagnation and poverty. It is assumed that the impact of institutional quality variables will show various results, because of the different estimation methods and different indicators of poverty and countries level of income.

Model and data description

Based on the observed literature, the following econometric model is built:

Pi,t= в0 + в1IQi,t + в2popgrowthi,t + в3GDPi,t +

в4humcapi,t + еi,t.,

where P denotes all poverty measures as headcount index, poverty gap, severity of poverty. The main independent variable IQ denotes all institutional measures as Control of Corruption (CO), Government Effectiveness (GE), Political Stability and Absence of Violence (PSAV), Regulatory Quality (RQ), Rule of Law (RL), Voice of Accountability (VA). The variable popgrowth measures percentage change in population; GDP is a measurement of economic growth; humcap is percentage of population attending secondary school. The sing е indicates an error term, i denotes country and t is time dimension.

Panel data analysis reduces collinearity between independent variables because it includes large data sets and offers more degrees of freedom (Hsiao, 2003). However, the current research considers data on different countries with different characteristics, which renders heteroscedasticity problem. The simple OLS estimator might present inconsistent and biased results (Lee and Azali, 2010; Wooldridge, 2015). The standard estimators for panel data analysis are fixed effects, random effects or first difference estimator (Wooldridge, 2015). The fixed effects remove the causes of time-invariant characteristics. On the other hand, the problem of endogeneity exists in the model, since institutional variables and economic growth have causal relationship to poverty. To eliminate this problem, the existing literature suggests using instrumental variable technique or running the regression through GMM estimator (Perera et al., 2013).

The current work aims to deal with all possible shortcomings in estimations and get unbiased results. According to La Porta et al. (1999), Acemoglu et al. (2001) latitude of a country is an appropriate instrument for institutional quality. Tebaldi and Mohan (2010) suggest applying Anderson-Rubin test to reveal whether an instruments weak or not. The same techniques are used in the current work, fixed effects with instrumental variable is employed to control for unobserved heterogeneity and eliminate endogeneity issue. To check robustness of results, one of the GMM methods is appropriate. Following Perera et al. (2013), the study also applies the panel unit root test (Levin, Lin, and Chu, 2002) to check for non-stationarity or existence of unit root.

Results

Fixed effects with instrumental variable. Fixed effects (FE) and first difference (FD) models are usually used as a solution for the problem of unobserved heterogeneity of the panel data (Wooldridge, 2015).