Principal investigator: Adriano Valladão Pires Ribeiro
Article title: INEQUALITY AND GROWTH: WHAT CAN THE DATA SAY?
Article authors: Abhijit V. Banerjee and Esther Duflo
Location of the intervention: Country panel
Sample size: 226 observations
Main theme: Economic Policy and Governance
Primary variable of interest: Growth
Evaluation method: Kernel density estimate
Policy Problem
The relationship between social inequality and economic growth, although easily stated, is difficult to measure. The question is basically whether greater (or lesser) inequality is associated with greater or lesser economic growth. First, it is difficult to determine causality in this relationship. Second, comparing the particularities of each country creates problems, since isolated characteristics would affect both inequality and growth. Third, assuming the wrong form for the relationship can lead to distorted conclusions. The study presented below seeks to solve the last problem mentioned.
Evaluation Context
In economic literature, several studies have attempted to estimate the relationship between inequality and growth. The standard procedure assumes a linear relationship between the two, implying that changes in one variable would always be associated with changes in the same proportion in the other. Furthermore, the technique used to estimate the relationship also matters; the result obtained can be negative, positive, or nonexistent. That is, depending on the procedure, one might find that greater inequality is related to lower, higher, or no future economic growth. Finally, it is not possible to speak of causality in the results obtained.
Details of the Intervention
In the theoretical field, there are two classes of arguments about the causal relationship between inequality and growth. The first, within the realm of political economy, starts from the premise that inequality leads to wealth redistribution, and that redistribution negatively affects growth. The second is about a wealth effect, assuming there is a relationship between present wealth and future wealth.
To put the principles of political economy simply, we have two political groups competing for a country's wealth. At each point in time, a new opportunity for growth arises, but this opportunity would require structural changes and could be blocked by one of the groups. On the other hand, instead of blocking, this group could accept the change in exchange for transferring some of the other group's wealth. The time it takes for the groups to reach an agreement diminishes the potential for growth, and the final share of wealth held by each group will measure the level of inequality. Thus, not only are inequality and growth strictly linked, but changes in inequality have causal effects on growth.
The wealth effect argument stems from the idea that each individual can spend or invest their wealth, and growth would come from the decision to invest wealth. People with different levels of wealth would invest different amounts, thus creating a connection between inequality and growth. Several implications can be drawn from this thinking: (i) beyond a certain level of wealth, individual investment would not change; (ii) increased wealth dispersion reduces the growth rate; (iii) inequality and growth tend to decrease over time. Therefore, growth would decrease with both an increase and a decrease in inequality; that is, observable changes in inequality imply a decrease in the growth rate.
details of Exclusive
The data used to measure the relationship between inequality and growth are the Gini Index and the GDP growth rate. The Gini Index measures income concentration and ranges from 0 to 1, where values close to zero represent greater equality and values close to one represent greater inequality. There are several ways to capture this relationship; the five methods below are derived from the discussion of the two arguments made in the previous section, with a brief intuition for each method. Due to the difficulty in actually establishing causality, the main objective is to capture the non-linear effects between inequality and growth; therefore, none of the specifications assumes... a priori that the relationship is linear.
Way 1: The GDP growth rate as a function of current GDP and past changes in inequality. It captures the idea that changes in inequality have causal effects on growth, and also allows for these effects to vary according to the level of wealth.
Way 2: Variations in inequality as a function of GDP levels and past inequality. It assumes the relationship between inequality and growth described above to be valid, and notes that changes in inequality are caused by the level of inequality.
Way 3: The square of the change in inequality as a function of the previous level of inequality. The important factor would not be the change in inequality itself, as in the previous case, but the absolute value of the change, since changes upwards or downwards would imply less growth.
Way 4: The growth rate as a function of GDP levels and inequality from the previous period. It attempts to directly measure the effect of past inequality levels on economic growth.
Way 5: The growth rate as a function of GDP, inequality, and changes in inequality. According to the wealth effect, changes in inequality have causal effects on growth, just as in the first form. However, the current specification is more general, as it allows the effect to differ depending on the level of inequality.
Results
From the study of the methods used to obtain the relationship in the previous section, some results stand out. First, variations in inequality are related to subsequent economic growth in a non-linear way; the relationship is an inverted U-shape. This means that variations in inequality in either direction, increase or decrease, are associated with lower growth, and the greater the variation in inequality, the greater the fall in growth. Second, the relationship between variation in inequality and past levels of inequality has a strong negative correlation, while the square of the variation and the past level of inequality is positive. This tells us that smaller variations in inequality are related to higher and lower levels of inequality, while larger variations are associated with values around 0.45 of the Gini Index. Finally, the U-shaped relationship between growth and the level of inequality seems to mirror the relationship between variation in inequality and growth.
Lessons Public Policy
The main lesson lies in limiting misinterpretations about the relationship between inequality and growth. There is evidence that the relationship is non-linear, so variations in inequality could be associated with both an increase and a decrease in growth. Another point is that, due to the limitations in isolating the causal effects of inequality on growth, nothing can be said about greater or lesser inequality implying greater or lesser economic growth. Create an account on the Sportingbet platform. It's simple, just locate the registration button in the upper right corner of the website and then fill in a list of information such as name, surname, date of birth, nationality, contact information, etc.
Reference
Banerjee, Abhijit V.; Duflo, Esther. “Inequality And Growth: What Can The Data Say?,” Journal of Economic Growth, v8(3,Sep), 267-299. 2003
