Data Science 2021

by Marina Barats

    

    

Project

Topic: Is higher social mobility associated with a more favourable business climate?


Social mobility refers to the movement of individuals either up or down the socioeconomic ladder relative to the position of their parents. Global Social Mobility Index (SMI) measures the impact of socio-economic background on one’s outcomes in life. It is determined based on the following indicators: health, education, technology access, work opportunities, working conditions and social protection. This project seeks to understand the relationship between business climate and social mobility. This is done by evaluating entrepreneurship indicators around the world and plotting SMI and related indicators against them.

As evident from Figure 1, on average, SMI is higher in Europe and lower in African and Caribbean states. SMI varies the most in Asia, with countries like Bangladesh having SMI as low as 40, and states like Japan scoring over 70.

Find the Colab notebook for Figure 1 here. SMI is scraped from the Business Insider website ( here is the Colab notebook), and poverty data is scraped from Wikipedia (here is the Colab notebook). Poverty values are in percentage and are not recognised as numeric in Vega-Lite. I remove the “%” sign, change the data type to numeric and create a new data frame. Then I join it with the original data frame, as they have one column in common.

The data for Figure 2 is downloaded using the World Bank library in Python. The Colab notebook is availible here. As evident from Figure 2, the cost of opening a business in states with the lowest SMI (Guatemala, Senegal, Cameroon) has been declining significantly over the last two decades. Meanwhile, for countries with the highest SMI (Canada, UK, Denmark), it slightly fluctuated but stayed below 4% of income per capita. Clearly, there is a negative relationship between SMI and the cost of opening a business.

Judging from Figure 3, there is some positive relationship between the cost of starting a business and the days it takes. Most Western, some Asian and South American states follow this pattern. However, there are many outliers, such as South Africa, where it takes almost the longest to start a business, but it costs the least.

The data for Figure 3 is accessed via the World Bank API using For loops in Python to iterate through the country list and obtain name and value for each country. See the Colab notebooks for cost, days to open a business, and combining the two in one dataset.

The poverty level is negatively correlated with SMI, which is evident from this regression analysis. A positive relationship between the cost of starting a business and the proportion of population in poverty is clear from Figure 4.

Cost is acquired via the World Bank API using For loops in Python. The cost data frame is then joined with an existing social mobility and poverty dataset. Here is the Colab notebook.

Global Entrepreneurship Index (GEI) is calculated by the Global Entrepreneurship and Development Institute and is scraped from their website. The regression equation for Figure 5 was calculated using the regplot Seaborn function in Colab and then was plugged into Vega-Lite to generate a regression plot. Here is the Colab notebook. R squared (R2) is calculated in Vega-Lite. R2 is the square of the correlation between the independent and dependent variables, it measures the proportion of variation in the dependent variable that can be attributed to the independent variable. R2 of 0.72 means that 72% of the variation in SMI is explained by GEI, which suggests a strong positive correlation between the two indicators.

New business registrations data was downloaded from the World Bank website and edited in Colab. It is then joined with the data on social mobility and poverty. Here is the Colab notebook. There is a positive correlation between new business registrations and SMI. As anticipated, Cyprus, Luxemburg and Malta, countries with low tax rates, have more business registrations for their SMI relative to other states in the sample.

Conclusion:

Several entrepreneurship indicators have been plotted against the Social Mobility Index and poverty level. The research has identified a negative relationship between the cost of starting a business and SMI, a positive correlation between the cost of starting a business and poverty, a positive correlation between the Global Entrepreneurship Index and SMI and a positive correlation between business registrations and SMI. Therefore, there is a clear positive correlation between social mobility and a favorable business climate.

This project does not intend to establish causal connections between the indicators, which may be investigated in further research. Furthermore, the relationship between social mobility and entrepreneurship may be subject to omitted variable bias from not accounting for variables like GDP per capita or political stability. This should be considered in research aiming to estimate causal relationship between the two.