Comparing fiscal policies and GDP to poverty rates in different countries.

gdp.csv

government_spending.csv

poverty_rates.csv

ruthless.py

# Michael A. Alcorn (airalcorn2@gmail.com)
# Data from:
#  -https://en.wikipedia.org/wiki/List_of_countries_by_tax_rates.
#  -https://en.wikipedia.org/wiki/Government_spending#As_a_percentage_of_GDP
#  -https://en.wikipedia.org/wiki/List_of_countries_by_GDP_(PPP)
#  -https://en.wikipedia.org/wiki/List_of_countries_by_percentage_of_population_living_in_poverty

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import re
import statsmodels.api as sm

from csv import DictReader

reader = DictReader(open("tax_rates.csv"))

country_data = {}

# Read in tax data.
tax_per = r"\d+\.?\d*%"

for row in reader:
    country = row["Country"]
    corporate_tax = re.search(tax_per, row["Corporate tax"])
    if corporate_tax is not None:
        corporate_tax = float(corporate_tax.group()[:-1])
    
    individual_tax = re.search(tax_per, row["Individual (max)"])
    if individual_tax is not None:
        individual_tax = float(individual_tax.group()[:-1])
    
    country_data[country] = {"corporate_tax": corporate_tax, "individual_tax": individual_tax}

# Read in goverment spending data.
reader = DictReader(open("government_spending.csv"))
for row in reader:
    country = row["Country"]
    govt_spend = re.search(r"\d+\.?\d*", row["Govt. expend. % GDP"])
    if govt_spend is not None:
        govt_spend = float(govt_spend.group())
    
    if country in country_data:
        country_data[country]["govt_spend"] = govt_spend

# Read in GDP data.
reader = DictReader(open("gdp.csv"))
for row in reader:
    country = row["Country"]
    gdp = row["GDP (millions of Int$)"].replace(",", "")
    gdp = float(gdp)
    
    if country in country_data:
        country_data[country]["gdp"] = gdp

# Read in poverty data.
reader = DictReader(open("poverty_rates.csv"))

data = []

for row in reader:
    country = row["Country"]
    poverty_rate = re.search(r"\d+\.?\d*", row["CIA"])
    if poverty_rate is not None:
        poverty_rate = float(poverty_rate.group())
    
    if country not in country_data:
        continue
    
    row = country_data[country]
    row["country"] = country
    row["poverty_rate"] = poverty_rate
    data.append(row)

# Put data in data frame.
df = pd.DataFrame(data)

# Take log of GDP.
df["log_gdp"] = np.log(df["gdp"])


def plot_scatter(col, label):
    (fig, ax) = plt.subplots()
    plt.scatter(df[col], df["poverty_rate"])
    for (i, txt) in enumerate(df["country"]):
        ax.annotate(txt, (df[col].iloc[i], df["poverty_rate"].iloc[i]))
    
    plt.xlabel(label)
    plt.ylabel("Poverty Rate")
    plt.show()


# Plot data.
plot_scatter("individual_tax", "Individual Tax (max)")
plot_scatter("corporate_tax", "Corporate Tax")
plot_scatter("govt_spend", "Government Spend")
plot_scatter("log_gdp", "Log GDP")
# plot_scatter("gdp", "GDP")

# Put poverty on zero to one interval.
df["poverty_p"] = df["poverty_rate"] / 100

# Build logistic regression model.
logit = sm.Logit(df["poverty_p"], df[["individual_tax", "corporate_tax", "govt_spend", "log_gdp"]], missing = "drop")
result = logit.fit()
result.summary()

tax_rates.csv