# load libraries

import os
os.getcwd()
os.chdir("C:\\Users\\user\\Downloads\\Thesis_paper_1\\dados")

import numpy as np
import matplotlib.pyplot as plt
import time
import pandas as pd
import datetime 
import statsmodels.api as sm
import wbdata 

Load IBRE dataset

df = pd.read_excel('prod_trab.xlsx', sheet_name='prod_hr')

Average Produtivity Growth

The table shows that productivity of industry and services are stagnant.Despite years of schooling increasing.

# slice df

ch = df[['Anos', 'Agropecuária', 'Industria Total', 'Serviços Total', 'Total', 'hc']]


# calculate pct

pct = ch.iloc[:,1:].pct_change().iloc[1: , :]

pct3 = pct

# define periods

ind1 = ['1995-2003', '2003-2011', '2011-2016', '2016-2019']


# set year

pct['ano'] = df['Anos']



# create boolean

pct['my_code']= 0
bol1 = (pct['ano']>=1996) & (pct['ano']<=2003)
pct.loc[bol1, ['my_code']] = 1
bol2 = (pct['ano']>=2003) & (pct['ano']<=2011)
pct.loc[bol2, ['my_code']] = 2
bol3 = (pct['ano']>=2011) & (pct['ano']<=2016)
pct.loc[bol3, ['my_code']] = 3
bol4 = (pct['ano']>=2016) & (pct['ano']<=2019)
pct.loc[bol4, ['my_code']] = 4


# group by total

pct['my_code2'] = 0
bol5 = (pct['ano']>=1996) & (pct['ano']<=2019)
pct.loc[bol5, ['my_code2']] = 1

pct2 = pct.groupby(['my_code2']).mean().iloc[:, 0:5]*100

pct2 = pct2.round(4)


# group by mean

pct = pct.groupby(['my_code']).mean().iloc[:, 0:5]*100
pct.index = ind1

pct.rename(columns= {'Industria Total': 'Indústria', 
                     'Serviços Total': 'Serviços', 
                     'hc': 'Anos de escolaridade', 
                    'Total': 'Produtividade agregada'}, inplace=True)


pct = pct.round(4)

#print(pct.to_latex(index=True))
#print(pct2.to_latex(index=True))
pct

	Agropecuária	Indústria	Serviços	Produtividade agregada	Anos de escolaridade
1995-2003	5.9721	-2.2409	-0.4629	0.1612	3.8692
2003-2011	6.6982	0.9278	1.5636	2.2208	1.7670
2011-2016	8.9577	-0.3645	0.0181	0.7520	1.9576
2016-2019	5.3128	1.4422	-1.0949	-0.0805	-0.0790

Estimate regression using statsmodel

The regression show that there is no relation between years of schooling and productivity in Brazil.

mod = sm.OLS(pct3.loc[:,['Total']], pct3.loc[:,['hc']] )
res = mod.fit()

print(res.summary())

                                 OLS Regression Results                                
=======================================================================================
Dep. Variable:                  Total   R-squared (uncentered):                   0.101
Model:                            OLS   Adj. R-squared (uncentered):              0.062
Method:                 Least Squares   F-statistic:                              2.581
Date:                Mon, 04 Jan 2021   Prob (F-statistic):                       0.122
Time:                        10:43:52   Log-Likelihood:                          59.799
No. Observations:                  24   AIC:                                     -117.6
Df Residuals:                      23   BIC:                                     -116.4
Df Model:                           1                                                  
Covariance Type:            nonrobust                                                  
==============================================================================
                 coef    std err          t      P>|t|      [0.025      0.975]
------------------------------------------------------------------------------
hc             0.2482      0.154      1.607      0.122      -0.071       0.568
==============================================================================
Omnibus:                        2.117   Durbin-Watson:                   1.483
Prob(Omnibus):                  0.347   Jarque-Bera (JB):                1.053
Skew:                           0.495   Prob(JB):                        0.591
Kurtosis:                       3.272   Cond. No.                         1.00
==============================================================================

Notes:
[1] R² is computed without centering (uncentered) since the model does not contain a constant.
[2] Standard Errors assume that the covariance matrix of the errors is correctly specified.

# slice dataframe

df2 = df[['Anos', 'Agropecuária', 'Industria Total', 'Serviços Total', 'Total', 'hc']]

You can conclude this in figure below.

fig,ax = plt.subplots(figsize=(10,8))
ax.plot(df2.Anos, df2.Agropecuária, label ='Agropecuária', color='gray', marker='o', markersize=10)
ax.plot(df2.Anos, df2[['Industria Total']], label='Indústria', color='crimson', marker='d', markersize=10)
ax.plot(df2.Anos, df2[['Serviços Total']], label='Serviços', color='orange', marker='s', markersize=10)
ax.plot(df2.Anos, df2['Total'],  label='Total', color='blue', marker='v', markersize=10)
ax.set_ylabel("Produtividade do trabalho", fontsize=18)
ax.set_xlabel("Anos", fontsize=18)
plt.legend(loc="center left", prop={'size': 16})
plt.xticks(np.arange(min(df2.Anos), max(df2.Anos)+1, 2.0), fontsize=20, rotation=45)
plt.yticks(fontsize=20)
plt.grid(True)
ax2=ax.twinx()
ax2.plot(df2.Anos, df2.hc, label = 'Anos de escolaridade', color='limegreen', marker='X', markersize=10)
ax2.set_ylabel("Anos de escolaridade", fontsize=18)
plt.yticks(fontsize=20)
plt.legend(loc="lower right", prop={'size': 16})
plt.tight_layout()    
#plt.savefig("C:/Users/user/Downloads/Thesis_paper_1/paper_tex/fig1.eps", format='eps', dpi=2000)

Stagnant productivity is a global trend

Using PWT dataset we verify that TFP of various countries was decrease, when we compare 1998 and 2017. See figure below.

df3 = pd.read_excel('pwt91.xlsx', sheet_name='Data')

df3.head()

	countrycode	country	currency_unit	year	rgdpe	rgdpo	pop	emp	avh	hc	...	csh_x	csh_m	csh_r	pl_c	pl_i	pl_g	pl_x	pl_m	pl_n	pl_k
0	ABW	Aruba	Aruban Guilder	1950	NaN	NaN	NaN	NaN	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
1	ABW	Aruba	Aruban Guilder	1951	NaN	NaN	NaN	NaN	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
2	ABW	Aruba	Aruban Guilder	1952	NaN	NaN	NaN	NaN	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
3	ABW	Aruba	Aruban Guilder	1953	NaN	NaN	NaN	NaN	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
4	ABW	Aruba	Aruban Guilder	1954	NaN	NaN	NaN	NaN	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN

5 rows × 52 columns

df4 = df3[['countrycode', 'year', 'rtfpna']]

x = np.array( df4[df4.year==1980]['rtfpna'] )
y = np.array( df4[df4.year==2017]['rtfpna'] )
name = np.array( df4[df4.year==2017]['countrycode'] )

hx = x[np.argmax(1*(name=='BRA') )]+0.08
hy = y[np.argmax(1*(name=='BRA') )]

fig, ax = plt.subplots(figsize=(10, 8))
ax.scatter(x, y, s=10)
plt.scatter(hx, hy, s=1100, color='yellow')
plt.plot([0.5, 1.5], [0.5, 1.5], 'k-', lw=2, label='45° line')
plt.xticks(fontsize=20)
plt.yticks(fontsize=20)
for tt, txt in enumerate(name):
    ax.annotate(txt, (x[tt], y[tt]), size=20)  
plt.grid(True)
plt.legend(loc="lower right", prop={'size': 20})
plt.xlabel("TFP - 1980", fontsize=20)
plt.ylabel("TFP - 2017", fontsize=20)
plt.tight_layout()    
#plt.savefig("C:/Users/user/Downloads/Thesis_paper_1/dados/fig2.eps", format='eps', dpi=2000)

Total Productivity Factors in Brazil (PWT data)

The TPF in Brazil between 1955 and 1980. From 1980 it started to decrease.

df5 =  df3[df3.countrycode=='BRA'][['rtfpna', 'year']] 

x2 = df5['rtfpna']
y2 = df5['year']

fig,ax = plt.subplots(figsize=(10,8))
ax.plot(y2, x2)
plt.grid(True)
plt.xticks(np.arange(min(y2), max(y2)+5, 5), fontsize=20, rotation=45 )
plt.yticks(fontsize=20)
plt.xlabel("Year", fontsize=20)
plt.ylabel("TFP", fontsize=20)
plt.tight_layout()    
#plt.savefig("C:/Users/user/Downloads/Thesis_paper_1/dados/fig3.eps", format='eps', dpi=2000)

Spend in education (World Bank Data)

The spend in education in Brazil has increase since 2004. And since 2006 it's bigger than high income countries. See figure below.

# https://wbdata.readthedocs.io/en/stable/

data_date = datetime.datetime(1995, 1, 1), datetime.datetime(2021, 1, 1) 
country = ['BRA', 'HIC', 'LMC', 'UMC', 'LIC']

ind = {'SE.XPD.TOTL.GD.ZS': 'spend'}

df6 = wbdata.get_dataframe(ind, country = country,  data_date=data_date)

df6 = df6.reset_index()

df6 = df6.pivot(index="date", columns="country", values="spend")

df6.reset_index(inplace=True)

df6.rename(columns={"date": "Anos", "Brazil": "Brasil",
                   "High income": "Renda alta", "Low income": "Renda Baixa",
                   "Lower middle income": "Renda média baixa",
                   "Upper middle income": "Renda média alta"}, inplace=True)

#pd.to_datetime(df6.Anos, format = "%Y")

#df['col'] = pd.to_datetime(df['col'])

df6.Anos = df6.Anos.astype(str).astype(int)

fig,ax = plt.subplots(figsize=(10,8))
ax.plot(df6.Anos, df6[["Renda média alta"]], label= "Renda média alta", color='orange', marker='v', markersize=10)
ax.plot(df6.Anos, df6[["Renda média baixa"]], label= "Renda média baixa", color='blue', marker='X', markersize=10)
ax.plot(df6.Anos, df6[["Renda Baixa"]], label="Renda Baixa", color='crimson', marker='s', markersize=10)
ax.plot(df6.Anos, df6[["Renda alta"]], label='Renda alta', color='gray', marker='d', markersize=10)
ax.plot(df6.Anos, df6[['Brasil']], label='Brasil', color='green', marker='o', markersize=10)
ax.set_ylabel("Gastos do governo em educação (% do PIB)", fontsize=18)
ax.set_xlabel("Anos", fontsize=18)
plt.legend(loc="upper left", prop={'size': 16})
plt.xticks(np.arange(min(df6.Anos), max(df6.Anos)+1, 2.0), fontsize=20, rotation=45)
plt.yticks(fontsize=20)
plt.grid(True)
plt.tight_layout()    
#plt.savefig("C:/Users/user/Downloads/Thesis_paper_1/paper_tex/fig4.eps", format='eps', dpi=2000)

fig,ax = plt.subplots(figsize=(10,8))
ax.plot(df6.Anos, df2.Agropecuária/df6.Brasil, label ='Agropecuária', color='gray', marker='o', markersize=10)
ax.plot(df6.Anos, df2['Industria Total']/df6.Brasil, label='Indústria', color='crimson', marker='d', markersize=10)
ax.plot(df6.Anos, df2['Serviços Total']/df6.Brasil, label='Serviços', color='orange', marker='s', markersize=10)
ax.plot(df6.Anos, df['Total']/df6.Brasil,  label='Total', color='blue', marker='v', markersize=10)
ax.set_ylabel("Produtividade relativa aos gastos do governo \n em educação", fontsize=18)
ax.set_xlabel("Anos", fontsize=18)
plt.legend(loc="upper right", prop={'size': 16})
plt.xticks(np.arange(min(df6.Anos), max(df6.Anos)-1, 2.0), fontsize=20, rotation=45)
plt.yticks(fontsize=20)
plt.grid(True)
plt.tight_layout()    
#plt.savefig("C:/Users/user/Downloads/Thesis_paper_1/dados/fig5.eps", format='eps', dpi=2000)