by Shona Carter
Poverty is has been a global challenge for many years as
leaders around the world work together to combat its devastating effects. Although success in poverty reduction has
been seen globally, there still remains persistent challenges to combating the
issue. Poverty remains a big problem in Latin America. As globalization spreads
throughout the world, partnerships regarding politics and economics aim to
reduce poverty because it can increase shared prosperity. Financial globalization has been considered a way of improving struggling economies and thereby
reducing poverty. I hypothesize that more financial integration would lead to lower levels of poverty. I will assess the impact of financial openness on poverty in
Latin America because poverty remains an issue in the region and there strong data exists regarding financial integration in the region. The independent variable was obtained from a financial openness
index produced by economists Menzie Chinn and Hiro Ito. The index measures the
extent to which a country’s capital account is open or has unrestricted capital account transactions. Previous measures of financial openness were mostly binary variables that tell if a country is “open” or “closed”, without
accounting for capital accounts that are partially open or partially closed. To
overcome this, the Chinn-Ito index uses previous measures and 4 binary
variables that indicate a country’s financial openness: the presence of
multiple exchange rates, restrictions on current account transactions,
restrictions on capital account transactions, and export proceeds surrender
requirements. These variables together measure the intensity of capital
controls better than a binary variable that only considers whether a capital
account is open or closed.
The variable for poverty, which was obtained from the World
Bank’s listing of economic indicators, is a very scarce variable that is
defined by the percentage of people living on less than $1.25 per day. Due to
the scarcity of the poverty variable, I created a new variable using factor
analysis. Two additional variables, life expectancy and infant mortality were
used to construct a new measure of poverty due to their high correlation with
the original poverty variable. I will
use this new variable to run a regression with financial openness variable and
poverty in order to get a meaningful idea of the correlation between financial
openness and poverty in Latin America. Due to the negative relationship between
infant mortality and life expectancy, the factor variable has a negative
relationship to the original poverty variable.
I will control for social and political indicators of globalization,
as they may influence the degree to which a country is financially
integrated. The social globalization
variable was made using information regarding international connections, such
as tourism, also the flow of information, such as internet users. It also uses cultural
proximity which could be defined using the number of international businesses per capita. Social
and Political globalization variables come from a Quality of Governance
dataset. Political globalization measures the degree to which a country has a relationship
with other countries on a government to government level. This variable
includes information such as the number of embassies a country has and its
memberships within international institutions. I
use cross section time series analysis to create multple regressions using the
STATA software. I also account for issues that arise with panel data such as
autocorrelation.
Cross Section Time Series
Analysis
Methods and Tables:
The statistical method that I
used required me to set up the data by the country and year so that I can make
the regressions. I ran a hausman test
which helps determine whether I should used fixed effects or random effects in
my regression. These effects would help me to obtain a valid statistical
inference when there is unobserved heterogeneity ( varying unobserved components of the effect of my estimation) The
hausman test supports my using fixed effects in my regressions.
Table 1 shows 5 models. The main independent variable,
financial integration, and two additional independent/control variables, social
and political globalization, are regressed with the original poverty variable.
They are also regressed with the factor variable, as well as the two dependent
variables, infant mortality and life expectancy, which are highly correlated
with poverty. In the second model, only
the observations for the poverty variable are regressed with the factor
variable.
Table 1: Financial
Integration’s Impact on Poverty in Latin America (year effects)
|
(1)
|
(2)
|
(3)
|
(4)
|
(5)
|
|
|
VARIABLES
|
Poverty
|
Factor(for
Poverty)
|
Factor
|
Infant Mort.
|
Life Expect.
|
|
|
|
|
|
|
Financial Integration
|
-4.633**
|
0.0195
|
-0.117***
|
2.654**
|
-0.913***
|
|
(1.967)
|
(0.0744)
|
(0.0328)
|
(1.183)
|
(0.239)
|
Political Globalization
|
-0.00791
|
0.0102***
|
0.00347***
|
-0.173***
|
0.0122
|
|
(0.0667)
|
(0.00252)
|
(0.00129)
|
(0.0471)
|
(0.00954)
|
Social Globalization
|
-0.304**
|
0.00725
|
0.0150***
|
-0.434***
|
0.137***
|
|
(0.128)
|
(0.00485)
|
(0.00215)
|
(0.0777)
|
(0.0157)
|
Constant
|
17.70**
|
-0.188
|
-1.296***
|
89.19***
|
56.46***
|
|
(7.392)
|
(0.358)
|
(0.128)
|
(4.632)
|
(0.940)
|
|
|
|
|
|
|
Observations
|
148
|
148
|
997
|
1,021
|
1,008
|
Number of id1
|
14
|
14
|
27
|
27
|
27
|
Standard
errors in parentheses
***
p<0.01, ** p<0.05, * p<0.1
The first table shows the regressions with year fixed
effects and the second table shows the regressions with year and country
effects. Table 1 and 2 produce similar
results with the coefficients aligning in the same directions.
Table
2: Financial Integration’s Impact on Poverty in Latin America (country effects)
|
(1)
|
(2)
|
(3)
|
(4)
|
(5)
|
|
|
VARIABLES
|
Poverty
|
Factor
(for poverty)
|
Factor
|
Infant
Mortality
|
Life
Expectancy
|
|
|
|
|
|
|
Financial Integration
|
-4.173**
|
0.655***
|
-0.178***
|
6.702***
|
-1.307***
|
|
(1.819)
|
(0.0550)
|
(0.0463)
|
(1.623)
|
(0.323)
|
Political Globalization
|
-0.0681
|
|
0.0225***
|
-0.674***
|
0.143***
|
|
(0.0566)
|
|
(0.00161)
|
(0.0509)
|
(0.0106)
|
Social Globalization
|
-0.139
|
|
0.0474***
|
-1.656***
|
0.374***
|
|
(0.0971)
|
|
(0.00233)
|
(0.0749)
|
(0.0156)
|
Constant
|
25.68***
|
-0.0624**
|
-2.885***
|
138.5***
|
46.03***
|
|
(3.550)
|
(0.0287)
|
(0.0787)
|
(3.439)
|
(0.814)
|
|
|
|
|
|
|
Observations
|
148
|
1,189
|
997
|
1,021
|
1,008
|
R-squared
|
0.203
|
0.109
|
0.623
|
|
|
Number of id1
|
14
|
32
|
27
|
27
|
27
|
Standard
errors in parentheses
***
p<0.01, ** p<0.05, * p<0.1
Table 3: Financial
Integration’s impact on Poverty in Latin America (lagged independent variables)
|
(1)
|
(2)
|
|
|
VARIABLES
|
Factor
|
Factor
|
|
|
|
Financial
Integration
|
0.168
|
0.0432
|
|
(0.103)
|
(0.0312)
|
Lag
Social Globalization
|
0.0118**
|
0.0137***
|
|
(0.00558)
|
(0.00178)
|
Political
Globalization
|
0.00987***
|
0.00481***
|
|
(0.00254)
|
(0.000848)
|
Lag
Political Globalization
|
0.0135***
|
0.00488***
|
|
(0.00251)
|
(0.000865)
|
Social
Globalization
|
0.0343***
|
0.0131***
|
|
(0.00543)
|
(0.00171)
|
Lag
Financial Integration
|
-0.364***
|
0.0261
|
|
(0.102)
|
(0.0302)
|
Constant
|
-2.854***
|
-1.011***
|
|
(0.0791)
|
(0.178)
|
Standard
Error/Auto. Adjustment?
|
No
|
Yes
|
|
|
|
Observations
|
972
|
972
|
R-squared
|
0.635
|
0.605
|
Number
of id1
|
27
|
27
|
Standard errors in
parentheses
*** p<0.01, **
p<0.05, * p<0.1
In these regressions for table 3, I decided to lag the
independent variables in order to get a more accurate estimate of the true effect
of financial integration. By lagging the independent variable, the models consider
previous values of X for the effect on Y, and could produce the most meaningful
of all the models. Tables 1, 2, and 3
show a progression of obtaining the most accurate and meaningful results.
Results:
Based on the most controlled and accountable regression in
table 3, financial integration has a weak or non-significant effect on poverty.
The simpler regressions in the first 2 tables that do not
include lagged variables or account for auto regression, show a significant yet
inconsistent effect on the scarce poverty variable. Keeping in mind that the
factor variable and poverty should have a negative relationship, these
regressions do not show the expected results. For example, since financial
integration is associated with a decrease in poverty in model 1 on the first 2
tables, then financial integration should lead to a positive coefficient in the
factor variable in model 3. The expected direction of the factor variable is
only shown in model 2 where it is regressed solely with the same units for the
poverty variable. The fact that the coefficients for the factor variable and
poverty go in the same direction in models 1 and 3 in the first two tables,
means that those results are inconsistent.
In the third table I set out to get the most accurate
regression to account for the discrepancies in tables 1 and 2 by using lagged
variables with standard error adjustments. In model 1 of this table with lagged variables,
I used the same process for working with panel data that I used in the previous
two models. In model 2 of this table, I used a different process to calculate
panel corrected standard error estimates that help with the issue of
auto correlation.
The results from the first tables show that financial
integration has a significant effect on the scarce variable for poverty, loosely
showing that financial integration leads to a decrease in poverty, however
since the poverty variable is scarce and the factor variable produces
inconsistent results, model 1 is not enough to allow me to reject the null
hypothesis. Therefore I would fail to reject the null hypothesis that financial
integration has no relationship to poverty in Latin America. The 2 other
independent variables however, show a significant relationship to life
expectancy and infant mortality in the expected directions.
Conclusion:
The regressions produce inconsistent results
despite high levels of significance. After making the proper adjustments to the
regression over time as I have gradually done from tables 1 through 3, the
effects of financial integration remain volatile. This tells me that either
there is likely more to this analysis than what these particular variables
contribute or financial integration is not a strong predictor of poverty levels.
It does not mean that financial integration is definitely not associated with
reductions in poverty, but rather my regression does not produce a stable
result that controls for the issues associated with panel data.
Appendix:
Summary Statistics
Variable
|
|
Mean
|
Std.
Dev.
|
Min
|
Max
|
Observations
|
poverty
|
overall
|
14.1763
|
15.71471
|
0.03
|
68.49
|
N = 165
|
|
between
|
|
15.20486
|
1.62
|
61.75917
|
n = 14
|
|
within
|
|
5.382453
|
2.527136
|
36.82714
|
T = 11.7857
|
Factor
|
overall
|
1.20E-10
|
1
|
-4.004508
|
1.487515
|
N = 1684
|
|
between
|
|
0.6822429
|
-2.025807
|
0.9966946
|
n = 36
|
|
within
|
|
0.741246
|
-2.262154
|
1.790473
|
T = 46.7778
|
financial
integration
|
overall
|
0.4315984
|
0.3442466
|
0
|
1
|
N = 1313
|
|
between
|
|
0.2168614
|
0.1247907
|
1
|
n = 34
|
|
within
|
|
0.2723456
|
-0.245791
|
1.130454
|
T-bar = 38.6176
|
Political
Globalization
|
overall
|
51.74875
|
19.25492
|
10.19148
|
94.72778
|
N = 1273
|
|
between
|
|
17.81613
|
18.27359
|
85.22592
|
n = 33
|
|
within
|
|
9.09341
|
25.41777
|
78.00189
|
T = 38.5758
|
Social
Globalization
|
overall
|
39.82963
|
11.57027
|
12.46371
|
66.42154
|
N = 1273
|
|
between
|
|
9.886109
|
14.93896
|
63.53885
|
n = 33
|
|
within
|
|
6.339785
|
25.87805
|
66.74578
|
T = 38.5758
|
infant
mortality
|
overall
|
45.07957
|
34.22633
|
5
|
202
|
N = 1806
|
|
between
|
|
22.89857
|
13.7125
|
116.6556
|
n = 36
|
|
within
|
|
25.74855
|
-20.65191
|
136.4481
|
T-bar = 50.1667
|
Life
expectancy
|
overall
|
67.26714
|
7.067833
|
42.159
|
79.70502
|
N = 2027
|
|
between
|
|
5.206262
|
52.75044
|
77.01164
|
n = 43
|
|
within
|
|
5.132249
|
51.33151
|
79.94593
|
T-bar = 47.1395
|
***************************************************************************
Graph: Financial Integration and Poverty
A strong
effect would show a clear positive relationship between financial integration
and the factor variable which is the same as a negative relationship between
financial integration and poverty.
Sources:
http://qog.pol.gu.se/data/datadownloads/qogstandarddata
http://data.worldbank.org/indicator/SI.POV.DDAY
http://web.pdx.edu/~ito/Chinn-Ito_website.htm
Governments throughout the years have allocated a significant amount of their GDP to education, in hopes that in the long run it improves upon the overall well being of the country. History has proven that investing in education prevents a vast array of social issues, such as: unemployment, poverty, environmental issues, hunger and some issues related to health; but does investing in education help reduce gender equality? This study is a cross time series analysis from 1946-2014 on all
countries in the Quality of Governance Data, measuring the effect on total public expenditure on education on gender equality. I control for GDP per capita, and assistance from international agencies. 
As aforementioned, education is associated with many development outcomes and interacts with other factors to contribute to achievements related to the Millennium Development Goals set by the United Nations. According to UNICEF and UNESCO, education underpins all social progress. If the general education level worldwide is improved, global unemployment problems can be hugely tacked leading to improvements in poverty and general standards of living, including reducing the gender inequality gap (2013). Not only is it impossible to achieve gender equality without education,
expanding education opportunities for all can help stimulate
productivity and reduce the economic vulnerability of poor households. Gender equity in and through education examines education processes in
education as well as outcomes in education to help a more gender
equitable, productive and peaceful society. Education is an important
tool in creating gender equity in society. Efforts to achieve gender
equality through informal and formal education (2012). 
The social significance of gender is that it is the device by which society controls its members, it opens and closes doors to power, property and even prestige. Like social class, it is a structural feature of society. This stratification is often referred to as the sex/gender system These types of differentiations occur not only on an interpersonal level between individuals, but also on a structural level within a given society. The shape and form of these systems can vary historically and cross culturally, as well as across populations. Thus closing the gender gap and making women equal to men has been such important part of societal development, even included in the Millennium Development Goals (2007). Additionally, it has been extensively proven through literature that improving the gender inequality gap in recent years, has increased participation of women in income-generating activities, education for girls and women becomes particularly important, especially in redressing gendered patters of discrimination. 
.png)
The data shows that historically there is not a significant relationship between public expenditure on education and gender equality. Although the relationship is not statistically significant, the charts and graphs show a positive relationship. For every dollar spend on education, there is a .00099 increase in gender equality. With this in mind the data suggests that if we want to progress upon the Millennium Development Goal of improving gender equality, perhaps it would be more valuable to allocate those specific resources to other initiatives besides educational program. For example, the data shows a positive and significant relationship between gender equality and GDP per capita, thus, perhaps investing in initiatives that incentivize financial growth result more effective when tackling the issue of gender inequality. Therefore, a more complete follow-up to this study could integrate a regression of other initiatives suggested by the literature that have the potential of improving gender equality across nations. Also, the study could be done separating countries by region, in which perhaps expenditure on education might have a bigger effect. Finally, this study poses helpful information for policymakers, specially in the departments of the government working towards gender equality to explore different alternatives besides educational initiatives. 
