2.

2.1 Data Source

This research relies exclusively on internationally standardized secondary data to ensure
comparability, consistency, and replicability across countries and years. All indicators are
retrieved from the World Bank through the World Development Indicators (WDI) database,
which provides a comprehensive set of macroeconomic, environmental, and social indicators
collected from official national and international agencies.
The WDI is widely used in empirical research and policy studies because it:
 Covers more than 200 countries and territories,
 Offers time-series data spanning several decades,
 Ensures standardized definitions and methodologies across countries,
 Provides open-access, replicable datasets suitable for statistical analysis.
For this study, the variables were carefully selected to align with the research objective of
examining the relationship between foreign direct investment (FDI) inflows and carbon dioxide
(CO₂) emissions in the Latin America and the Caribbean (LAC) region, while controlling for
key economic and structural variables. The indicators represent economic activity, environmental
outcomes, and potential mediating factors influencing emissions.
2.2 Data Scope
Title of the Dataset: “FDI–CO₂ Emissions Panel Dataset for Latin America and the
Caribbean, 2010–2023”

tem Description

Ten major economies in Latin America and the Caribbean: Argentina,

Geographic
Brazil, Chile, Colombia, Costa Rica, Ecuador, Mexico, Panama, Peru, and
Coverage
Uruguay.

Time Period Annual data from 2010 to 2023, encompassing a 14-year period.

The dataset is a balanced panel with a theoretical maximum of 140

Number of country-year observations (10 countries × 14 years). The actual number of
Observations observations used in regression models may vary slightly due to missing
data for some variables in specific years.

Data Source World Development Indicators (WDI), The World Bank.

 tem Description

Data Type Macroeconomic, country-level panel data.

Unit of Analysis Country-year.

Rationale for the chosen scope:

The Latin America and Caribbean (LAC) region presents a particularly interesting case for
studying the environmental impacts of foreign direct investment because:
1. It has experienced steady inflows of FDI since the early 2000s, particularly in sectors
such as manufacturing, resource extraction, energy, and infrastructure.
2. The region is highly vulnerable to climate change impacts and simultaneously under
pressure to achieve economic growth.
3. Many LAC economies depend on natural resource-based industries, which are often
carbon-intensive.
4. The period 2010–2023 captures both pre- and post-pandemic economic trends,
including the recovery period after 2021, providing meaningful temporal variation.
This well-defined scope ensures the dataset is large enough to conduct reliable statistical
analyses, yet focused enough to reflect regional dynamics accurately.

2.3 Variable Description

The study uses one dependent variable (CO₂ emissions), one key explanatory variable (FDI
inflows), and five control variables commonly used in environmental economics literature to
isolate the relationship of interest.
Variable Description and Rationale
The variables were selected based on their theoretical relevance to the research question
concerning the relationship between Foreign Direct Investment (FDI) and CO₂ emissions,
drawing from environmental economics and the Pollution Haven Hypothesis literature.

Variable Rationale for

WDI Code Description Unit
Name Inclusion

CO2 [Link] CO2 Metric Dependent

Emissions emissions tons Variable.
(metric tons Directly measures
Variable Rationale for
WDI Code Description Unit
Name Inclusion

the environmental
per capita) outcome of
interest.

Main
Explanatory
Foreign direct Variable.
investment, Captures the scale
FDI Inflow [Link] % of GDP
net inflows (% of foreign capital
of GDP) inflow relative to
the economy's
size.

Control
Variable. A core
variable in the
Environmental
GDP per
Kuznets Curve
GDP per capita
[Link] US$ (EKC)
capita (constant 2015
hypothesis; level
US$)
of economic
development is a
key driver of
emissions.

Control
Variable.
Urbanization is
Urban
associated with
Urban population (%
[Link] % concentrated
Population of total
energy use,
population)
transportation
demands, and
industrial activity.

Industry [Link] Industry % of GDP Control Variable.

Value (including The industrial
Added construction), sector is typically
Variable Rationale for
WDI Code Description Unit
Name Inclusion

energy-intensive
value added and a primary
(% of GDP) source of CO2
emissions.

Control Variable.
Directly measures
Energy use the scale of
Energy (kg of oil Kg of oil energy
[Link]
Use equivalent per equivalent consumption,
capita) which is the
primary source of
CO2 emissions.

Trade Trade (% of
[Link]
Openness GDP)

All variables are annual, country-level indicators, and logarithmic transformations will be
considered in subsequent regression analysis to address skewness and interpret elasticities more
effectively.
2.4 Summary Statistics

Variable Mean Std. Dev. Min Max Obs

CO₂ Emissions (kt) 88,764.2 125,340.5 1,230.0 492,000.0 140

FDI Inflows (% of
4.2 2.5 0.1 12.8 140
GDP)

GDP per capita

9,860.7 6,540.2 2,100.5 25,300.0 140
(USD)

Trade Openness
54.7 25.3 17.3 128.6 140
(% of GDP)
Variable Mean Std. Dev. Min Max Obs

Industry Share (%
26.4 8.1 12.5 41.7 140
of GDP)

Energy Use (kg oil

1,610.2 856.4 732.3 3,420.6 140
eq.)

Population 46,200,00
38,500,000 3,200,000 215,000,000 140
(people) 0

The wide standard deviations in emissions, energy use, and population illustrate the
heterogeneity of countries in the sample, which is important for regression analysis and
heterogeneity testing in later sections.
2.5 Data Visualization
a) Scatter Plot – FDI Inflows vs CO₂ Emissions
Interpretation:
The scatter plot suggests a moderately positive relationship between FDI inflows and CO₂
emissions across the 10 countries from 2010 to 2023. Countries with higher FDI inflows (e.g.,
Brazil, Mexico) tend to have higher emission levels. However, the dispersion indicates that other
factors such as energy structure, regulatory quality, and economic scale may also shape this
relationship.
b) Trend Plot – Regional Trends in FDI and CO₂ Emissions
4
Interpretation:
Between 2010 and 2019, both FDI inflows and CO₂ emissions followed an upward trajectory,
reflecting the region’s economic growth and increasing integration into global production chains.
During 2020–2021, both variables declined due to the global economic slowdown caused by the
COVID-19 pandemic, followed by a partial recovery in 2022–2023. This co-movement provides
preliminary evidence supporting the hypothesis that FDI is associated with emission trends,
although causality requires formal econometric testing.
2.6 Correlation Analysis
Variables CO₂ FDI GDPpc Trade Industry Energy Pop
CO₂ 1.000 0.325 0.615 0.271 0.511 0.714 0.889
emissions
FDI 0.325 1.000 0.318 0.492 0.203 0.311 0.267
inflows
(% GDP)
GDP per 0.615 0.318 1.000 0.523 0.433 0.601 0.571
capita
Trade 0.271 0.492 0.523 1.000 0.361 0.214 0.300
openness
Industry 0.511 0.203 0.433 0.361 1.000 0.510 0.498
share
 Energy 0.714 0.311 0.601 0.214 0.510 1.000 0.620
use
Population 0.889 0.267 0.571 0.300 0.498 0.620 1.000

Interpretation:
 Thestrong positive correlation between CO₂ emissions and population (0.889) confirms the
expected scale effect: larger countries emit more CO₂.
A moderate positive correlation between FDI inflows and CO₂ emissions (0.325) suggests
that higher levels of foreign investment may be associated with increased emissions.
 Energyuse (0.714) and GDP per capita (0.615) are also strongly correlated with CO₂
emissions, highlighting the role of energy intensity and economic scale.
 Trade openness shows a weaker but still positive correlation (0.271), indicating that global
integration may indirectly influence emissions through trade and investment flows.
 No pairwise correlations exceed 0.9 except for CO₂ and population, which may require
transformation or robustness checks to avoid multicollinearity issues in the regression
model.
2.7 Discussion of Patterns and Preliminary Insights
The exploratory data analysis provides valuable insights that help formulate expectations for the
upcoming econometric modeling:
1. FDI as a driver of emissions:
The upward trend and positive correlation between FDI and CO₂ emissions are consistent
with the “pollution haven” and “scale effect” hypotheses in environmental economics.
FDI inflows often target sectors such as manufacturing, mining, and energy, which can
increase emissions if not accompanied by green technology transfer or strict regulation.
2. Energy consumption and industrial structure:
High correlations between energy use, industrial value added, and CO₂ emissions
confirm the central role of the energy-industrial nexus in the region. The heavy reliance
on fossil fuels in power generation and production processes amplifies the environmental
impact of investment.
3. Population and scale effects:
Population stands out as the strongest predictor of emissions in the correlation matrix,
reflecting the importance of controlling for scale effects in any empirical model. Larger
countries naturally emit more due to higher energy demand, transportation use, and
industrial output.
4. Trade and globalization linkages:
Trade openness shows moderate positive correlations with both FDI and CO₂, indicating
that globalization may have indirect environmental consequences. Trade expansion often
accompanies investment flows and industrial activities that contribute to emissions.
 5. Heterogeneity across countries:
The scatter plot indicates significant variation across countries, which may be explained
by differences in institutional quality, environmental regulations, energy mix, and
technological capacity. This justifies the inclusion of interaction terms (e.g., FDI ×
institutional quality) in later hypothesis testing.
6. Policy relevance:
These results highlight the need for policy frameworks that balance investment
promotion with environmental protection. Governments may consider environmental
regulations, green investment incentives, and renewable energy promotion to decouple
economic growth from emissions.
Concluding Remark
This section establishes the empirical foundation for the study. The dataset provides sufficient
variation across countries and over time to allow for meaningful econometric analysis in the
subsequent section. The patterns observed in exploratory analysis are consistent with existing
literature on environmental impacts of investment flows and point toward a positive association
between FDI inflows and CO₂ emissions in Latin America and the Caribbean.
However, these findings remain descriptive and do not imply causality. Further regression
analysis in Question 3 will specify the model formally, test hypotheses, and assess the magnitude
and significance of the effects, including potential moderating roles of institutional quality.
[Link]