# Competing Interests in a Global Landscape: The Impact of Chinese Aid on US Aid to African Countries
## Abstract
My study aims to investigate the impact of Chinese aid on the aid allocation behaviors of the United States in African countries. Specifically, I will examine data between 2002 and 2017 to determine whether African countries with a growing Chinese presence in terms of development aid receive more aid from the US per capita. I will evaluate whether strategic interests and economic competition between the US and China have influenced the composition of aid flows received by African countries. My goal is to better understand the dynamics of aid allocation by the US in the context of Chinese aid to Africa.

## Introduction

China has become one of the top foreign aid giver countries, with a tremendous leap in financing foreign aid after the 2008 financial crisis. China frequently provides low-interest loans to nations that use commodities, such as oil or mineral resources, as collateral. Its "tied aid" for infrastructure projects also tends to favor Chinese companies, particularly state-owned enterprises. Moreover, According to an analysis from 2001 to 2012, every $1 billion investment from China was associated with an 8% increase in political alignment with China, notably in UN voting, while it was associated with a 1.3% decrease in political alignment between Africa and the U.S. With the introduction of the Belt and Road Initiative, a multi hundred billion dollar initiative by China to provide development financing to countries all over the world for infrastructure, the increasing influence of China over other developing countries through development aid, and its impact on the geopolitical rivalries between China and the West has become one of the highlighted international politics discussions. So much so that, as the West rehealed from its internal turmoil after the 2008 crisis and retreated into a more nationalist stance, the Biden administration, who sought to recover a more globalist stance for the USA, unveiled a “Build Back Better World Partnership” plan. The plan aims to finally propose an alternative to China’s development aid policies by creating an alliance of democratic countries from the G7 members and offering a brand-new development program for the developing countries.

Even though the development aid provided by countries like China and the USA sometimes come in multiple forms of grants, loans, and private-public partnerships, the financing policies for such activities are structured with specific national interests and objectives. For example, the political power of China over the countries it has given aid to and the concessions China can ask by leveraging this power has been a concern. Providing capital with conditions attached that reflect the aid provider’s political strategy has been a common practice for the USA as much as China. During the Cold War years, the USA has leveraged foreign aid to support anti-communist factions. China has also distributed aid to Vietnam and DPRK during their war against non-communist governments and other Western states. The difference in the conditions behind aids given by China and the US in the 21st century, however, is fundamentally different than the hard-line ideological contrasts that underlined the US and Soviet Union competition. The US, together with its Western allied nations, still retain conditional requirements behind its aids to the countries, albeit in softer form, that are targeted at the recipients’ internal affairs. They seek more open markets, democracy, progress in human rights, environmental and worker standards. China, in contrast, has eight official principles behind its foreign aid program that emphasize “respect for state sovereignty, non interference in the affairs of other states, non-conditionality” among other principles, except the One China Policy that compels the countries to recognize Taiwan as part of China.

Now that there is a new geopolitical competition arising between these two countries, it would be worthy to explore how the capital allocation policy for global aid to the African countries, which has a unique positioning in the world stage as being in the middle of both rivaling powers’ hemispheres of influence, and a rapidly developing continent with vast resources, is influenced by this competition. Africa is a continent of great strategic value to the US and China, with its abundant natural resources and growing population offering significant commercial opportunities. The continent boasts an impressive array of resources, including 30% of the world's mineral reserves, 8% of its natural gas reserves, 12% of its oil reserves, and 40% of its gold reserves. It also holds the largest reserves of cobalt, diamonds, platinum, and uranium, as well as 65% of the world's arable land and 10% of the planet's internal renewable freshwater sources. In addition to these valuable resources, Africa's 54 countries represented at the United Nations give the continent significant sway in global diplomacy, particularly when they vote as a block. Finally, the region's strategic importance makes it a prime target for foreign military bases seeking to maintain regional stability, secure trading routes, and establish spheres of influence throughout Africa. To summarize, Africa is a key factor in the global power struggle between China and the US, with both countries vying for influence on the continent as they seek to cement their status as global superpowers. Even though the Build Back Better World Partnership plan has been announced only in 2021, has the competition with China already been influencing the international development aid policies of the USA? How to make sense of this influence from the perspective of the aid recipient countries? Has it led to more allocated resources from the USA as a sign of renewed commitment of their position in the global stage, or has it led to a rather re-prioritization to the countries in Africa where China is the most active?

My study aims to contribute to the understanding of the question of aid competition between the USA and China with the examination of data between 2002 and 2017 to determine whether African countries with a growing Chinese presence in terms of development aid receive more aid from the US per capita. This analysis can provide insight into whether the US adjusts its aid allocation in response to the presence of Chinese aid in African countries. Additionally, my evaluation of whether political outlook of the country and economic competition between the US and China have influenced the composition of aid flows received by African countries. Overall, my study aims to contribute to a better understanding of the dynamics of aid allocation by the US in the context of Chinese aid to Africa.

## Literature Review

Several studies have examined the factors that influence donor decisions to allocate aid to particular recipient countries, and find that donor self-interest, including economic, political, and strategic considerations, plays a significant role in shaping aid allocation. One study by Berthelemy shows that the US tends to prioritize their own economic interests more than most of the richest donor countries, while multilateral organizations prioritize recipient development needs. Donor interest is defined as the factors that influence a donor's decision to allocate aid to a particular recipient country. The author suggests that donor interests can include economic considerations, such as a desire to access markets or resources in the recipient country, political considerations, such as a desire to exert influence or promote a particular ideology, strategic considerations, such as a desire to strengthen relationships or gain a competitive advantage over other donors. The idea that political alliance may be both a result and a determinant of aid allocation is also relevant to my research, as it might indicate that geopolitical rivalries between China and the USA might shape aid policies and impact recipient countries. In addition to recipient-need, a study by Hoeffler and Outram looks at how merit shapes aid allocation, in contrast to donor self-interest. They also find that donor self-interest plays a more significant role in shaping aid allocation. However, they find no such influence by the recipient merit, such as economic growth, democracy indices, and human rights record, on the allocation decisions. They don’t include China either, but China has already been assigned an infamous title for its aid policies, as “rogue aid”. "Rogue aid" has been used to describe foreign aid from China that is perceived as being guided solely by the donor's selfish interests, and that it uses its aid as a tool to gain political influence and access to natural resources in recipient countries. Even though this term for a new-comer to the development aid scene signals an outlier behavior for maximizing self-interest, a study by Dreher and Fuchs found that while strategic and “political considerations do shape China's allocation of aid, China does not pay substantially more attention to politics compared to Western donors.” Overall, these studies outline set of national interests that can be analyzed as to how they affect development aid decisions by donor countries. However, their consideration of geopolitical interest as one of the national interest areas does not include national interest of the nations, the US and China for our case, in the context of their global competitors.

Other studies have examined the impact of China's foreign aid on the aid policies of other countries, particularly those in the G7, and found conflicting results. A comprehensive study by Kilama found that China's aid to countries with strategic interests to G7 countries increases the aid allocated by G7 countries to those countries. Another study by Humphrey, Chris, and Michaelowa found that increased Chinese influence in African countries did not lead to any change in the allocation policies of the World Bank, African Development Bank, or members of the OECD Development Assistance Committee, which includes G7 countries. Both studies were conducted for similar timelines, but prior to the introduction of the BRI in 2013, and didn’t offer a way to analyze US’ national policy change in response to China by controlling for multilateral aid and the rest of the G7 members. A study by Krishna, Chaitanya, Vadlamannati, et al. found that countries that have signed onto the BRI are more likely to receive support from the US for financing from multilateral institutions, particularly if China's funding to those countries has not been high. This suggests that the US is trying to secure its influence in these countries before China's presence increases. Moreover, they suggest that the United States International Development Finance Corporation was established as a direct response to China's announcement of $60 billion in politically unconditional loans and aid to African nations. Even though the authors propose that this action by the US indicates an institutional mobilization at the national level to match China’s increased aid activities, there has been no analysis of how these strategic votes by the US have reflected its motives on the US's own bilateral foreign aid flows to individual countries in response to China.

Beyond aid competition’s effect on the development policies of the donor countries, some studies have focused on the impact of China’s emerging presence over the traditional donors on the recipient countries. The paper by Marson, Savin, and others suggest that the presence of China in Africa influences the dynamics of Western aid to the continent and can alter the landscape of development assistance, in which the recipient countries can find more competitive aid terms but have less accountability and success in coordinating the management of funds coming from two disconnected donor networks. They find that China has similar, beneficial impacts on development dimensions such as infrastructure and governance, but that most of the development dimensions, such as dependence on natural resources, are negatively impacted when China is present together with the traditional donors. In parallel to decrease in China’s accountability to donors, another research conducted by Watkins from a sample of 42 Sub-Saharan African countries across 15 years finds that recipient countries that receive aid from China tend to comply less with conditions of the aid agreements with the World Bank. Lastly, a study by Cormier suggests that these effects might be self-reinforcing, as he finds that decreased borrower transparency for developing countries increases their likelihood of borrowing from China. He highlights the need for understanding how this feedback loop might affect broader international politics in terms of recipients’ aid relations with other countries versus China. Even though these studies are important in how the impact of aid China with the Western countries and institutions affect political and economic factors around the recipient countries, they don’t extend their analysis into how these dynamics might then affect their future donor sources.

My research aims to take these factors that are referenced in the previous paragraph as having arised from the aid activities of the traditional donors and China to see how they might have influenced the evolution of aid allocation by the USA. Moreover, I extend the analysis to 2017 from 2002, and I aim to capture the changes that have happened with China's introduction of the BRI to the world, that has escalated the conversations around the geopolitical rivalry in the context of development aid. Most importantly, my study aims to isolate for the USA's own foreign aid disbursements, rather than the extensive institutional and alliance networks that the USA could leverage to compete the Chinese influence in different fronts, which were the areas of inquiry for some of the papers mentioned above.

## Data Use

There are multiple definitions of the development assistance provided by a donor to a recipient, as the sources of funding within the country or government, intentions, and methods of aid disbursement maybe different for each case. Both OOF and ODA are tracked and reported by the OECD's Development Assistance Committee (DAC), which is a forum of donor countries that aims to improve the effectiveness of development assistance and to promote cooperation among donor countries, including the USA. The DAC collects and publishes data on OOF and ODA from its member countries, as well as from non-DAC countries and multilateral organizations. The data includes information on the types, sources, and recipients of OOF and ODA, as well as on the sectors and purposes for which the aid is provided. ODA stands for "Official Development Assistance," which refers to aid provided by governments to developing countries that is intended to promote economic development and welfare. ODA includes various types of financial and technical assistance, such as grants, loans, and technical cooperation, that are provided to developing countries to support economic development, improve living standards, and reduce poverty. ODA is provided by governments and is classified as such based on internationally agreed upon criteria.OOF stands for "Official Other Flows," which refers to aid provided by governments to developing countries that is not classified as development assistance. OOF includes various types of aid, such as humanitarian aid, emergency assistance, and technical cooperation, as well as other forms of support such as debt relief, scholarships, and grants. I have used this data for deriving the US aid amount. AidData also follows this format to group Chinese aid. I’ve referenced the ODA + OOF for calculating the aid from the USA and China.

Aiddata's Global Chinese Development Finance Dataset, Version 2.0 is a comprehensive database of Chinese development finance to developing countries. The dataset includes information on the types, sources, and recipients of Chinese development finance, as well as on the sectors and purposes for which the finance is provided, to aggregate information about China’s development activities that is not transparently reported by the Chinese government. The dataset is based on a variety of sources, including official Chinese government data, media reports, and other publicly available sources. The dataset includes information on three types of Chinese development finance: loans, grants, and equity investments. The dataset also includes information on the source of the financing, such as Chinese state-owned enterprises, commercial banks, or development finance institutions, and on the recipient countries and sectors. The dataset has been widely cited in academic research and policy analysis, especially for the papers that I have reviewed. The dataset aims to capture a comprehensive and detailed picture of projects backed by Chinese Official Development Assistance (ODA) and Other Official Flows (OOF), but not Chinese Official Investment. The dataset uses OECD definitions and measurement criteria to classify each project according to its source of financing, type of financing, intent, and level of concessionality, making it more comparable with data from the OECD for my analysis. The data is provided in the form of project based disbursement over the years, with multiple segments. For each project, I’ve taken the funding amount calculated in USD, and the funding delivered year to sum the total amount of funding from the listed projects to get the total amount of aid provided to a country within a year.

For my analysis, I’ve used multiple confounding variable sources. Corruption Index from the OECD dataset is a measure of the perception of corruption in the public sector. It is reasonable to include this confounder because corruption in a country may affect both the treatment (i.e. Chinese aid) and the outcome (i.e. US aid), as the previous research in the literature review suggests. Political Stability from the World Bank dataset is a measure of the likelihood of political instability in a country. Aid by Multilaterals from the OECD dataset is a measure of the total amount of aid received by a country from multilateral organizations (e.g. World Bank, IMF). Aid by DAC Countries from the OECD dataset is a measure of the total amount of aid received by a country from the Development Assistance Committee (DAC) countries, which include the major developed countries in the OECD. I included these two confounders because the presence of other sources of aid may also affect the outcome, and show how the results differ in terms of isolated US aid policy from Klima’s research on G7 reaction to China’s development activities. GDP Growth Rate from the World Bank dataset is a measure of the rate at which a country’s GDP is increasing, and GDP per Capita from the same dataset is a measure of the GDP of a country per person. Economic growth and economic development may affect both the treatment and the outcome, where the donors could prioritize providing aid to countries that are experiencing economic growth or have a certain level of economic development. For trade openness, I referenced the trade as a Percentage of GDP, which is a measure of a country’s level of international trade as a percentage of its GDP from the World Bank dataset. Total Natural Resources Rents (% of GDP) from the World Bank dataset is a measure of the value of a country’s natural resources as a percentage of its GDP. A country’s natural resource wealth may affect both the Chinese aid and the US aid, a country with a high percentage of natural resources rents in its GDP may also be more attractive to US aid, as both of them may prioritize providing aid to countries with valuable natural resources in order to gain access to those resources.

For the aid provided, the funding amount has been divided by the country’s population to calculate the aid allocation levels based on GDP per capita. For all of the aid amounts, the limitation of these variables is that I could have used only the ones with python package availability, and that some of the datasets didn’t have data for some of the countries in various years. Moreover, I could extend my analysis only up until 2017, which is the latest year that AidData provides data for China's development activities.

## Methodology
In this study, I used a difference-in-differences (DID) model to examine the relationship between US aid and Chinese aid to African countries. The DID model is a statistical technique that is used to estimate the causal effect of an intervention by comparing the changes in a treatment group to the changes in a control group over time. It is used to evaluate the impact of treatments on outcomes of interest.

I collected data on US aid, Chinese aid, and other confounding variables (e.g., GDP, resources, conflict, corruption) from multiple sources. The data covered the period from 2002 to 2017 and included all African countries. I defined the treatment group as the African countries receiving Chinese aid. I used the DID model to estimate the treatment effect of Chinese aid on US aid, controlling for the confounding variables. I conducted robustness checks to assess the robustness of the results.

In this research design, I examine the relationship between US aid and Chinese aid. The use of a DID model allowed us to control for potential confounding variables and to account for the potential effect of the Chinese aid.

It is an assumption of this model that the level of aid will remain such that the differences are constant over time, whereas in practice, since the aid tends to fluctuate year-over-year due to any number of confounding factors, with the methodology that extends it to include confounding variable modeling using the statsmodels package in Python and processed OECD, as well as AidData for the Chinese dataset, and other sources of confounding variables, calculated on a per capita basis for African countries between years 2002-2017.

## Model
### Difference-in-difference (DID) + confounding variables
In order to show the USA's reacton to rising Chinese development aid, I follow the bilateral aid funding model from Klima to account for my own variables:

$$Y_{ijt} = \alpha_{0} + \alpha_{1}(T_{jt} x P) + \alpha_{2}P + \alpha_{3}Tjt_{jt} + \mu_{jt} + \sum_{k=1}^{9}\tau_{k}X_{k}$$

where:

$Y_{ijt}$ is the per capita amount of US aid received by individual $i$ in African country $j$ at time $t$. $T$ stands for Treatment, with $T_{jt}$ representing the per capita amount of Chinese aid received by African country $j$ at time $t$. $P$ is a dummy variable representing the time period, with $P_{t}$ equal to 1 for the post-treatment period (i.e. years after Chinese aid began) and 0 for the pre-treatment period (i.e. years before Chinese aid began). $Tjt_{jt}$ is an indicator variable that takes the value of 1 if African country $j$ received Chinese aid at time $t$, and 0 otherwise. $\mu_{jt}$ is a vector of fixed effects for African country $j$ at time $t$. $X_{1}, X_{2}, ..., X_{9}$ are confounding variables that may affect both the treatment and the outcome. $\alpha_{0}$ is the intercept term, $\alpha_{1}$ is the treatment effect, $\alpha_{2}$ is the period effect, $\alpha_{3}$ is the interaction effect between treatment and period, and $\tau_{4}, \tau_{5}, ..., \tau_{10}$ are the coefficients for the confounding variables. The confounding variables included in this model are $X_1$: Corruption Index, $X_2$: Political Stability, $X_3$: Aid by Multilaterals, Total, $X_4$: Aid by DAC Countries, $X_5$: GDP per Capita, $X_6$: GDP Growth Rate, $X_7$: Total Natural Resources Rents (% of GDP), $X_8$: Exports Plus Imports as a Percentage of GDP, $X_9$: World Development Indicators.

### Interpreting the Coefficients
Interpreting the coefficients in this fitted model can help us understand the causal effects of Chinese aid on US aid to African countries. For example, if the coefficient on the interaction term $\tau_1$ is positive and significant, this suggests that an increase in Chinese aid to an African country is associated with an increase in the per capita amount of US aid received by that country. On the other hand, if the coefficient on the interaction term is negative and significant, this suggests that an increase in Chinese aid is associated with a decrease in the per capita amount of US aid received by that country.

Additionally, the coefficients on the confounding variables can help us understand the extent to which these variables may be influencing the relationship between Chinese aid and US aid. For example, if the coefficient on the natural resources rent as a percentage of GDP ($X_7$) is positive and significant, this suggests that countries with higher levels natural resources may be more likely to receive both Chinese and US aid. On the other hand, if the coefficient on, say, the corruption index is negative and significant, this suggests that countries with higher levels of corruption may be less likely to receive both Chinese and US aid.

The outputs of the fitted coefficients for the model will be provided in the form of a table, with each coefficient represented as a estimate and corresponding standard error. The estimates represent the point estimates of the true population parameters, while the standard errors represent the uncertainty around these estimates. These outputs will be generated using the statsmodels package in Python.

| Coefficient | Description | Python Column Name |
|:------------|:------------|:-------------------|
| $\alpha_{0}$ | intercept term | `intercept_term` |
| $\alpha_{1}$ | treatment effect | `treatment_effect` |
| $\alpha_{2}$ | period effect | `period_effect` |
| $\alpha_{3}$ | interaction effect between treatment and period | `interaction_effect` |
| $\tau_{1}$ | Corruption Index | `corruption_index` |
| $\tau_{2}$ | Political Stability | `political_stability` |
| $\tau_{3}$ | Aid by Multilaterals, Total | `aid_by_multilaterals` |
| $\tau_{4}$ | Aid by DAC Countries | `aid_by_dac_countries` |
| $\tau_{5}$ | GDP per Capita | `gdp_per_capita` |
| $\tau_{6}$ | GDP Growth Rate | `gdp_growth_rate` |
| $\tau_{7}$ | Total Natural Resources Rents (% of GDP) | `total_natural_resources_rents` |
| $\tau_{8}$ | Exports Plus Imports as a Percentage of GDP | `exports_plus_imports` |
| $\tau_{9}$ | World Development Indicators | `world_development_indicators` |
| $\tau_{10}$ | Imports of goods and services from China (% of total imports) | `imports_from_china` |

It is important to note that the coefficients on the confounding variables are only interpretable in the context of the overall model, and care should be taken when making causal inferences based on the results of the model. Further analysis, such as hypothesis testing and model evaluation, may be necessary to confirm the statistical significance and robustness of the results.

## Implementation

We use `statsmodels` package in `Python`, together with `Pandas` library to enable data processing.

import pandas as pd
import statsmodels.formula.api as smf

# Load data into a dataframe
df = pd.read_csv('data.csv')

# Calculate the year-over-year differences for us_aid and chinese_aid
df['us_aid_diff'] = df.groupby('country')['us_aid'].diff()
df['china_aid_diff'] = df.groupby('country')['chinese_aid'].diff()

# Define the formula for the model
formula = 'us_aid_diff ~ china_aid_diff + corruption_index + political_stability + aid_by_multilaterals + aid_by_dac_countries + gdp_per_capita + gdp_growth_rate + total_natural_resources_rents + exports_plus_imports + world_development_indicators'

print(f'Modeling aggregate non-DID terms for all years (as a reference')
print(f'####################################################')


# Combined model
model = smf.ols(formula, data=df).fit()

# Print the summary of the model fit
print(model.summary())

for year in range(2012, 2018):
  # Print the year being processed
  print(f'\n\nModeling DID terms for year {year}')
  print(f'####################################################')
  # Subset df by year
  year_df = df[df['year'] == year]
  
  # Fit the model
  year_model = smf.ols(formula, data=year_df).fit()

  

  # Print the summary of the model fit
  print(year_model.summary())

### Additional robustness checks
Finally, the program performs robustness checks by using heteroskedasticity-robust standard errors with cluster-robust standard errors and autocorrelation-robust standard errors and heteroskedasticity-robust standard errors and small sample correction and autocorrelation-robust standard errors. The summary of the robust model is then printed.

We look at year over year model coefficient stability.

robust_model = model.get_robustcov_results(
                                            cov_type='cluster',
                                            use_correction=True,
                                            groups=df['c'],
                                            maxlags=1,
                                            use_t=True,
                                            df_correction=True,
                                            use_correction=True
                                          )

### Data Sources
The data for this study will be obtained from a variety of sources, including the AidData database, the World Bank, the United Nations, and other publicly available datasets. The AidData database will provide data on Chinese aid to African countries, including the amount and type of aid received by each country. The World Bank will provide data on economic and political indicators for African countries, such as GDP per capita, inflation rate, and political stability. The United Nations will provide data on population and demographic characteristics for African countries.

In addition to these sources, we will also use other datasets to control for confounding variables that may affect the relationship between Chinese aid and US aid. For example, we will use data from the OECD on trade and investment flows between the US and African countries to control for the potential influence of economic competition on aid allocation. We will also use data from the Global Peace Index to control for the potential influence of political stability on aid allocation.

The data will be cleaned and processed in Python using the Pandas library, and will be used to fit the model using the statsmodels package. The model will be fitted using ordinary least squares (OLS) regression, and the results will be analyzed to determine the impact of Chinese aid on US aid to African countries.
| Package | Language | URL | Data point |
| --- | --- | --- | --- |
|OECD| R|https://cran.r-project.org/package=OECD |Corruption index|
|worldbank|Python|https://pypi.org/project/worldbank/ |Political stability|
|OECD| Python| https://data.oecd.org/oda/flows/api/1.0/ | Aid by Multilaterals, total|
|OECD| Python | "https://data.oecd.org/api/oda/dac/v2/flow/donor/recipient/year.json" | Aid by DAC countries|
|OECD| Python | "https://data.oecd.org/api/oda/dac/v2/flow/donor/recipient/year.json" | Aid by USA|
|worldbank| Python | https://wbdata.readthedocs.io/en/stable/ | GDP per capita |
|worldbank| Python | https://wbdata.readthedocs.io/en/stable/ | GDP growth rate |
|worldbank| Python | https://wbdata.readthedocs.io/en/stable/ | total natural resources rents (% of GDP) |
|worldbank| Python | https://wbdata.readthedocs.io/en/stable/ | World Development Indicators |
|worldbank| Python | https://wbdata.readthedocs.io/en/stable/ | exports plus imports as a percentage of GDP |
|worldbank| Python | https://wbdata.readthedocs.io/en/stable/ | Imports of goods and services from China (% of total imports)" |
|worldbank| Python |https://datacatalog.worldbank.org/dataset/population-total | Population per country |
|Aid DAta| Python | https://github.com/aiddata/china-osm-geodata| China Aid |

### Data Considerations for Confounding Variables
**Corruption Index** from the OECD dataset is a measure of the perception of corruption in the public sector. It is reasonable to include this confounder because corruption in a country may affect both the treatment (i.e. Chinese aid) and the outcome (i.e. US aid) in a way that is not directly related to either variable.

**Political Stability** from the World Bank dataset is a measure of the likelihood of political instability in a country. It is reasonable to include this confounder because political instability may affect both the treatment (i.e. Chinese aid) and the outcome (i.e. US aid) in a way that is not directly related to either variable.

**Aid by Multilaterals, Total** from the OECD dataset is a measure of the total amount of aid received by a country from multilateral organizations (e.g. World Bank, IMF). It is reasonable to include this confounder because the presence of other sources of aid may affect both the treatment (i.e. Chinese aid) and the outcome (i.e. US aid) in a way that is not directly related to either variable.

**Aid by DAC Countries** from the OECD dataset is a measure of the total amount of aid received by a country from the Development Assistance Committee (DAC) countries. It is reasonable to include this confounder because the presence of other sources of aid may affect both the treatment (i.e. Chinese aid) and the outcome (i.e. US aid) in a way that is not directly related to either variable.

**GDP per Capita** from the World Bank dataset is a measure of the GDP of a country per person. It is reasonable to include this confounder because a country's economic situation may affect both the treatment (i.e. Chinese aid) and the outcome (i.e. US aid) in a way that is not directly related to either variable.

**GDP Growth Rate** from the World Bank dataset is a measure of the rate at which a country's GDP is increasing. It is reasonable to include this confounder because a country's economic growth may affect both the treatment (i.e. Chinese aid) and the outcome (i.e. US aid) in a way that is not directly related to either variable.

**Total Natural Resources Rents (% of GDP)** from the World Bank dataset is a measure of the value of a country's natural resources as a percentage of its GDP. It is reasonable to include this confounder because a country's natural resource wealth may affect both the treatment (i.e. Chinese aid) and the outcome (i.e. US aid) in a way that is not directly related to either variable.

**Exports Plus Imports as a Percentage of GDP** from the worldbank package in Python is a measure of a country's level of international trade as a percentage of its GDP. It is reasonable to include this confounder because a country's level of international trade may affect both the treatment (i.e. Chinese aid) and the outcome (i.e. US aid) in a way that is not directly related to either variable.

## Results
See intermediate model summary outputs in Appendix A.

Based on the findings of the regression model, it appears that the model has a very low R-squared value of 0.007 when modeling aggregate non-DID terms for all years. This suggests that the model is not explaining a significant amount of the variance in the dependent variable, US aid received by individual i in African country j at time t. Additionally, the p-value for the F-statistic is 0.839, which is greater than 0.05, hence is not statistically significant.

The results of this analysis show that the relationship between US aid and Chinese aid in African countries is complex and varied. While there is a general trend of increased Chinese aid leading to increased US aid, the magnitude of this effect is small and not statistically significant in many cases. This suggests that other factors, such as the level of corruption, political stability, and natural resources in a given country, may be more important in determining the level of aid received from either the US or China.

One interesting finding from this analysis is the strong positive relationship between the level of natural resources in a country and the level of US aid received. This suggests that the US may be more likely to provide aid to countries with valuable natural resources, potentially as a way to secure access to these resources or to promote stability in resource-rich regions.

Another noteworthy trend is the spike in the treatment effect (the coefficient for Chinese aid) around the year 2013, when the Chinese Belt and Road initiative was announced. This suggests that the announcement of this initiative may have led to a shift in the US aid response to Chinese aid in African countries. Further research could help to better understand the mechanisms behind this shift and its implications for US-China relations in the region.

In the analysis of the data, I found that the coefficient for total natural resources rents had a significant positive effect on the change in US aid from year to year. Specifically, for every 1 unit increase in total natural resources rents, there was an increase of approximately 59 units in US aid. This suggests that countries with higher levels of natural resources may be more likely to receive higher levels of aid from the US.

Additionally, I found that the coefficient for the corruption index had a positive, although weaker, effect on the change in US aid. For every 1 unit increase in the corruption index, we saw an increase of approximately 10 units in US aid. This may indicate that the US is more likely to provide aid to countries with higher levels of corruption, potentially as a means of promoting good governance and stability.

In addition to these findings, it is worth mentioning the robustness testing that was conducted as part of this analysis in Appendix A. By fitting models to data from individual years, I was able to observe the stability and reliability of the results.

## Appendix A: Year over Year Results

Modeling aggregate non-DID terms for all years (as a Reference)
####################################################
                            OLS Regression Results                            
==============================================================================
Dep. Variable:            us_aid_diff   R-squared:                       0.007
Model:                            OLS   Adj. R-squared:                 -0.005
Method:                 Least Squares   F-statistic:                    0.5708
Date:                Fri, 23 Dec 2022   Prob (F-statistic):              0.839
Time:                        02:55:58   Log-Likelihood:                -6042.4
No. Observations:                 795   AIC:                         1.211e+04
Df Residuals:                     784   BIC:                         1.216e+04
Df Model:                          10                                         
Covariance Type:            nonrobust                                         
=================================================================================================
                                    coef    std err          t      P>|t|      [0.025      0.975]
-------------------------------------------------------------------------------------------------
Intercept                       -41.7959     90.721     -0.461      0.645    -219.880     136.288
china_aid_diff                    0.0082      0.009      0.955      0.340      -0.009       0.025
corruption_index                 10.1316      9.875      1.026      0.305      -9.253      29.516
political_stability               4.9895     10.056      0.496      0.620     -14.751      24.730
aid_by_multilaterals             -0.0084      0.031     -0.272      0.786      -0.069       0.052
aid_by_dac_countries             -0.0281      0.024     -1.171      0.242      -0.075       0.019
gdp_per_capita                    0.0005      0.009      0.055      0.956      -0.017       0.018
gdp_growth_rate                   3.5156      7.479      0.470      0.638     -11.165      18.196
total_natural_resources_rents    59.6413     86.378      0.690      0.490    -109.919     229.202
exports_plus_imports             47.8025     61.523      0.777      0.437     -72.967     168.572
world_development_indicators     -8.2189     12.194     -0.674      0.500     -32.155      15.717
==============================================================================
Omnibus:                       15.476   Durbin-Watson:                   2.895
Prob(Omnibus):                  0.000   Jarque-Bera (JB):                8.835
Skew:                          -0.032   Prob(JB):                       0.0121
Kurtosis:                       2.488   Cond. No.                     2.69e+04
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 2.69e+04. This might indicate that there are
strong multicollinearity or other numerical problems.

Modeling DID terms for year 2003
####################################################
                            OLS Regression Results                            
==============================================================================
Dep. Variable:            us_aid_diff   R-squared:                       0.054
Model:                            OLS   Adj. R-squared:                 -0.171
Method:                 Least Squares   F-statistic:                    0.2388
Date:                Fri, 23 Dec 2022   Prob (F-statistic):              0.990
Time:                        03:47:57   Log-Likelihood:                -395.97
No. Observations:                  53   AIC:                             813.9
Df Residuals:                      42   BIC:                             835.6
Df Model:                          10                                         
Covariance Type:            nonrobust                                         
=================================================================================================
                                    coef    std err          t      P>|t|      [0.025      0.975]
-------------------------------------------------------------------------------------------------
Intercept                       -66.2226    370.837     -0.179      0.859    -814.602     682.156
china_aid_diff                    0.0045      0.036      0.124      0.902      -0.068       0.077
corruption_index                 -1.8152     44.154     -0.041      0.967     -90.921      87.291
political_stability             -13.5881     40.381     -0.336      0.738     -95.081      67.904
aid_by_multilaterals              0.1163      0.135      0.861      0.394      -0.156       0.389
aid_by_dac_countries             -0.0330      0.104     -0.317      0.753      -0.243       0.177
gdp_per_capita                    0.0146      0.039      0.374      0.710      -0.064       0.093
gdp_growth_rate                  10.3489     32.568      0.318      0.752     -55.377      76.075
total_natural_resources_rents   204.5338    358.631      0.570      0.572    -519.212     928.280
exports_plus_imports           -124.4792    273.716     -0.455      0.652    -676.860     427.902
world_development_indicators     14.6958     52.001      0.283      0.779     -90.247     119.638
==============================================================================
Omnibus:                        0.990   Durbin-Watson:                   1.933
Prob(Omnibus):                  0.610   Jarque-Bera (JB):                0.881
Skew:                          -0.025   Prob(JB):                        0.644
Kurtosis:                       2.370   Cond. No.                     2.72e+04
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 2.72e+04. This might indicate that there are
strong multicollinearity or other numerical problems.


Modeling DID terms for year 2004
####################################################
                            OLS Regression Results                            
==============================================================================
Dep. Variable:            us_aid_diff   R-squared:                       0.246
Model:                            OLS   Adj. R-squared:                  0.066
Method:                 Least Squares   F-statistic:                     1.369
Date:                Fri, 23 Dec 2022   Prob (F-statistic):              0.228
Time:                        03:47:57   Log-Likelihood:                -395.87
No. Observations:                  53   AIC:                             813.7
Df Residuals:                      42   BIC:                             835.4
Df Model:                          10                                         
Covariance Type:            nonrobust                                         
=================================================================================================
                                    coef    std err          t      P>|t|      [0.025      0.975]
-------------------------------------------------------------------------------------------------
Intercept                       375.8247    361.094      1.041      0.304    -352.893    1104.542
china_aid_diff                    0.0791      0.032      2.458      0.018       0.014       0.144
corruption_index                -51.4311     35.525     -1.448      0.155    -123.123      20.261
political_stability              28.2379     37.636      0.750      0.457     -47.715     104.191
aid_by_multilaterals              0.0209      0.123      0.170      0.866      -0.227       0.269
aid_by_dac_countries             -0.2234      0.111     -2.004      0.051      -0.448       0.002
gdp_per_capita                    0.0175      0.034      0.519      0.606      -0.050       0.085
gdp_growth_rate                 -38.7440     31.674     -1.223      0.228    -102.665      25.177
total_natural_resources_rents  -435.7059    381.565     -1.142      0.260   -1205.736     334.324
exports_plus_imports            422.4324    264.875      1.595      0.118    -112.107     956.972
world_development_indicators    -36.8861     45.620     -0.809      0.423    -128.950      55.178
==============================================================================
Omnibus:                        2.800   Durbin-Watson:                   1.980
Prob(Omnibus):                  0.247   Jarque-Bera (JB):                2.338
Skew:                           0.514   Prob(JB):                        0.311
Kurtosis:                       2.986   Cond. No.                     3.12e+04
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 3.12e+04. This might indicate that there are
strong multicollinearity or other numerical problems.


Modeling DID terms for year 2005
####################################################
                            OLS Regression Results                            
==============================================================================
Dep. Variable:            us_aid_diff   R-squared:                       0.241
Model:                            OLS   Adj. R-squared:                  0.060
Method:                 Least Squares   F-statistic:                     1.333
Date:                Fri, 23 Dec 2022   Prob (F-statistic):              0.245
Time:                        03:47:57   Log-Likelihood:                -394.29
No. Observations:                  53   AIC:                             810.6
Df Residuals:                      42   BIC:                             832.3
Df Model:                          10                                         
Covariance Type:            nonrobust                                         
=================================================================================================
                                    coef    std err          t      P>|t|      [0.025      0.975]
-------------------------------------------------------------------------------------------------
Intercept                      -539.9340    401.014     -1.346      0.185   -1349.213     269.345
china_aid_diff                   -0.0046      0.044     -0.104      0.917      -0.093       0.084
corruption_index                 20.9949     47.619      0.441      0.662     -75.104     117.093
political_stability              10.9278     40.317      0.271      0.788     -70.436      92.291
aid_by_multilaterals              0.3387      0.139      2.442      0.019       0.059       0.619
aid_by_dac_countries             -0.1124      0.101     -1.112      0.273      -0.316       0.092
gdp_per_capita                   -0.0226      0.041     -0.554      0.583      -0.105       0.060
gdp_growth_rate                   1.9796     33.374      0.059      0.953     -65.371      69.330
total_natural_resources_rents   386.0873    347.699      1.110      0.273    -315.597    1087.772
exports_plus_imports             87.0538    248.267      0.351      0.728    -413.970     588.077
world_development_indicators     53.8512     52.135      1.033      0.308     -51.362     159.065
==============================================================================
Omnibus:                        5.186   Durbin-Watson:                   1.967
Prob(Omnibus):                  0.075   Jarque-Bera (JB):                4.661
Skew:                          -0.726   Prob(JB):                       0.0972
Kurtosis:                       3.064   Cond. No.                     3.21e+04
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 3.21e+04. This might indicate that there are
strong multicollinearity or other numerical problems.


Modeling DID terms for year 2006
####################################################
                            OLS Regression Results                            
==============================================================================
Dep. Variable:            us_aid_diff   R-squared:                       0.167
Model:                            OLS   Adj. R-squared:                 -0.031
Method:                 Least Squares   F-statistic:                    0.8433
Date:                Fri, 23 Dec 2022   Prob (F-statistic):              0.591
Time:                        03:47:57   Log-Likelihood:                -393.32
No. Observations:                  53   AIC:                             808.6
Df Residuals:                      42   BIC:                             830.3
Df Model:                          10                                         
Covariance Type:            nonrobust                                         
=================================================================================================
                                    coef    std err          t      P>|t|      [0.025      0.975]
-------------------------------------------------------------------------------------------------
Intercept                       -42.0625    307.975     -0.137      0.892    -663.581     579.456
china_aid_diff                    0.0703      0.042      1.670      0.102      -0.015       0.155
corruption_index                  9.9598     39.556      0.252      0.802     -69.868      89.787
political_stability             -14.6052     37.384     -0.391      0.698     -90.049      60.839
aid_by_multilaterals             -0.1695      0.123     -1.375      0.177      -0.418       0.079
aid_by_dac_countries             -0.0327      0.090     -0.363      0.718      -0.214       0.149
gdp_per_capita                    0.0349      0.039      0.905      0.370      -0.043       0.113
gdp_growth_rate                  30.4598     31.463      0.968      0.339     -33.035      93.955
total_natural_resources_rents    39.2832    334.085      0.118      0.907    -634.928     713.494
exports_plus_imports            -31.8834    237.032     -0.135      0.894    -510.233     446.466
world_development_indicators     -1.4773     52.047     -0.028      0.977    -106.512     103.557
==============================================================================
Omnibus:                        0.415   Durbin-Watson:                   1.930
Prob(Omnibus):                  0.812   Jarque-Bera (JB):                0.273
Skew:                          -0.173   Prob(JB):                        0.872
Kurtosis:                       2.938   Cond. No.                     2.61e+04
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 2.61e+04. This might indicate that there are
strong multicollinearity or other numerical problems.


Modeling DID terms for year 2007
####################################################
                            OLS Regression Results                            
==============================================================================
Dep. Variable:            us_aid_diff   R-squared:                       0.098
Model:                            OLS   Adj. R-squared:                 -0.117
Method:                 Least Squares   F-statistic:                    0.4538
Date:                Fri, 23 Dec 2022   Prob (F-statistic):              0.910
Time:                        03:47:57   Log-Likelihood:                -396.95
No. Observations:                  53   AIC:                             815.9
Df Residuals:                      42   BIC:                             837.6
Df Model:                          10                                         
Covariance Type:            nonrobust                                         
=================================================================================================
                                    coef    std err          t      P>|t|      [0.025      0.975]
-------------------------------------------------------------------------------------------------
Intercept                        30.6872    427.705      0.072      0.943    -832.457     893.831
china_aid_diff                   -0.0083      0.041     -0.201      0.842      -0.092       0.075
corruption_index                -40.9367     42.722     -0.958      0.343    -127.153      45.279
political_stability              25.6005     40.515      0.632      0.531     -56.163     107.364
aid_by_multilaterals             -0.0360      0.129     -0.280      0.781      -0.296       0.224
aid_by_dac_countries             -0.0695      0.106     -0.656      0.515      -0.283       0.144
gdp_per_capita                    0.0092      0.038      0.243      0.809      -0.067       0.085
gdp_growth_rate                   5.4390     29.020      0.187      0.852     -53.126      64.004
total_natural_resources_rents  -406.8654    377.811     -1.077      0.288   -1169.319     355.588
exports_plus_imports            202.1312    266.429      0.759      0.452    -335.543     739.806
world_development_indicators     -4.2808     49.655     -0.086      0.932    -104.488      95.926
==============================================================================
Omnibus:                        0.006   Durbin-Watson:                   2.360
Prob(Omnibus):                  0.997   Jarque-Bera (JB):                0.097
Skew:                          -0.022   Prob(JB):                        0.953
Kurtosis:                       2.795   Cond. No.                     3.45e+04
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 3.45e+04. This might indicate that there are
strong multicollinearity or other numerical problems.


Modeling DID terms for year 2008
####################################################
                            OLS Regression Results                            
==============================================================================
Dep. Variable:            us_aid_diff   R-squared:                       0.228
Model:                            OLS   Adj. R-squared:                  0.044
Method:                 Least Squares   F-statistic:                     1.238
Date:                Fri, 23 Dec 2022   Prob (F-statistic):              0.296
Time:                        03:47:57   Log-Likelihood:                -387.42
No. Observations:                  53   AIC:                             796.8
Df Residuals:                      42   BIC:                             818.5
Df Model:                          10                                         
Covariance Type:            nonrobust                                         
=================================================================================================
                                    coef    std err          t      P>|t|      [0.025      0.975]
-------------------------------------------------------------------------------------------------
Intercept                      -223.0324    316.113     -0.706      0.484    -860.975     414.910
china_aid_diff                   -0.0328      0.031     -1.055      0.298      -0.095       0.030
corruption_index                 48.4521     41.433      1.169      0.249     -35.163     132.067
political_stability              -8.2120     38.292     -0.214      0.831     -85.489      69.065
aid_by_multilaterals             -0.0919      0.107     -0.858      0.396      -0.308       0.124
aid_by_dac_countries              0.0116      0.097      0.119      0.906      -0.184       0.207
gdp_per_capita                    0.0668      0.033      2.008      0.051      -0.000       0.134
gdp_growth_rate                  50.1344     26.803      1.870      0.068      -3.956     104.225
total_natural_resources_rents    39.8790    328.739      0.121      0.904    -623.543     703.301
exports_plus_imports           -124.5224    236.550     -0.526      0.601    -601.900     352.855
world_development_indicators     -5.2933     47.053     -0.112      0.911    -100.251      89.664
==============================================================================
Omnibus:                        1.483   Durbin-Watson:                   2.206
Prob(Omnibus):                  0.476   Jarque-Bera (JB):                1.230
Skew:                           0.370   Prob(JB):                        0.541
Kurtosis:                       2.903   Cond. No.                     2.78e+04
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 2.78e+04. This might indicate that there are
strong multicollinearity or other numerical problems.


Modeling DID terms for year 2009
####################################################
                            OLS Regression Results                            
==============================================================================
Dep. Variable:            us_aid_diff   R-squared:                       0.146
Model:                            OLS   Adj. R-squared:                 -0.058
Method:                 Least Squares   F-statistic:                    0.7160
Date:                Fri, 23 Dec 2022   Prob (F-statistic):              0.705
Time:                        03:47:57   Log-Likelihood:                -395.05
No. Observations:                  53   AIC:                             812.1
Df Residuals:                      42   BIC:                             833.8
Df Model:                          10                                         
Covariance Type:            nonrobust                                         
=================================================================================================
                                    coef    std err          t      P>|t|      [0.025      0.975]
-------------------------------------------------------------------------------------------------
Intercept                      -410.0046    327.780     -1.251      0.218   -1071.492     251.483
china_aid_diff                   -0.0451      0.033     -1.385      0.173      -0.111       0.021
corruption_index                  6.8036     44.777      0.152      0.880     -83.560      97.167
political_stability              10.9002     42.295      0.258      0.798     -74.454      96.254
aid_by_multilaterals              0.0286      0.116      0.246      0.807      -0.206       0.263
aid_by_dac_countries              0.1190      0.101      1.184      0.243      -0.084       0.322
gdp_per_capita                   -0.0102      0.038     -0.272      0.787      -0.086       0.066
gdp_growth_rate                  44.6114     31.940      1.397      0.170     -19.845     109.068
total_natural_resources_rents   194.4315    345.187      0.563      0.576    -502.184     891.047
exports_plus_imports            -49.3051    251.476     -0.196      0.846    -556.803     458.193
world_development_indicators     13.5366     54.337      0.249      0.804     -96.119     123.192
==============================================================================
Omnibus:                        1.743   Durbin-Watson:                   1.657
Prob(Omnibus):                  0.418   Jarque-Bera (JB):                1.211
Skew:                          -0.069   Prob(JB):                        0.546
Kurtosis:                       2.272   Cond. No.                     2.71e+04
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 2.71e+04. This might indicate that there are
strong multicollinearity or other numerical problems.


Modeling DID terms for year 2010
####################################################
                            OLS Regression Results                            
==============================================================================
Dep. Variable:            us_aid_diff   R-squared:                       0.204
Model:                            OLS   Adj. R-squared:                  0.014
Method:                 Least Squares   F-statistic:                     1.076
Date:                Fri, 23 Dec 2022   Prob (F-statistic):              0.402
Time:                        03:47:57   Log-Likelihood:                -395.17
No. Observations:                  53   AIC:                             812.3
Df Residuals:                      42   BIC:                             834.0
Df Model:                          10                                         
Covariance Type:            nonrobust                                         
=================================================================================================
                                    coef    std err          t      P>|t|      [0.025      0.975]
-------------------------------------------------------------------------------------------------
Intercept                         2.6020    327.874      0.008      0.994    -659.075     664.279
china_aid_diff                   -0.0184      0.036     -0.506      0.616      -0.092       0.055
corruption_index                 48.1261     38.565      1.248      0.219     -29.702     125.954
political_stability             -65.0569     42.400     -1.534      0.132    -150.623      20.509
aid_by_multilaterals             -0.0716      0.140     -0.511      0.612      -0.354       0.211
aid_by_dac_countries             -0.1335      0.108     -1.241      0.221      -0.351       0.084
gdp_per_capita                   -0.0497      0.036     -1.381      0.175      -0.122       0.023
gdp_growth_rate                  48.7100     31.367      1.553      0.128     -14.591     112.011
total_natural_resources_rents   670.7321    352.538      1.903      0.064     -40.719    1382.183
exports_plus_imports           -200.4000    241.490     -0.830      0.411    -687.747     286.947
world_development_indicators      9.3103     49.547      0.188      0.852     -90.680     109.301
==============================================================================
Omnibus:                        2.372   Durbin-Watson:                   1.476
Prob(Omnibus):                  0.305   Jarque-Bera (JB):                2.131
Skew:                           0.483   Prob(JB):                        0.345
Kurtosis:                       2.827   Cond. No.                     2.65e+04
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 2.65e+04. This might indicate that there are
strong multicollinearity or other numerical problems.


Modeling DID terms for year 2011
####################################################
                            OLS Regression Results                            
==============================================================================
Dep. Variable:            us_aid_diff   R-squared:                       0.124
Model:                            OLS   Adj. R-squared:                 -0.084
Method:                 Least Squares   F-statistic:                    0.5958
Date:                Fri, 23 Dec 2022   Prob (F-statistic):              0.808
Time:                        03:47:57   Log-Likelihood:                -399.48
No. Observations:                  53   AIC:                             821.0
Df Residuals:                      42   BIC:                             842.6
Df Model:                          10                                         
Covariance Type:            nonrobust                                         
=================================================================================================
                                    coef    std err          t      P>|t|      [0.025      0.975]
-------------------------------------------------------------------------------------------------
Intercept                      -555.4491    474.071     -1.172      0.248   -1512.162     401.264
china_aid_diff                   -0.0354      0.048     -0.733      0.468      -0.133       0.062
corruption_index                 43.8755     47.884      0.916      0.365     -52.759     140.510
political_stability              48.3575     41.231      1.173      0.247     -34.850     131.565
aid_by_multilaterals              0.0361      0.119      0.302      0.764      -0.205       0.277
aid_by_dac_countries              0.0109      0.103      0.106      0.916      -0.196       0.218
gdp_per_capita                   -0.0159      0.037     -0.428      0.671      -0.091       0.059
gdp_growth_rate                  -6.1782     37.196     -0.166      0.869     -81.243      68.887
total_natural_resources_rents   502.7817    369.727      1.360      0.181    -243.358    1248.922
exports_plus_imports            345.7519    326.215      1.060      0.295    -312.577    1004.081
world_development_indicators    -27.6749     59.932     -0.462      0.647    -148.623      93.273
==============================================================================
Omnibus:                        1.462   Durbin-Watson:                   1.752
Prob(Omnibus):                  0.481   Jarque-Bera (JB):                1.211
Skew:                          -0.367   Prob(JB):                        0.546
Kurtosis:                       2.903   Cond. No.                     3.48e+04
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 3.48e+04. This might indicate that there are
strong multicollinearity or other numerical problems.

Modeling DID terms for year 2012
####################################################
                            OLS Regression Results                            
==============================================================================
Dep. Variable:            us_aid_diff   R-squared:                       0.114
Model:                            OLS   Adj. R-squared:                 -0.097
Method:                 Least Squares   F-statistic:                    0.5410
Date:                Fri, 23 Dec 2022   Prob (F-statistic):              0.851
Time:                        02:55:58   Log-Likelihood:                -404.99
No. Observations:                  53   AIC:                             832.0
Df Residuals:                      42   BIC:                             853.7
Df Model:                          10                                         
Covariance Type:            nonrobust                                         
=================================================================================================
                                    coef    std err          t      P>|t|      [0.025      0.975]
-------------------------------------------------------------------------------------------------
Intercept                      -314.5468    518.954     -0.606      0.548   -1361.838     732.744
china_aid_diff                   -0.0415      0.043     -0.968      0.338      -0.128       0.045
corruption_index                 -0.1436     44.829     -0.003      0.997     -90.612      90.324
political_stability              39.4433     50.467      0.782      0.439     -62.403     141.289
aid_by_multilaterals              0.0099      0.159      0.062      0.951      -0.311       0.331
aid_by_dac_countries              0.0811      0.116      0.701      0.487      -0.153       0.315
gdp_per_capita                    0.0367      0.041      0.895      0.376      -0.046       0.120
gdp_growth_rate                  14.4673     36.582      0.395      0.694     -59.358      88.293
total_natural_resources_rents  -190.4429    380.632     -0.500      0.619    -958.589     577.703
exports_plus_imports            138.2222    303.086      0.456      0.651    -473.430     749.874
world_development_indicators    -66.4344     51.924     -1.279      0.208    -171.221      38.353
==============================================================================
Omnibus:                        0.383   Durbin-Watson:                   1.960
Prob(Omnibus):                  0.826   Jarque-Bera (JB):                0.184
Skew:                          -0.144   Prob(JB):                        0.912
Kurtosis:                       2.991   Cond. No.                     3.61e+04
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 3.61e+04. This might indicate that there are
strong multicollinearity or other numerical problems.


Modeling DID terms for year 2013
####################################################
                            OLS Regression Results                            
==============================================================================
Dep. Variable:            us_aid_diff   R-squared:                       0.297
Model:                            OLS   Adj. R-squared:                  0.130
Method:                 Least Squares   F-statistic:                     1.778
Date:                Fri, 23 Dec 2022   Prob (F-statistic):             0.0951
Time:                        02:55:58   Log-Likelihood:                -394.41
No. Observations:                  53   AIC:                             810.8
Df Residuals:                      42   BIC:                             832.5
Df Model:                          10                                         
Covariance Type:            nonrobust                                         
=================================================================================================
                                    coef    std err          t      P>|t|      [0.025      0.975]
-------------------------------------------------------------------------------------------------
Intercept                       236.9031    398.864      0.594      0.556    -568.037    1041.843
china_aid_diff                   -0.0037      0.030     -0.121      0.905      -0.065       0.058
corruption_index                  0.7919     35.630      0.022      0.982     -71.113      72.696
political_stability              78.6325     41.904      1.876      0.068      -5.933     163.198
aid_by_multilaterals             -0.3395      0.137     -2.472      0.018      -0.617      -0.062
aid_by_dac_countries             -0.0892      0.094     -0.952      0.347      -0.278       0.100
gdp_per_capita                    0.0065      0.038      0.173      0.864      -0.069       0.082
gdp_growth_rate                   0.3727     30.466      0.012      0.990     -61.110      61.855
total_natural_resources_rents  -708.3941    353.717     -2.003      0.052   -1422.225       5.436
exports_plus_imports            212.2838    255.258      0.832      0.410    -302.848     727.415
world_development_indicators     87.9388     49.588      1.773      0.083     -12.133     188.011
==============================================================================
Omnibus:                        2.017   Durbin-Watson:                   2.150
Prob(Omnibus):                  0.365   Jarque-Bera (JB):                1.285
Skew:                          -0.012   Prob(JB):                        0.526
Kurtosis:                       2.238   Cond. No.                     3.36e+04
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 3.36e+04. This might indicate that there are
strong multicollinearity or other numerical problems.


Modeling DID terms for year 2014
####################################################
                            OLS Regression Results                            
==============================================================================
Dep. Variable:            us_aid_diff   R-squared:                       0.143
Model:                            OLS   Adj. R-squared:                 -0.061
Method:                 Least Squares   F-statistic:                    0.7012
Date:                Fri, 23 Dec 2022   Prob (F-statistic):              0.718
Time:                        02:55:58   Log-Likelihood:                -402.71
No. Observations:                  53   AIC:                             827.4
Df Residuals:                      42   BIC:                             849.1
Df Model:                          10                                         
Covariance Type:            nonrobust                                         
=================================================================================================
                                    coef    std err          t      P>|t|      [0.025      0.975]
-------------------------------------------------------------------------------------------------
Intercept                       329.1024    395.143      0.833      0.410    -468.329    1126.534
china_aid_diff                   -0.0097      0.039     -0.247      0.806      -0.088       0.069
corruption_index                  3.7128     47.015      0.079      0.937     -91.167      98.593
political_stability             -96.6520     48.701     -1.985      0.054    -194.934       1.630
aid_by_multilaterals              0.0096      0.161      0.060      0.953      -0.315       0.334
aid_by_dac_countries             -0.0099      0.106     -0.093      0.926      -0.223       0.204
gdp_per_capita                    0.0212      0.040      0.535      0.595      -0.059       0.101
gdp_growth_rate                 -10.0106     39.675     -0.252      0.802     -90.078      70.056
total_natural_resources_rents   -39.5836    435.459     -0.091      0.928    -918.376     839.209
exports_plus_imports           -117.3107    280.743     -0.418      0.678    -683.873     449.252
world_development_indicators    -61.0425     58.412     -1.045      0.302    -178.922      56.837
==============================================================================
Omnibus:                        3.431   Durbin-Watson:                   1.604
Prob(Omnibus):                  0.180   Jarque-Bera (JB):                1.750
Skew:                          -0.092   Prob(JB):                        0.417
Kurtosis:                       2.129   Cond. No.                     2.78e+04
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 2.78e+04. This might indicate that there are
strong multicollinearity or other numerical problems.


Modeling DID terms for year 2015
####################################################
                            OLS Regression Results                            
==============================================================================
Dep. Variable:            us_aid_diff   R-squared:                       0.120
Model:                            OLS   Adj. R-squared:                 -0.090
Method:                 Least Squares   F-statistic:                    0.5702
Date:                Fri, 23 Dec 2022   Prob (F-statistic):              0.829
Time:                        02:55:58   Log-Likelihood:                -396.05
No. Observations:                  53   AIC:                             814.1
Df Residuals:                      42   BIC:                             835.8
Df Model:                          10                                         
Covariance Type:            nonrobust                                         
=================================================================================================
                                    coef    std err          t      P>|t|      [0.025      0.975]
-------------------------------------------------------------------------------------------------
Intercept                        75.4967    354.102      0.213      0.832    -639.111     790.104
china_aid_diff                    0.0033      0.032      0.102      0.919      -0.061       0.068
corruption_index                 16.2461     38.602      0.421      0.676     -61.657      94.149
political_stability              39.9740     47.321      0.845      0.403     -55.523     135.471
aid_by_multilaterals              0.0004      0.124      0.003      0.998      -0.249       0.250
aid_by_dac_countries             -0.0201      0.095     -0.212      0.833      -0.211       0.171
gdp_per_capita                   -0.0717      0.035     -2.071      0.045      -0.142      -0.002
gdp_growth_rate                   9.0403     31.469      0.287      0.775     -54.466      72.547
total_natural_resources_rents   225.9087    421.254      0.536      0.595    -624.216    1076.033
exports_plus_imports            -64.6298    256.978     -0.251      0.803    -583.231     453.972
world_development_indicators     21.8266     47.308      0.461      0.647     -73.644     117.297
==============================================================================
Omnibus:                        0.937   Durbin-Watson:                   1.535
Prob(Omnibus):                  0.626   Jarque-Bera (JB):                0.976
Skew:                           0.200   Prob(JB):                        0.614
Kurtosis:                       2.469   Cond. No.                     2.93e+04
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 2.93e+04. This might indicate that there are
strong multicollinearity or other numerical problems.


Modeling DID terms for year 2016
####################################################
                            OLS Regression Results                            
==============================================================================
Dep. Variable:            us_aid_diff   R-squared:                       0.254
Model:                            OLS   Adj. R-squared:                  0.077
Method:                 Least Squares   F-statistic:                     1.431
Date:                Fri, 23 Dec 2022   Prob (F-statistic):              0.200
Time:                        02:55:58   Log-Likelihood:                -401.58
No. Observations:                  53   AIC:                             825.2
Df Residuals:                      42   BIC:                             846.8
Df Model:                          10                                         
Covariance Type:            nonrobust                                         
=================================================================================================
                                    coef    std err          t      P>|t|      [0.025      0.975]
-------------------------------------------------------------------------------------------------
Intercept                       358.1952    390.864      0.916      0.365    -430.601    1146.991
china_aid_diff                    0.0287      0.035      0.816      0.419      -0.042       0.100
corruption_index                 95.5761     53.209      1.796      0.080     -11.804     202.956
political_stability             -33.8464     47.889     -0.707      0.484    -130.490      62.798
aid_by_multilaterals             -0.1107      0.166     -0.667      0.509      -0.446       0.224
aid_by_dac_countries             -0.1242      0.121     -1.026      0.311      -0.368       0.120
gdp_per_capita                    0.0033      0.043      0.075      0.940      -0.084       0.091
gdp_growth_rate                 -81.6975     34.902     -2.341      0.024    -152.132     -11.263
total_natural_resources_rents    23.4190    361.329      0.065      0.949    -705.773     752.611
exports_plus_imports            -86.3494    259.964     -0.332      0.741    -610.977     438.279
world_development_indicators     15.4040     53.781      0.286      0.776     -93.129     123.938
==============================================================================
Omnibus:                        2.296   Durbin-Watson:                   2.037
Prob(Omnibus):                  0.317   Jarque-Bera (JB):                1.397
Skew:                          -0.064   Prob(JB):                        0.497
Kurtosis:                       2.215   Cond. No.                     2.54e+04
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 2.54e+04. This might indicate that there are
strong multicollinearity or other numerical problems.


Modeling DID terms for year 2017
####################################################
                            OLS Regression Results                            
==============================================================================
Dep. Variable:            us_aid_diff   R-squared:                       0.173
Model:                            OLS   Adj. R-squared:                 -0.024
Method:                 Least Squares   F-statistic:                    0.8759
Date:                Fri, 23 Dec 2022   Prob (F-statistic):              0.562
Time:                        02:55:58   Log-Likelihood:                -395.63
No. Observations:                  53   AIC:                             813.3
Df Residuals:                      42   BIC:                             834.9
Df Model:                          10                                         
Covariance Type:            nonrobust                                         
=================================================================================================
                                    coef    std err          t      P>|t|      [0.025      0.975]
-------------------------------------------------------------------------------------------------
Intercept                       196.1850    363.406      0.540      0.592    -537.199     929.569
china_aid_diff                    0.0259      0.037      0.708      0.483      -0.048       0.100
corruption_index                  4.0989     44.356      0.092      0.927     -85.415      93.613
political_stability             -15.6815     39.534     -0.397      0.694     -95.464      64.101
aid_by_multilaterals              0.1506      0.138      1.094      0.280      -0.127       0.429
aid_by_dac_countries             -0.0996      0.105     -0.950      0.347      -0.311       0.112
gdp_per_capita                   -0.0337      0.041     -0.819      0.417      -0.117       0.049
gdp_growth_rate                  -1.8980     27.072     -0.070      0.944     -56.531      52.735
total_natural_resources_rents    96.0068    399.396      0.240      0.811    -710.007     902.021
exports_plus_imports            174.3151    252.587      0.690      0.494    -335.426     684.056
world_development_indicators    -84.1049     43.486     -1.934      0.060    -171.863       3.653
==============================================================================
Omnibus:                        1.185   Durbin-Watson:                   2.068
Prob(Omnibus):                  0.553   Jarque-Bera (JB):                1.227
Skew:                          -0.302   Prob(JB):                        0.542
Kurtosis:                       2.564   Cond. No.                     3.24e+04
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 3.24e+04. This might indicate that there are
strong multicollinearity or other numerical problems.

### Data
- `oced_oaf_oof.csv` - https://stats.oecd.org
- https://foreignassistance.gov/data
- https://www.aiddata.org/data/aiddatas-global-chinese-development-finance-dataset-version-2-0
- GDP per capita: https://data.worldbank.org/indicator/NY.GDP.PCAP.CD

### Notebook
Hypothesis: Increase in foreign aid provided to a country per capita in Africa by China leads to an increase in foreign aid provided per capita to that country by the USA.

Null hypothesis: if we are wrong, then it should be possible to reliably observe increase in aid by the US _not_ following an increase in aid by CCP.

Need to analyze interaction effects between the confounding variables

## Sources

1. "Despite Its Reputation, Chinese Aid is Quite Effective." The Economist, 12 October 2017. https://www.economist.com/china/2017/10/12/despite-its-reputation-chinese-aid-is-quite-effective. Accessed 2 October 2022.

2. "China's Aid to Africa: Monster or Messiah?" Brookings Institution, https://www.brookings.edu/opinions/chinas-aid-to-africa-monster-or-messiah/. Accessed 2 December 2022.

3. "Chinese Economic Engagement in Africa." Foreign Policy Research Institute, https://www.fpri.org/article/2022/01/chinese-economic-engagement-in-africa/. Accessed 2 December 2022.

4. Chatzky, Andrew, and James McBride. “China's Massive Belt and Road Initiative.” Council on Foreign Relations, 28 January 2020, https://www.cfr.org/backgrounder/chinas-massive-belt-and-road-initiative. Accessed 12 November 2022.

5. “FACT SHEET: President Biden and G7 Leaders Launch Build Back Better World (B3W) Partnership.” The White House, 12 June 2021, https://www.whitehouse.gov/briefing-room/statements-releases/2021/06/12/fact-sheet-president-biden-and-g7-leaders-launch-build-back-better-world-b3w-partnership/. Accessed 2 October 2022.

6. Regilme, Salvador Santino F., and Obert Hodzi. “Comparing US and Chinese Foreign Aid in the Era of Rising Powers.” The International Spectator, vol. 56, no. 2, 2021, pp. 114-131, https://doi.org/10.1080/03932729.2020.1855904.

7. Berthélemy, J.-C. (2006), Bilateral Donors’ Interest vs. Recipients’ Development Motives in Aid Allocation: Do All Donors Behave the Same?. Review of Development Economics,10:179-194.https://doi-org.proxy.library.nyu.edu/10.1111/j.1467-9361.2006.00311.x

8. Hoeffler, A. and Outram, V. (2011), Need, Merit, or Self-Interest—What Determines the Allocation of Aid?. Review of Development Economics, 15: 237-250. https://doi-org.proxy.library.nyu.edu/10.1111/j.1467-9361.2011.00605.x

9. Dreher, A. and Fuchs, A. (2015), Rogue aid? An empirical analysis of China's aid allocation. Canadian Journal of Economics/Revue canadienne d'économique, 48: 988-1023. https://doi-org.proxy.library.nyu.edu/10.1111/caje.12166

10. Humphrey, Chris, and Katharina Michaelowa. “China in Africa: Competition for traditional development finance institutions?” World Development, vol. 120, 2019, pp. 15-28, (https://www.sciencedirect.com/science/article/pii/S0305750X19300725).

11. Marta Marson, Ivan Savin, Complementary or adverse? Comparing development results of official funding from China and traditional donors in Africa. Structural Change and Economic Dynamics, vol. 62, 2022, pp. 189-206, https://doi.org/10.1016/j.strueco.2022.04.010.

12. Watkins, M. Undermining conditionality? The effect of Chinese development assistance on compliance with World Bank project agreements. Rev Int Organ 17, 667–690 (2022). https://doi-org.proxy.library.nyu.edu/10.1007/s11558-021-09443-z

13. Cormier, B. Chinese or western finance? Transparency, official credit flows, and the international political economy of development. Rev Int Organ (2022). https://doi.org/10.1007/s11558-022-09469-x

14. “Despite its reputation, Chinese aid is quite effective.” The Economist, 12 October 2017, https://www.economist.com/china/2017/10/12/despite-its-reputation-chinese-aid-is-quite-effective. Accessed 2 October 2022.

15. Chatzky, Andrew, and James McBride. “China's Massive Belt and Road Initiative.” Council on Foreign Relations, 28 January 2020, https://www.cfr.org/backgrounder/chinas-massive-belt-and-road-initiative. Accessed 12 November 2022.

16. “FACT SHEET: President Biden and G7 Leaders Launch Build Back Better World (B3W) Partnership.” The White House, 12 June 2021, https://www.whitehouse.gov/briefing-room/statements-releases/2021/06/12/fact-sheet-president-biden-and-g7-leaders-launch-build-back-better-world-b3w-partnership/. Accessed 2 October 2022.

17. Regilme, Salvador Santino F., and Obert Hodzi. “Comparing US and Chinese Foreign Aid in the Era of Rising Powers.” The International Spectator, vol. 56, no. 2, 2021, pp. 114-131, https://doi.org/10.1080/03932729.2020.1855904.

18. Non-Interference 2.0: China's Evolving Foreign Policy towards a Changing Africa : https://journals.sagepub.com/doi/10.1177/186810261504400105 Appendix

19. "Corruption Index." Organisation for Economic Co-operation and Development, https://cran.r-project.org/package=OECD.

20. "Political Stability." World Bank, https://pypi.org/project/worldbank/.

21. "Aid by Multilaterals, total." Organisation for Economic Co-operation and Development, https://data.oecd.org/oda/flows/api/1.0/.

22. "Aid by DAC countries." Organisation for Economic Co-operation and Development, https://data.oecd.org/api/oda/dac/v2/flow/donor/recipient/year.json.

23. "Aid by USA." Organisation for Economic Co-operation and Development, https://data.oecd.org/api/oda/dac/v2/flow/donor/recipient/year.json.

24. "GDP per capita." World Bank, https://data.worldbank.org/indicator/NY.GDP.PCAP.CD.

25. "GDP growth rate." World Bank, https://data.worldbank.org/indicator/NY.GDP.MKTP.KD.ZG.

26. "Total natural resources rents (% of GDP)." World Bank, https://data.worldbank.org/indicator/NY.GDP.TOTL.RT.ZS.

27. "World Development Indicators." World Bank, https://databank.worldbank.org/reports.aspx?source=2&series=LP.LPI.INFR.XQ&country=.

28. "Trade as a percentage of GDP." World Bank, https://data.worldbank.org/indicator/NE.TRD.GNFS.ZS?most_recent_year_desc=true/.

30. "Population per country." World Bank, https://datacatalog.worldbank.org/dataset/population-total.

31. "China Aid." AidData, https://www.aiddata.org/data/aiddatas-global-chinese-development-finance-dataset-version-2-0.

32. "Our Work in Africa. (n.d.). United Nations Environment Programme. Dec 22,2022 from https://www.unep.org/regions/africa/our-work-africa"

33. Sany, J., Sheehy, T. P. (2021, April 28). Sidestepping Great Power Rivalry: U.S.-China competition in Africa. United States Institute of Peace. Retrieved December 22, 2022, from https://www.usip.org/publications/2021/04/sidestepping-great-power-rivalry-us-china-competition-africa

34. Jones, C. D. (2022, January 24). Chinese economic engagement in Africa: Implications for U.S. policy. Foreign Policy Research Institute. Retrieved December 22, 2022, from https://www.fpri.org/article/2022/01/chinese-economic-engagement-in-africa/