Dollar Power: Can Machine Learning and Econometrics Solve BRICS? | By Jared Bilbury | Jun 2023

Machine Learning


Photo by Kyle Glenn on Unsplash

econometric model

By definition, econometrics is the application of statistical methods to economic data to give economic relations an empirical content. And while some scholars consider this to be the original data science, it is also great for analyzing economic theories and turning them into quantifiable metrics.

So I did what any curious analyst would do. That is, we created a set of linear models to explore the relationships between variables/phenomena within the BRICS economy. The determining factors are:

Trade balance (in billions)

Percentage of each country’s GDP

terms of trade

effective exchange rate

annual growth rate

The data I use starts from 2009 (the year the first BRICs summit was held) to gain insight into how national economies have behaved over the past 12-13 years. South Africa also joined him in 2010, allowing him to represent the model more accurately. The dataset is here.

Finally, the analysis was done in R by creating a series of linear regression models and analyzing the determinants for US GDP.

linear regression

The basic command I started with was a simple linear regression on two variables. Almost immediately, I was able to find a series of correlations between the two key variables. The code and results are below.

Figure 4: Linear model between US GDP and China’s effective exchange rate

As you can see, there is a direct correlation between US GDP and China’s effective exchange rate. A country’s effective exchange rate represents how that country’s currency values ​​relative to indices of other major currencies. In other words, The richer America gets, the stronger China’s currency gets.. This is directly related to what was shown above for real GDP between China and the US. This also shows why China is wise to maintain a productive relationship with us, despite some differences.

So, we know about China and how big its influence is, but what about the rest of the BRICS countries? More insight into the distribution of was also revealed.

Figure 5: Linear model between China’s annual growth rate and India’s annual growth rate
Figure 6: Linear model between China’s trade balance and Russia’s effective exchange rate

At this point, you may notice patterns in the data. Within the BRICS countries, most variables in the model tend to be mutually correlated. In this case, we find that India’s annual growth rate is directly correlated with China’s annual growth rate. However, the United States not yet India’s largest trading partner. Basically, what we are looking at is the difference in how these two markets create value for India. The United States imports more goods from India than it exports instead.However, India has almost double More than the amount imported from the United States. Combine this with China’s exchange rate appreciation and US GDP growth, and the results are easy to analyze.

As for Russia, we see the flip side of this relationship: China’s expanding trade balance is having a negative impact on Russia’s exchange rate. But the model becomes more complicated when we analyze the economies China and Russia work together. Russia has gradually been isolated through a series of sanctions against some of the world’s most resource-rich countries (Australia, Canada and the United States). For this reason, the country relies heavily on China for technical and economic support. However, China is diversifying its economic portfolio to not only include these countries, but also to become some of its major trading partners.. Ultimately, a perfect storm of factors (oil and gas, restrictive sanctions, and the current war with Ukraine) seems to have contributed to the model’s results, but this likely doesn’t include China directly. have a nature. In addition to this, finding good economic data for Russia can be somewhat complicated.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *