I’ve been seeing a lot of YouTube videos lately about whether there will be a recession in 2023, how bad it will be, how long it will last, etc.
They look at charts like the yield curve, interest rates, unemployment, … and try to use what they see “intuitively” determine whether or not there will be a recession.
Statements such as:
- “The yield curve is going down, which usually happens before a recession”
- “But unemployment is down, which indicates that there might NOT be a recession”
Of course, looking at only a single feature at a time might lead to contradictions such as the above and hence is suboptimal.
This sounds like a problem for data science and machine learning.
Machine learning models can incorporate ALL input features at once, and model the dependency of the outcome (recession) as a function of those inputs.
Check out the video on this here:
To get the full code, note that it was just added to my course on Financial Engineering.
This course goes over many important aspects of finance, such as the efficient market hypothesis (whether or not future events like interest rate hikes are already “priced in” to the market), the random walk hypothesis (a statistical statement about the predictability of future stock prices based on the past), portfolio theory (how to optimally invest in a basket of assets), and algorithmic trading (how to use computer algorithms to choose the best time to buy and sell).