What secrets lie in balance sheets? Today, we’re trying to find out. This season on the failure of Silicon Valley Bank (SVB) we’ve bounced between a few different datasets, including macroeconomic data and headline assets and liabilities. This week’s episode takes a more detailed look at the bank’s balance sheet data over the last ten years and tries to place that in the context of the wider banking industry.

Neeraj has spent the week wrangling with balance sheet information to try and make it add up, eventually using data published by the bank itself to break down the contributors to its total assets and liabilities. The results support the accepted conclusion that Silicon Valley Bank held too much of its assets in illiquid long term investments. To try and contextualise these results, Nick has been throwing machine learning at the problem, using dimensionality reduction techniques to try and visualise banks with similar balance sheet breakdowns and how these have varied over time. On the face of it, there are the big banks and then everyone else. But look closer and you can see Silicon Valley Bank on a journey of its own!

References

SVB balance sheet, CAR over time, and t-SNE balance sheet Datalore notebook

https://twitter.com/KobeissiLetter/status/1641474423269081088?t=ZUMQNOlLIgZnkmUG2RjwMw&s=33

This week’s Jupyter notebook on Github

Thumbnail DALL·E prompt: "cartoon goose holding a long scroll of paper and painting a graph with a red line and black grid lines on a canvas that is sitting on an easel”

Please join us on our website and Discord:
https://www.wildtruthchase.com

https://discord.gg/gqM54csKSB

In season five, episode four of a finance-focused podcast, the hosts analyze the balance sheets of Silicon Valley Bank using two approaches: one using AlphaVantage to compare the bank's data with other banks, and the other comparing the bank's capital adequacy ratios with other banks. The analysis reveals a shift in the bank's asset ratio and a low capital adequacy ratio, which could cause issues with converting assets into cash. The podcasters also discuss the industry-wide decrease in capital adequacy ratios for banks due to the pandemic, highlighting several regional banks that may be in trouble. The speakers use machine learning to visualize the balance sheets of different banks and discuss how Silicon Valley Bank stands out on the plot. The conversation concludes with a discussion of future analysis, including looking at macroeconomic indicators and other failed banks. The code used for the analysis is available on GitHub, and the AlphaVantage API can be used to obtain data.

#financepodcast #bankinganalysis #AIforfinance

Timestamps

[00:00:00] "Silicon Valley Bank's Balance Sheets: A Recap"

[00:00:40] "LM Usage: Chat GPT Example and Episode Description Comparisons"

[00:02:21] Falling Number of US Banks: Opening vs. Closing

[00:04:35] Analyzing Silicon Valley Bank's balance sheet changes over the past decade

[00:08:11] "Capital Adequacy Ratio and Troubled Banks"

[00:12:48] "Visualizing High Dimensional Data with Dimensionality Reduction Animation"

[00:16:38] The Financial Dangers of Silicon Valley Bank in 2020