Small Caps, Overlaps and ChatGPT

While it’s clear that conversational artificial intelligence is the new technological breakthrough everyone is talking about, using it to improve financial decisions could be tricky because there is a lot of garbage on the internet about finance and investments, that may be confusing or even detrimental if used as inputs to a learning model.

We thought we’d take our first spin in ChatGPT by asking it an important question to which we believe we already know the ballpark answer thanks to our decades of capital markets experience.

Our basic thesis is that once you own an index like the NIFTY and maybe one more large-cap mutual fund, there is no further benefit to owning any more large- and mid-cap funds in your equity portfolio, since many of these funds own many of the same stocks. If you want more equity exposure, it is best to select one good small-cap fund in your portfolio.

We started our experiment by asking “CGPT” how one could find the most frequently overlapping stocks between the leading Indian mutual funds.

(Editorial note: We are presenting excerpts of our conversation with CGPT, which is a third-party tool and we don’t claim to verify or endorse any specific findings or data that it provides)

CGPT lists out steps to perform the type of analysis that we at Shepherd’s Hill have done. Basically, you take the top mutual funds, their schemes, the holdings in the schemes and count the frequency with which the holdings appear in these schemes.

So, we asked it to list the largest mutual funds in the country. It listed ten mutual fund asset management companies (AMCs), ordering them (without being prompted) by assets under management, the top three being HDFC, ICICI and SBI. As many of you may know, CGPT qualifies its investment-related answers with the following comment: “my knowledge cutoff is [September 2021]”. This is okay for us because we are looking for a fundamental truth, not a short-term stock tip.

Next, we asked it to list the top 10 funds or “schemes” (by AUM) of the top 3 AMCs. Then, we asked it to list the top 10 holdings of each of the top 3 schemes of the top 3 AMCs. 

Finally, we asked “Can you now take these top holdings lists and list them by descending order of frequency with which they appear in the top 10 holdings of the top 3 schemes of these three AMCs?”

And to tie a bow on it, we said “Can you add a comment ‘NIFTY stock’ to each of the holdings in the list if they are indeed a part of the NIFTY 50 index?”

Here is what CGPT came up with (remember, its data is as of September 2021):

• HDFC Bank Ltd. – 7 (NIFTY stock)

• Reliance Industries Ltd. – 6 (NIFTY stock)

• ICICI Bank Ltd. – 6 (NIFTY stock)

• Housing Development Finance Corporation Ltd. – 5 (NIFTY stock)

• Infosys Ltd. – 4 (NIFTY stock)

• State Bank of India. – 3 (NIFTY stock)

• Axis Bank Ltd. – 3 (NIFTY stock)

• Bharti Airtel Ltd. – 3 (NIFTY stock)

• Larsen & Toubro Ltd. – 3 (NIFTY stock)

• Tata Consultancy Services Ltd. – 3 (NIFTY stock)

If CGPT is correct, then each of these top 10 companies are owned by a large percentage of top funds in the country. In that case, there is no point in owning more than a couple of these funds because they overlap so much, and you will end up buying the same product in different packaging.

Nice work, CGPT!