Tata Asset Management company is launching a new fund “Tata Quant Fund”. This fund is going to use machine learning to aid stock selection and portfolio allocation. It is planning to use statistical techniques to identify patterns and their correlation to performance of stocks. Using same, a strategy would be formalized. This model would keep learning and validating based on feedback and will thrive to improve continuously. Sounds interesting? Let’s check if you should invest.
1. Basic Investment Details
Let’s first start with basic investment related details of this fund. To be honest, there is nothing significantly different operationally from other funds in India. But just for reference and good to know view, below are details.
|Type||Open Ended||You can invest anytime even after NFO|
|Important Dates||NFO opens on 3rd Jan 2020 and closes on 17th Jan 2020||Would re open for regular subscription on 28th January, 2020|
|Plans available||Growth and Dividend, Direct and Regular|
|Investment Options||Lumpsum, SIP|
|Minimum Investment||Lumpsum - Rs. 5000|
SIP - Rs. 500
|Load||Entry load - nil. Exit Load: 1% if redeemed within 1 year (365 days) of allotment|
|Important Documents||Scheme Broucher||Key Information cum Application form|
|Scheme Information Document||Scheme Presentation|
|This document is what you should try to understand before investing|
|Who can not invest||• A person who falls within the definition of the term “U.S” Person” under the US Securities Act of 1933 and corporations or other entities organised under the laws of the U.S.
• A person who is resident of Canada.
• OCB (Overseas Corporate Bodies) as defined under Income Tax Act, 1061 and under Foreign Exchange Management Act, 1999
|Benchmark||BSE 200 TRI|
2. Investment Objective
Objective of the scheme is to generate medium to long-term capital appreciation by investing in equity and equity related instruments. These will be selected based on a quantitative model (that’s why name – Quant). Of course, there is no assurance or guarantee that the investment objective of the Scheme will be achieved. So scheme does not assure or guarantee any returns.
3. Portfolio & Asset allocation of Tata Quant
Under normal circumstances, the asset allocation of the Scheme will be as mentioned in below table. What it means in simple term is mostly fund will remain invested full in Indian equities and related instruments. This is inline with most of other equity funds. So not much of difference here. Overall your risk profile should be at least medium and preferably high to be suitable.
|Instruments||Indicative allocations (% of net assets)||Risk Profile|
|Equity & Equity related instruments||80||100||Medium to High|
|Debt & Money Market instruments||0||20||Low to Medium|
|Units issued by REITs and InvITs||0||10||Medium to High|
Investment would be in BSE 200 stocks which means portfolio would consist well established companies. Stock selection will be based on machine learning algorithm and also prediction of bearish or bullish nature of upcoming period. Presentation mentions stocks would be selected on top scores using alpha, value and quality bucket.
Parameters fund intends to use for scoring are standard ones. These are ROCE (Return on Capital Employed), Price to Earning (P/E), Price to Book Value (P/B), Dividend yield (D/P) parameters, Return on Equity (ROE), Debt-to Equity (D/E) and Earning Per Share (EPS) growth. Difference is that this fund will be depending more of mathematical score for these factors. Other funds use fund manager judgement based on these values. So though fund manager may have final say, selection will be as per calculation of score of each company.
4. Know your Fund Manager
Since Fund Manager is key in actively managed funds, let’s understand who is fund manager of Tata Quant Fund.
- MBA (Finance) from Queensland University of Technology from Australia.
- More than 15 years of experience in fund management and broking,
- Joined Tata Asset Management in November 2018 as Fund Manager
- Prior stint includes IDFC Securities Ltd. where he was heading the Institutional Derivatives desk. (Equities).
Other schemes managed by him are as under. He joined Tata Asset Management company end of 2018. So we can refer performance of these schemes in 2019 to see his investment style and performance. Overall, I feel he has delivered in line or little better than category in most cases. But difference is not significantly high compared to category average. At same time he is not lagging behind the category. So I would be neutral in this fund from fund manager perspective. Also being an algorithmic fund, it is not clear how much say fund manager will have over machine recommendations.
- Tata Digital India Fund (11.85% vs 11.40%)
- Tata Equity Savings Fund (Equity Portfolio) (9.64% vs 8.80%)
- Tata India Pharma and Healthcare Fund (9.96% vs 7.41%)
- Tata Resources and Energy Fund (15.59% vs no fund to compare)
- Tata Arbitrage Fund (7.36% vs 6.56%)
- Tata Nifty Exchange Traded Fund (N.A. as passive fund)
- Tata Balanced Advantage Fund (hedged equity portfolio) (3.17 % vs 4.05% for six months)
- Tata Nifty Private Bank Exchange traded fund (N.A. as passive fund)
4. USP of this fund – Quant / AI / Machine Learning?
Lets now look at unique selling point of this fund. What is differentiating factor in this fund from all the sea of funds that are available in market. Well this fund is going to use new technology of Machine Learning which is subset of AI or Artificial Intelligence. For common person, this means that based on data and trends, there will be a model created for stock selection. In Machine learning, Model is more or less a mathematical formula which evaluates various options and does recommendations.
Here plan is to divide stocks as per different strategies referred as alpha, value, quality. Then use machine learning to give score to each stock in BSE 200 and select stocks that are top scorers. Additionally, fund plans to use machine learning and predict if next month will be bullish or bearish. Based on that, fund could invest more or hedge more. Overall a new concept. This may turn out promising in future but should put your money and be guinea pig during this experiment.
5. Who are peers of Tata Quant Fund
There are couple of funds which are currently using mathematical models viz Nippon India Quant Fund and DSP Quant Fund. Nippon fund delivered around 9% compared to 13% of BSE 200 TRI which is far below. However DSP fund is new kid in block and yet to complete a year. It has delivered 9.65% vs 3.15% of BSE 200 TRI which is comprehensively beating benchmark. However its a new fund and need to be further observed. So looking at peers, we don’t get conviction that such quant strategy is sure shot winner.
6. Risk factors
1) First and foremost, you should understand underlying asset is equity. So all the risks associated with equity are applicable. However good model may be, it will have its share of volatility.
2) Being a Quant Fund, Investment strategies mathematical based on historical data. This strategy haven’t shown any success story in India over long term.
3) Machine learning and AI is still evolving and this is one of initial attempt in mutual fund in India so we do not have any reference to confirm this will work. What if model does not work with real data and how investors will be safeguarded is not clear.
4) This fund may invest in REITs and InvITs which have high risk.
“Never invest in a business you cannot understand.” – Warren Buffett
Well, if above quote does not give enough ideas about my opinion, please read on.
Should you invest in Tata Quant Fund during NFO? My opinion is “Not now. Please wait…” at this time. You should wait for say six months to see portfolio as well as performance of fund against benchmark Index. Only then you should put your money on this new concept. Main reasons for my recommendation are as under:
- Product is very complex to understand. Not sure about you but I am not very clear how this all is going to work. I would blame mainly on marketing team for this. They were able to create some sample relevant examples to explain approach. In their presentation, they took a simplistic approach of comparing with cricket team to explain how ML works. Instead of that taking one real scenario in past where the model was tested would have helped. Explaining how model recommended stock composition at that time would have given some confidence.
- It talks using ML for predicting if it will be bullish or bearish next month. It however does not provide any stats how accurate model has been on test data.
- I am in general against putting money in open ended fund NFO. These NFOs are not like a stock IPO that it would get listed with some premium. So its generally better to wait to get it open. We can see the portfolio of fund and check if it fits in our overall portfolio. There is really no need to rush and put money in any open ended NFO.
- You have option of DSP Quant Fund if you want to add quantitative fund right now in your portfolio. It has at least six months track record. BTW, It would be good contest to watch these two funds for next six months and chose winner.
- Fund manager at this stage has not delivered any great performing funds.
- Machine learning based stock selection strategies are still evolving and do not have long term track records in India.
Hope this was helpful for you in some way. If you like the information, please subscribe to blog. This will help you get notified when a new article is written. It will also motivate me and keep me going. Also share blog within your network if you think it could benefit them.